Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease

Mohammed Saeed Alzahrani; Fawaz Waselallah Alsaade

doi:10.3390/agronomy13051184

Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease

Alzahrani, Mohammed Saeed;Alsaade, Fawaz Waselallah 2023-04-22 00:00:00 agronomy Article Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease Mohammed Saeed Alzahrani and Fawaz Waselallah Alsaade * College of Computer Science and Information Technology, King Faisal University, Al-Ahsa P.O. Box 4000, Saudi Arabia * Correspondence: falsaade@kfu.edu.sa Abstract: Plant diseases pose the greatest threat to food supply integrity, and it is a signiﬁcant challenge to identify plant diseases in their earliest stages to reduce the potential for associated economic damage. Through the use of computer vision, a cutting-edge artiﬁcial intelligence is offered as a solution to this problem so that tomato leaf diseases may be classiﬁed. The proposed network is expected to provide excellent results. Transfer learning is also used to make the model efﬁcient and cost effective. Since tomato diseases may signiﬁcantly inﬂuence crop output and quality, early identiﬁcation and diagnosis of these diseases are essential for successful treatment. Deep learning has shown a great deal of promise in plant disease identiﬁcation, providing excellent accuracy and efﬁciency. In this investigation, we compared the performance of three different deep learning models—DenseNet169, ResNet50V2, and a transform model, namely ViT, with regard to diagnosing diseases affecting tomatoes. Both diseased and healthy tomato samples were included in the dataset of photos of tomato diseases used for training and testing the models. The DenseNet121 model had the best results, with a training accuracy of (99.88%) and a testing accuracy of (99.00%). This gave it the greatest overall accuracy. Both the ResNet50V2 and VIT models attained high levels of accuracy, with testing accuracies of (95.60% and 98.00%), respectively. Our results demonstrate deep learning’s potential for accurate and efﬁcient tomato disease detection, which could aid in early disease management and ultimately improve crop yield and quality. The experimental ﬁndings show that the suggested ensemble models stand out due to the short amount of time required for Citation: Alzahrani, M.S.; Alsaade, training and testing as well as their exceptional classiﬁcation performances. Because of this study, F.W. Transform and Deep Learning professionals will be able to facilitate the early diagnosis of plant diseases in a straightforward and Algorithms for the Early Detection expedient way, thereby preventing the emergence of new infections. and Recognition of Tomato Leaf Disease. Agronomy 2023, 13, 1184. Keywords: artiﬁcial intelligence; deep leaning; transform learning; food security; plant diseases https://doi.org/10.3390/ agronomy13051184 Academic Editors: Louis Kouadio and Luís Manuel Navas Gracia 1. Introduction Received: 7 March 2023 Plants are crucial to our survival because they give us sustenance and shield us from Revised: 14 April 2023 radiation. Without plants, it is difﬁcult to imagine there ever being life on Earth; they Accepted: 18 April 2023 not only supply food for all terrestrial creatures but also shield the ozone layer, which Published: 22 April 2023 blocks harmful ultraviolet light. Tomatoes are globally grown vegetables because they are nutrient rich and safe for human consumption [1]. Around 160,000,000 metric tons of tomatoes are consumed annually across the globe [2]. Many people believe that tomato sales could help rural communities earn much-needed income, which would signiﬁcantly Copyright: © 2023 by the authors. impact poverty levels [3]. Tomatoes have a high concentration of nutrients and are grown Licensee MDPI, Basel, Switzerland. all over the world; as a result, they have a major impact on the agricultural economy when This article is an open access article cultivated and harvested. Tomatoes are beneﬁcial in preventing diseases, such as gingival distributed under the terms and bleeding, hypertension, and hepatitis, due to their pharmacological properties [1], and conditions of the Creative Commons their anti-cancer properties are well documented. Tomatoes are increasingly in demand Attribution (CC BY) license (https:// because of their growing popularity. Statistically, almost 80.00% of agricultural output [2] creativecommons.org/licenses/by/ is attributed to small farmers. However, diseases and pests wipe out approximately half 4.0/). Agronomy 2023, 13, 1184. https://doi.org/10.3390/agronomy13051184 https://www.mdpi.com/journal/agronomy Agronomy 2023, 13, 1184 2 of 24 of these farmers’ harvests every year. Research into ﬁeld crop disease detection is crucial because the diseases and insects that parasitize tomatoes have a profound impact on their growth. According to a 2020 study compiled by the Food and Agricultural Organization Corporate Statistical Database (FAOSTAT), 186,821 million tons of tomatoes were produced globally [1]. Agriculture is the fundamental contributor to economic expansion as well as the foundation upon which human civilization was built. The many different kinds of plant diseases that farmers have to protect their crops against have proven to be a signiﬁcant obstacle for agricultural production. Understanding how to prevent plant diseases and implementing preventative measures are the two most important factors in crop yield optimization. Early detection of plant diseases is essential if one wants to achieve maximum agricultural output while at the same time preserving ﬁnancial resources and reducing the amount of crop loss. The fact that everything is handled by computers also makes it simple to put into action. On the one hand, accurate diagnosis and categorization of diseases within a reasonable amount of time is highly important for preserving both the quality and quantity of tomatoes. Environmental conditions may play a role in the development of a wide range of plant diseases. The disease triangle is a conceptual model that represents the relationship between three crucial factors: the environment, the host, and the infectious agent. The disease triangle was developed to explain how diseases spread. This type was created in the 1950s and has been widely used since its introduction. It follows that disease cannot occur, since it is impossible for it to take place if any one of these three requirements is absent from the triangle. Many abiotic factors, including air movement, temperature, humidity, pH, and watering, could potentially have signiﬁcant effects on plants. Fungus, bacteria, and viruses are some organisms that may attack a plant. The infectious agent is an organism that causes the plant to become diseased. The term “host” refers to a plant that is infected with a disease. Disease development is the consequence of all the risk factors converging simultaneously [3]. In general, infections are deﬁned by symptoms that manifest by working their way up through the plant, and as a result, the majority of diseases have rapid transmission rates once they have infected a plant. Pathogenic fungi, bacteria, and viruses, as well as poor climatic conditions, are some of the many potential causes of plant diseases. Because of the presence of these diseases, a plant’s fundamental functions, such as photosynthesis, pollination, fertilization, and germination, may be disrupted. This emphasizes the critical need to accurately diagnose diseases as early as medical science currently allows. Technology has advanced to the point that it is possible to use tools to diagnose whether a plant is diseased, and if so, what kind of disease it has, rather than relying only on the judgment of a human expert. Results from procedures such as object recognition and classiﬁcation, as well as image processing and artiﬁcial intelligence algorithms, are becoming increasingly high quality as the quality of photos obtained by technological equipment continues to improve. Machine learning (ML) and deep learning (DL) have proved much more effective than traditional methods of optimization and prediction. First, unlike traditional techniques that require humans to extract characteristics and are limited by the amount of data, new systems may learn automatically from large quantities of data. Second, ML and DL models may generalize well to data that have not yet been seen, which is a signiﬁcant improvement on prior methods. In contrast to more traditional methods, machine learning and deep learning models may pick up on nonlinear and intricate con- nections in the data. Hence, ML excels at handling situations with many moving parts, especially those with complex interactions. Nowadays, AI (Artiﬁcial intelligence) is exten- sively used in many ﬁelds, including communication, building, magnetism, physics, and biology [4–8]. It is crucial to accurately identify plant diseases and classify them in a timely manner in this approach [9]. AI has progressed to the point that it can now automatically identify plant diseases from raw images [10,11]. To date, there has been much research on determining the causes of plant diseases. Most studies used preexisting datasets, models, and libraries to conduct their experiments. Agronomy 2023, 13, 1184 3 of 24 To automate the processes of identifying and classifying plant leaf diseases, Singh and Misra [12] devised an algorithm for the picture segmentation technique. Using a genetic algorithm, they achieved an overall success rate of 97.60% across ﬁve different diseases while trying to identify them. Zhang et al. [13] analyzed a database of cucumber leaf samples to identify diseases. In this study, the researchers segmented diseased leaves using k-means clustering to obtain shape and color information that could be utilized for disease diagnosis. When classifying these damaged leaves using the sparse representation method, an 86.00% accuracy rate was attained. Convolutional neural network (CNN) models, a kind of AI, can be used for disease detection and diagnosis in plants [14]. Model training was performed using a dataset that included a total of 87,848 images. The collection also includes 58 unique plant–disease combinations, which are applied to 25 different plant species. A 99.53% success rate was calculated from the data. Image processing [15,16], pattern recognition [17,18], and computer vision [19,20] have all seen rapid development and use in agriculture in recent years, with a particular emphasis on automating disease- and pest-detection processes. These tasks often cause problems due to the complexity and time commitment involved in preprocessing and constructing picture features for traditional computer vision models. The accuracy with which the feature extraction processes and with which the learning algorithm have been built also affects the efﬁcacy of these systems [21–23]. Deep learning is gaining momentum in sickness detection as a result of increased processing power, increased storage capacity, and the availability of large datasets. This method has recently been applied to the detection of plant diseases; a problem that has proven difﬁcult to solve. As a subset of machine learning, deep learning is a term with a speciﬁc meaning. CNNs are among the most popular algorithms used in deep learning for tasks, including picture classiﬁcation, object recognition, and semantic segmentation [24,25]. CNNs are helpful for discovering patterns in images, objects, and sceneries because they learn to classify based on data from the image, so the user is no longer required to laboriously isolate the desired features within the image [26,27]. This article will assess and explain a selection of the many different approaches to deep learning that are currently in use. Although a number of studies have been conducted to investigate the impact of disease detection on tomato crops, the existing model still requires improvement. Consequently, we proposed a CNN model with two convolution layers, two max pooling layers, a disease in tomato plants by analyzing data from the hidden layer, and a ﬂattening layer for tomato plants. Farmers can now solve problems independently without seeking advice from agriculture industry specialists. This entails identifying the many diseases that might affect crops. Our model was developed to assist in early-stage diagnoses of plant diseases. This will increase the overall productivity of agricultural activities, and as a result, food availability. The need for an automated technique for diagnosing diseases that might impact tomato plants is the key driver of this project. The following are some of this study’s contributions that ﬁll some of the gaps in existing research: Evaluating a wide variety of characteristics, including a crop’s production, yield capac- ity, grain quality, and nutrient retention, leads to an accurate plant disease diagnosis. Providing advice on constructing new CNN models and creating new ensemble structures using the proposed CNNs. Using an advanced transform learning lie to detect tomato leaf disease is critical to saving food security. Increasing the accuracy percentage of the proposed model in comparison to earlier studies in the ﬁeld that were reported in academic journals. Increasing categorization reliability and reducing the amount of time required for training and examinations. This paper presents an architecture for early disease identiﬁcation and classiﬁcation in tomato leaves that is based on three different deep learning models—DenseNet169, ResNet50V2, and transform ViT model and data augmentation. To properly forecast the Agronomy 2023, 13, 1184 4 of 24 kind of diseases that will impact the tomato leaves, this study aims to build a trustworthy framework for screening photographs of tomato leaves for indicators of disease. 2. Study Background Researchers from a wide variety of institutions have developed automated diseases detection systems using cutting-edge technologies, such as machine learning and neural network designs like Inception V3 net, VGG 16 net, and Squeeze Net. They use highly accurate procedures to diagnose plant diseases in tomato leaf tissues. In order to detect and categorize tomato diseases, a pretrained network model has an accuracy of 94.00% to 95.00% [28,29]. Using a dataset of 300 images and the Tree Classiﬁcation Model and Segmentation, six types of tomato leaf disease were identiﬁed and classiﬁed [30]. It has been proposed [31] that we can classify leaf-affecting plant diseases with 93.75% accuracy using a speciﬁc approach. Plant leaf diseases may be accurately identiﬁed and categorized using image processing software and a classiﬁcation scheme [32]. A smartphone with an 8-megapixel camera was used to photograph a variety of conditions and then divide the resulting data into two groups: healthy and sick. The image processing procedure was composed of three distinct actions: boosting contrast, image segmentation, and distinctive feature locating. An artiﬁcial neural network with several layers and a feed-forward neural net was used to conduct classiﬁcation tasks, and the results of these networks are compared. Compared to the ﬁndings obtained by the multilayer perceptron (MLP) network and the radial basis function (RBF) network, they are much more favorable. Over the course of the study, the image of the plant blade was dissected into healthy and sick sections. However, this did not allow pinpointing the cause of the problem. In order to diagnose leaf diseases, researchers used color space analysis, color time, a histogram, and color coherence as a classiﬁer, which achieved 87.2% in terms of accuracy. Researchers have used AlexNet and VGG 19 models with a frame size of 13,262 to detect diseases that are hurting tomato harvests. The model was utilized to achieve 97.49% accuracy [33]. To obtain a 95.00% detection rate for viruses that damage dairy crops [34], we used transfer learning and a CNN model. With the goal of identifying and categorizing the conditions of tomato plants’ leaf surfaces, an AlexNet-based deep learning mechanism using neural network-trained transfer learning achieved an accuracy of 95.75% [35,36]. Resnet-50 was created to identify 1000 unique diseases that may harm tomato leaves by presenting 3000 images, each of which was tagged with a disease name, such as “lesion blight”, “late blight”, or “yellow curl leaf”. The ﬁrst convolution layer ’s kernel size was increased to 11 11, and the comparing network activation function was switched to Leaky- ReLU. The model has been reﬁned over many iterations, and its performance in classifying diseases has improved to an accuracy of 98.30% and a precision of 98.00% [37]. Simpliﬁed eight-layer CNN models have been proposed for disease detection and classiﬁcation in tomato leaves [38]. This research used the PlantVillage dataset [39], a compilation of data on various agricultural products. By applying deep learning to the tomato leaf dataset, the author focused on disease diagnoses to improve performance. In recent years, CNNs have emerged as a reliable tool for diagnosing plant diseases [40,41]. Some research [42,43] has concentrated on improving feature detection quality by removing obstacles caused by inconsistencies in illumination and background homogeneity in high-stakes scenarios. Feature detection improvements have been the subject of other investigations that have sought to implement them by detecting complicated contexts. Few authors have developed real-time models to hasten plant disease detection [44,45]. Other authors’ work has also led to the early identiﬁcation of plant diseases via model development [46,47]. In Ref. [48], the authors investigate digital photos of tomato leaves to determine the presence of various diseases. The authors implement a classiﬁcation model based on CNN and AI-derived algorithms that is 96.55% accurate in recognizing ﬁve different diseases. In several studies, deep neural network models have been used to identify diseases in tomato leaves. In Ref. [49], for example, the authors compare four alternative models (LeNet, VGG16, ResNet, and Xception) and conclude that the VGG16 model Agronomy 2023, 13, 1184 5 of 24 achieves the greatest performance (99.25% accuracy) when used to categorize nine distinct disorders. The effectiveness of deep neural network models for diagnosing diseases in tomato leaves has been studied in other research. According to Ref. [50], an identical issue was solved with 95.00% or higher accuracy using the AlexNet, GoogleNet, and LeNet models. Agarwal et al. [51] built CNN architecture and compared it to other ML models (such as random forest and decision trees) and DL models for classifying data into 10 groups (VGG16, Incep-tionv3, and MobileNet). The result was a 99.20% accuracy boost. Many research efforts have focused on improving classiﬁcation accuracy by combining random forests, support vector machines, and multinomial logistic regression, which are just a few examples of the classiﬁcation networks that may be employed with the obtained leaf characteristics [52]. With the help of MobileNetv2 and NASNetMobile, we were able to successfully extract leaf features. Researchers have shown that classiﬁcation accuracy may be greatly improved by combining these two techniques. Many studies have successfully diagnosed plant diseases using algorithms such as Mask R-CNN [53]. Computing costs and model sizes have been decreased via the use of several techniques namely K-nearest neighbors and Gabor ﬁlters and K-nearest neighbors (KNN). Both methods have been used to attempt to reduce the time and resources needed to run deep learning calculations. To reduce computational costs, the authors of Ref. [54] used a SqueezeNet architecture with only 33 ﬁlters. YOLO-Tomato, which is based on YOLOv3 and was utilized by the authors of Refs. [55–57] to improve tomato identiﬁcation, has been presented as a solution for dealing with these issues. YOLOv3 has been designed with a thick architecture in order to make the reuse of features easier and to assist in learning a model that is both more accurate and more compact. 3. Materials and Methods Due to tomato leaves’ complicated designs and the wide variety of diseases that affect tomatoes, the disease identiﬁcation process can be difﬁcult. Deep learning has emerged as a strong technology in recent years as support for the computer-assisted detection of diseases affecting tomatoes. This technique uses deep neural networks, which can learn complicated patterns from extensive volumes of picture data. The recommended approach for detecting tomato diseases via deep learning includes several essential stages. To begin, a tomato leaf picture dataset is gathered. This dataset will comprise examples of various diseased and healthy leaves. After preprocessing, the photos are utilized to teach a deep neural network, such as a CNN, to recognize intricate patterns within images. After the model has been trained, it can categorize newly acquired pictures as belonging to either the healthy class or one of the disease classes. Adjustments to the model’s parameters and the inclusion of new data may both enhance the model’s performance and make it more accurate. In conclusion, model performance may be tested using a test set of pictures to determine its accuracy and generalization capacity. The fundamental architecture of the suggested system for the classiﬁcation and detection of tomato plant leaf diseases is presented in Figure 1. 3.1. Dataset This study used the Tomato Leaf Diseases Dataset, consisting of 11,000 images of tomato leaves affected by 10 distinct diseases. Each class, including tomato mosaic virus, target spot, bacterial spot, tomato yellow leaf curl virus, late blight, leaf mold, early blight, spider mites (two-spotted spider mite), tomato healthy, and Septoria leaf spot, contains 1100 images. The dataset is publicly accessible on Kaggle. Table 1 displays the tomato leaf features, and the sample numbers of each class, and a sample of the tomato leaf diseases is presented in Figure 2. The dataset is available here https://www.kaggle.com/datasets/ kaustubhb999/tomatoleaf (accessed on 20 January 2023). Agronomy 2023, 13, 1184 6 of 24 Agronomy 2023, 13, x FOR PEER REVIEW 6 of 26 Figure Figure 1. 1. The The structure of the structure of the suggested suggested approach. approach. Table 1. Tomato leaf features. 3.1. Dataset This study used the Tomato Leaf Diseases Dataset, consisting of 11,000 images of Class Number of Samples Description tomato leaves aﬀected by 10 distinct diseases. Each class, including tomato mosaic virus, This is a viral disease that can cause signiﬁcant damage to tomato plants, target spot, bacterial spot, tomato yellow leaf curl virus, late blight, leaf mold, early blight, leading to reduced growth and yield. Symptoms include mosaic-like patterns spider mites (two-spott ed spider mite), tomato healthy, and Septoria leaf spot, contains of light and dark green on leaves, stunted growth, and deformed fruit. The Mosaic virus 1100 1100 images. The dataset is publicly accessible on Kaggle. Table 1 displays the tomato leaf disease is transmitted by contact with infected plant material and can be managed using resistant tomato varieties and cultural practices, such features, and the sample numbers of each class, and a sample of the tomato leaf diseases as sanitation. is presented in Figure 2. The dataset is available here htt ps://www.kaggle.com/da- tasets/kaustubhb999/tomatoleaf (accessed on 20 January 2023). This disease is caused by the fungus Corynespora cassiicola and can cause circular, sunken lesions with concentric rings on leaves and stems. It can lead Target spot 1100 Table 1. Tomato leafto features. defoliation and reduced fruit yield, but it can be managed using fungicides and cultural practices, such as crop rotation and sanitation. Class Number of Samples Description This is a common disease that can cause severe damage to tomato plants, This is a viral disease that can cause signiﬁcant damage to tomato leading to reduced yields and poor fruit quality. Symptoms include brown, plants, leading to reduced growth and yield. Symptoms include mo- sunken lesions on leaves and fruit, which eventually turn black and crusty. Bacterial spot 1100 The disease is caused by the bacterium Xanthomonas campestris pv. vesicatoria saic-like patt erns of light and dark green on leaves, stunted growth, and Mosaic virus 1100 and can be controlled using copper-based fungicides and cultural practices, deformed fruit. The disease is transmitt ed by contact with infected such as crop rotation and sanitation. plant material and can be managed using resistant tomato varieties and This is a viral disease that can cause signiﬁcant damage to tomato plants, cultural practices, such as sanitation. leading to stunted growth and reduced fruit yield. Symptoms include leaf This disease is caused by the fungus Corynespora cassiicola and can Yellow leaf curl virus 1100 yellowing and curling as well as distorted fruit. The disease is transmitted by cause circular, sunken lesions with concentric rings on leaves and the whiteﬂy Bemisia tabaci, and it can be managed using insecticides, resistant Target spot 1100 stems. It can lead to defoliation and reduced fruit yield, but it can be tomato varieties, and cultural practices, such as crop rotation and sanitation. managed using fungicides and cultural practices, such as crop rotation This disease is caused by the fungus Phytophthora infestans and can cause rapid and sanitation. and devastating damage to tomato plants. Symptoms include water-soaked Late blight 1100 lesions on leaves and stems, which quickly turn brown and necrotic. The This is a common disease that can cause severe damage to tomato disease can be managed using fungicides and cultural practices, such as crop plants, leading to reduced yields and poor fruit quality. Symptoms in- rotation, sanitation, and removal of infected plant material. clude brown, sunken lesions on leaves and fruit, which eventually turn Bacterial spot 1100 This disease is caused by the fungus Passalora fulva and can cause signiﬁcant black and crusty. The disease is caused by the bacterium Xanthomonas damage to tomato leaves, reducing plant growth and yield. Symptoms include campestris pv. vesicatoria and can be controlled using copper-based fun- Leaf mold 1100 yellowing and brown lesions on the plant’s lower leaves. The disease can be gicides and cultural practices, such as crop rotation and sanitation. managed using fungicides and cultural practices, such as crop rotation and sanitation. Agronomy 2023, 13, 1184 7 of 24 Table 1. Cont. Class Number of Samples Description This disease is caused by the fungus Alternaria solani and is characterized by concentric rings of dark brown spots on the plant’s lower leaves. It can cause Early blight 1100 defoliation and reduced fruit yield but can be managed using fungicides and cultural practices, such as crop rotation and pruning. These are tiny arachnids that can cause signiﬁcant damage to tomato plants by feeding on the undersides of leaves, causing yellowing and stunted growth. Spider mites 1100 They can be managed using predatory mites, insecticidal soaps, and cultural practices, such as crop rotation and sanitation. This class contains images of healthy tomato leaves, which can be used as Tomato healthy 1100 references for comparison with diseased leaves. This disease is caused by the fungus Septoria lycopersici and is characterized by small, circular lesions with dark brown centers and yellow halos on the lower Septoria leaf spot 1100 leaves of the plant. It can cause signiﬁcant defoliation and yield loss, but can Agronomy 2023, 13, x FOR PEER REVIEW be managed using fungicides and cultural practices, such as crop rotation 8 of 26 and sanitation. Figure 2. Demonstrated dataset samples. Figure 2. Demonstrated dataset samples. 3.2. Data Preprocessing 3.2. Data Preprocessing The data preprocessing step is essential because it improves both the data quality and The data preprocessing step is essential because it improves both the data quality and overall performance of the classiﬁcation methods used for the picture categorization process. overall performance of the classiﬁcation methods used for the picture categorization pro- The primary goal of data preprocessing is to prepare images for the deep learning model cess. The primary goal of data preprocessing is to prepare images for the deep learning by removing any discrepancies, noise, or outliers that could negatively impact the model’s model by removing any discrepancies, noise, or outliers that could negatively impact the accuracy. This is accomplished by organizing data in a way that makes it easier for the model’s accuracy. This is accomplished by organizing data in a way that makes it easier model to analyze. In deep learning models for image classiﬁcation, several common data for the model to analyze. In deep learning models for image classiﬁcation, several com- preprocessing approaches are applied, such as image resizing and normalization. Image mon data preprocessing approaches are applied, such as image resizing and normaliza- tion. Image resizing involves adjusting the size of the images to ensure that they are of a standard size for the model’s training. This helps reduce model complexity and ensures that it is trained on images of the same size. Images normalization entails adjusting the brightness and contrast levels of each picture so that they are uniform across all of them. This is done to guarantee that the model is trained on photos that have comparable qual- ities. Applying data preprocessing techniques results in a more consistent and high-qual- ity dataset, which can improve the model’s accuracy. These techniques can be imple- mented using various image-processing libraries, such as PIL and OpenCV. 3.3. Transfer Learning Algorithms Transfer learning approaches allow a model to be trained on one task and then ap- plied to a diﬀerent but related task. This is particularly useful when there is a lack of la- beled data for a speciﬁc task or when the task is similar to one that has already been solved. It can save a lot of time and resources by leveraging the knowledge and features learned from a pretrained model. Additionally, transfer learning can improve model performance by reducing the risk of overﬁtt ing and allowing the model to be bett er generalized to a new dataset. Both the quantity of labeled data available for the new task and the degree of similarity between the pretrained model and the new task should be taken into account Agronomy 2023, 13, 1184 8 of 24 resizing involves adjusting the size of the images to ensure that they are of a standard size for the model’s training. This helps reduce model complexity and ensures that it is trained on images of the same size. Images normalization entails adjusting the brightness and contrast levels of each picture so that they are uniform across all of them. This is done to guarantee that the model is trained on photos that have comparable qualities. Applying data preprocessing techniques results in a more consistent and high-quality dataset, which can improve the model’s accuracy. These techniques can be implemented using various image-processing libraries, such as PIL and OpenCV. 3.3. Transfer Learning Algorithms Transfer learning approaches allow a model to be trained on one task and then applied to a different but related task. This is particularly useful when there is a lack of labeled data for a speciﬁc task or when the task is similar to one that has already been solved. It can save a lot of time and resources by leveraging the knowledge and features learned from a pretrained model. Additionally, transfer learning can improve model performance by reducing the risk of overﬁtting and allowing the model to be better generalized to a new dataset. Both the quantity of labeled data available for the new task and the degree of similarity between the pretrained model and the new task should be taken into account before deciding to use transfer learning. It may be preferable to retrain the model from scratch if the new job is quite different from the one the pretrained model was built for, or if there is a huge quantity of labeled data available. Model performance may be enhanced via transfer learning, and the need for labeled data can be reduced. It is applicable to many different models, including CNNs, and may be implemented using either feature-based or ﬁne-tuning approaches. 3.3.1. DenseNet121 Model DenseNet121 is a popular CNN architecture that Huang et al. introduced in 2017 as part of the DenseNet family. The network is based on the concept of dense connections between layers, which allow information to ﬂow more efﬁciently through the network and improve its accuracy. The DenseNet121 architecture consists of multiple dense blocks, each containing a set of convolutional layers that are densely connected to all subsequent layers [58–61]. In this research, we employ a pretrained DenseNet121 model for disease identiﬁcation in tomatoes. It is common practice to apply transfer learning, and this was done to initialize the pretrained model using weights learned from the huge ImageNet dataset. A new set of layers, consisting of a global average pooling layer, two fully connected layers with 512 and 256 neurons, and two batch normalization layers, was added after the top layer of the pretrained model was eliminated and replaced. Once the second layer was completely linked, the ReLU activation function was applied to a new layer that served as an activation layer. Finally, a softmax activation layer was added to the output layer, which comprised of 10 neurons that corresponded to the 10 different types of tomato diseases. Figure 3 illustrates this point. This architecture was chosen because DenseNet models have shown strong perfor- mance in various image classiﬁcation tasks and have a compact architecture suitable for transfer learning. The global average pooling layer was used to reduce the number of pa- rameters in the model, and batch normalization layers were added to improve the stability and convergence of the training process. The important parameters of DenseNet121 model are presented in Table 2. Agronomy 2023, 13, 1184 9 of 24 Table 2. Describe DenseNet121 model parameters. DenseNet121 Parameters Value Batch_Size 8 Optimizer SGD Learning Rate (lr) 0.0001 Momentum 0.9 Early Stopping (es) Yes Model Checkpoint (mc) Yes Agronomy 2023, 13, x FOR PEER REVIEW 10 of 26 Number of Epochs 100 Figure 3. DenseNet121 model for tomato disease detection. Figure 3. DenseNet121 model for tomato disease detection. 3.3.2. ResNet50V2 Model 3.3.2. ResNet50V2 Model ResNet50V2 is a convolutional neural network architecture that has been used for ResNet50V2 is a convolutional neural network architecture that has been used for computer vision for analysis and classiﬁcation images, including disease classiﬁcation in computer vision for analysis and classiﬁcation images, including disease classiﬁcation in plants. The pretrained model was initialized with weights learned from the large image plants. The pretrained model was initialized with weights learned from the large image pr processi ocessing ng data dataset, set, which which is is a a common pra common practice ctice i in n tra transfer nsfer l learning. earning. The top The top la layer yer of of the the pretrained model was removed and replaced with a new set of layers. The classiﬁer that pretrained model was removed and replaced with a new set of layers. The classiﬁer that we we ad added ded on on top top of the pr of the pretrained etrained ResNet50V2 ResNet50V2 mode modell is de is designed signed t to o ad adapt apt the n the network etwork to to a new im a new image age clclassiﬁcation assiﬁcation task. The p task. The retraine pretrained d model model output is output passed through a is passed through Glob- a alAveragePooling2D layer that reduces feature maps’ spatial dimensions by averaging GlobalAveragePooling2D layer that reduces feature maps’ spatial dimensions by averaging them them alon along g t the he height an height and d width width axe axes. s. This results This results in in a a ﬁxed-length ﬁxed-length fe featur ature e v vector ector tthat hat encodes encodes the the most most important features important features of of the the input im input image. age. The The feature v feature vector ector is proce is processed ssed by two fu by two fully lly connect connected ed lay layers ers w with ith 51 512 2 and 2 and 256 56 n neu- eu- rrons, respectiv ons, respectively ely, wh , which ich learn to map learn to map the the fe featur ature v e vector ector to a hig to a high-level h-level representation representation that captures the discriminative information for the new image classiﬁcation task. Batch that captures the discriminative information for the new image classiﬁcation task. Batch normalization normalization lay layers ers ar are e added added after after e each ach full fully y conn connected ected lay layer er to to stabilize stabilize the the trainin training g process and accelerate convergence. After the second fully connected layer, nonlinearity process and accelerate convergence. After the second fully connected layer, nonlinearity is introduced into the network by adding an activation layer using the ReLU activation is introduced into the network by adding an activation layer using the ReLU activation function. function. This al This allows lows the network to the network to acquir acquire more e more nuanced repre nuanced representations. sentations. Finally, the second fully connected layer ’s output is passed through a fully connected Finally, the second fully connected layer’s output is passed through a fully connected layer with 10 neurons and a softmax activation function, which outputs a probability layer with 10 neurons and a softmax activation function, which outputs a probability dis- distribution over the 10 classes in the dataset. This layer learns to map the high-level tribution over the 10 classes in the dataset. This layer learns to map the high-level repre- representation to the ﬁnal output of the network, which is the predicted class label for the sentation to the ﬁnal output of the network, which is the predicted class label for the input image. The structure of ResNet50V2 model to detect tomato diseases is presented in Fig- ure 4. The parameters of developing the ResNet50V2 model for detecting tomato disease are presented in Table 3. Agronomy 2023, 13, 1184 10 of 24 Agronomy 2023, 13, x FOR PEER REVIEW 11 of 26 input image. The structure of ResNet50V2 model to detect tomato diseases is presented in Figure 4. The parameters of developing the ResNet50V2 model for detecting tomato disease are presented in Table 3. Figure 4. DenseNet121 model for tomato disease detection. Figure 4. DenseNet121 model for tomato disease detection. T Table able 3. 3. Describe Describe ResNet50V2 model parameters. ResNet50V2 model parameters. Parameters Values Parameters Values #Optimizer SGD #Optimizer SGD #Learning Rate (lr ) 0.0001 #Learning Rate (lr) 0.0001 #Momentum 0.9 #Momentum 0.9 #Early Stopping (es) Yes #Early Stopping (es) Yes #Model Checkpoint (mc) Yes #Model Checkpoint (mc) Yes #Number of Epochs 100 #Number of Epochs 100 3.3.3. Vision Transformer 3.3.3. Vision Transformer Vision transformer (VIT) is a new architecture for image classiﬁcation that has gained Vision transformer (VIT) is a new architecture for image classiﬁcation that has gained signiﬁcant attention in the deep learning community. Unlike traditional CNNs, VITs use signiﬁcant att ention in the deep learning community. Unlike traditional CNNs, VITs use transformer-based architecture, which has been highly successful in natural language transformer-based architecture, which has been highly successful in natural language pro- processing tasks. cessing tasks. VIT architecture consists of several key components. First, the input image is divided VIT architecture consists of several key components. First, the input image is divided into a sequence of ﬁxed-sized patches. Each patch is then ﬂattened into a 1D vector and into a sequence of ﬁxed-sized patches. Each patch is then ﬂatt ened into a 1D vector and passed through an embedding layer, which maps the patch to a higher-dimensional feature passed through an embedding layer, which maps the patch to a higher-dimensional fea- space. These patch embeddings are then fed into a transformer encoder, which applies ture space. These patch embeddings are then fed into a transformer encoder, which ap- a series of self-attention mechanisms to learn the contextual relationships between the plies a series of self-att ention mechanisms to learn the contextual relationships between patches, as shown in Figure 5. the patches, as shown in Figure 5. The following steps outline the process we used to train our data and input it into the VIT model for classiﬁcation. The input image is passed through a “patches” layer, which divides the image into a grid of non-overlapping 6 6 patches. This turns the 2D image into a 3D tensor of shape (batch size, patch size patch size number of channels) and the number of color channels (typically 3 for RGB images). The patch tensor is then fed through a “patch encoder” layer, which applies a learned linear transformation (via a dense layer) to each patch and adds a learnable position embedding to each patch. This step helps the model capture spatial relationships between patches and makes it aware of their relative positions within the image. Agronomy 2023, 13, x FOR PEER REVIEW 12 of 26 Agronomy 2023, 13, 1184 11 of 24 Figure 5. Image after transformer encoder. Figure 5. Image after transformer encoder. This layer produces the model’s predictions for each input image. During training, The following steps outline the process we used to train our data and input it into the the model is optimized using the AdamW optimizer, a variant of the Adam optimizer VIT model for classiﬁcation. The input image is passed through a “patches” layer, which that includes weight decay regularization. The loss function used is sparse categorical divides the image into a grid of non-overlapping 6 × 6 patches. This turns the 2D image Agronomy 2023, 13, x FOR PEER REVIEW 13 of 26 cross-entropy, which is a common choice for multiclass classiﬁcation problems. Figure 6 into a 3D tensor of shape (batch size, patch size × patch size × number of channels) and shows the structure of ViT transform model used to detect tomato disease. The signiﬁcant the number of color channels (typically 3 for RGB images). The patch tensor is then fed parameters of ViT model are shown in Table 4. through a “patch encoder” layer, which applies a learned linear transformation (via a dense layer) to each patch and adds a learnable position embedding to each patch. This step helps the model capture spatial relationships between patches and makes it aware of their relative positions within the image. This layer produces the model’s predictions for each input image. During training, the model is optimized using the AdamW optimizer, a variant of the Adam optimizer that includes weight decay regularization. The loss function used is sparse categorical cross- entropy, which is a common choice for multiclass classiﬁcation problems. Figure 6 shows the structure of ViT transform model used to detect tomato disease. The signiﬁcant pa- rameters of ViT model are shown in Table 4. Figure 6. ViT model for tomato disease detection. Figure 6. ViT model for tomato disease detection. Table 4. Describe ViT model parameters. Parameter Value Learning_Rate 0.001 Weight_Decay 0.0001 Batch_Size 128 Number_Epochs 50 Image_Size 72 Patch_Size 6 Number_Patches 144 Projection_Dimension 64 Number_Heads 4 Transformer_Units (128, 64) 4. Experiment The eﬀectiveness of DenseNet121, ResNet50V2, and VIT, three deep learning classi- ﬁcation algorithms, was assessed within the context of this research as they relate to the diagnosis of tomato diseases. The models were trained, tested, and validated using a da- taset that included photographs of both healthy and sick tomatoes. The conﬁguration for training a model was done on a laptop equipped with an 8th generation Core i7 CPU and a GPU 1070 utilized and 8GBs RAM. The TensorFlow and transformer architecture library was used for developing the models. Both the training and testing sets were created from the dataset utilized for this investigation. These two sets were then compared. This divi- sion’s mission was to verify that the deep learning models generated for this research could correctly categorize new photographs. To do this, every category in the dataset was broken up into 1000 photographs that would be used for training, and 100 images that would be used for testing. The primary goal of this division was to supply the deep learn- ing models with a diverse set of photographs from which they could learn, with the end goal of evaluating the models’ capacity to generalize their ﬁndings to images that they had not seen before. Agronomy 2023, 13, 1184 12 of 24 Table 4. Describe ViT model parameters. Parameter Value Learning_Rate 0.001 Weight_Decay 0.0001 Batch_Size 128 Number_Epochs 50 Image_Size 72 Patch_Size 6 Number_Patches 144 Projection_Dimension 64 Number_Heads 4 Transformer_Units (128, 64) 4. Experiment The effectiveness of DenseNet121, ResNet50V2, and VIT, three deep learning clas- siﬁcation algorithms, was assessed within the context of this research as they relate to the diagnosis of tomato diseases. The models were trained, tested, and validated using a dataset that included photographs of both healthy and sick tomatoes. The conﬁguration for training a model was done on a laptop equipped with an 8th generation Core i7 CPU and a GPU 1070 utilized and 8GBs RAM. The TensorFlow and transformer architecture library was used for developing the models. Both the training and testing sets were created from the dataset utilized for this investigation. These two sets were then compared. This division’s mission was to verify that the deep learning models generated for this research could correctly categorize new photographs. To do this, every category in the dataset was broken up into 1000 photographs that would be used for training, and 100 images that would be used for testing. The primary goal of this division was to supply the deep learning models with a diverse set of photographs from which they could learn, with the end goal of evaluating the models’ capacity to generalize their ﬁndings to images that they had not seen before. 4.1. Performance Measurement To measure the success of a deep learning model, its accuracy is often measured against a gold standard. It is a measure of the model’s predictive efﬁcacy that may be calculated by dividing the number of accurate forecasts by the total number of predictions. TP + TN Accuracy = 100% (1) FP + FN + TP + TN It is standard practice to utilize precision as a measure of a deep learning model’s success, particularly when the job at hand is one of categorization. This metric assesses the reliability of the model’s positive predictions by dividing the number of correct forecasts by the total number of correct predictions. It can be collated by Formula (2). TP Precision = 100 % (2) TP + FP Deep learning models make extensive use of the assessment statistic known as recall, particularly when dealing with categorization issues. It provides a numerical representation of the fraction of the dataset’s positive instances that correspond to accurate positive predictions generated by the model. In plainer language, it calculates the number of true positive predictions as a fraction of the total number of positive cases included in the dataset. TP Recall = 100 (3) TP + FN The F1 score is a commonly used evaluation metric in deep learning models, especially for classiﬁcation problems. It is a way to balance the trade-off between precision and recall Agronomy 2023, 13, 1184 13 of 24 by taking the harmonic mean of the two metrics. This makes it a useful metric when dealing with imbalanced datasets or when trying to identify rare classes. precision Sensitivity F1 score = 2 100 (4) precision + Sensitivity where: FP: false positive TP: True positive FN: false negative TN: true negative. 4.2. Results 4.2.1. Results of the DenseNet121 Model Table 5 shows the DenseNet121 model’s high performance in 90% training and 10 testing. The DenseNet121 model was trained using eight batch sizes and the SGD optimizer for 100 epochs. To prevent overﬁtting and minimize training time, the early-stop method was employed, which ended the training after 34 epochs. DenseNet121 scored high in accuracy (99.88%) in the training phase, whereas its testing accuracy was 0.99%. Table S1 shows the result of DenseNet121 with 90% training and 10 testing; it is observed that the accuracy of the DenseNet121model is (98%). Table 5. Results of DenseNet121 model (90% training and 10% testing). Class Name Precision% Recall% F1-Score% Support Bacterial_Spot 96 100 98 100 Early_Blight 99 100 00 100 Late_Blight 100 100 100 100 Leaf_Mold 100 100 100 100 Septoria_Leaf_Spot 97 100 99 100 Spider_Mites_Two-Spotted_Spider_Mite 99 99 99 100 Target_Spot 100 93 96 100 Yellow_Leaf_Curl Virus 100 98 99 100 Mosaic_Virus 99 100 100 100 Healthy 00 100 100 100 Accuracy of DenseNet121 99 1000 DenseNet121_Macro Avg 99 99 99 1000 DenseNet121_Weighted Avg 99 99 99 1000 Figure 7 illustrates how well DenseNet121 model performed throughout the 90% training and 10% validation phases for the purpose of recognizing and categorizing diseases that may affect tomato leaf tissue. A training accuracy beginning at 72.00% and increasing to 99.98% was attained as a consequence of these ﬁndings. According to the graphic representation, the accuracy at the validation phase started at 86.00% and reached 99.00%. The accuracy loss of the DenseNet121 model was 0.2 in validation. The performance of the DenseNet121 model with 80% training and 20% validation phases is presented in Figure S1; it is noted that the mode achieved accuracy 98.45% and validation loss was decreased from 0.8 to 0.1. The proposed Dense-Net121 model and its evaluation yielded a confusion matrix, as shown in Figures 8 and S2. The evaluation parameters were (90%, 80% as training and 10%, 20% as testing). The confusion matrix displays the tTP), TN, FP, and FN values that were gathered for each class. The target_spot class has high environment misclassiﬁcation based on the multiple-graphics processing unit (MGPU) architecture. In terms of leaf disease categorization, the suggested model attained a validation accuracy of 99.64% with 90% training and 10% testing whereas the Dense-Net121 model was attained accuracy 98.45% with 80% training and 20% testing. Agronomy 2023, 13, x FOR PEER REVIEW 15 of 26 Septoria_Leaf_Spot 97 100 99 100 Spider_Mites_Two-Spotted_Spider_Mite 99 99 99 100 Target_Spot 100 93 96 100 Yellow_Leaf_Curl Virus 100 98 99 100 Mosaic_Virus 99 100 100 100 Healthy 00 100 100 100 Accuracy of DenseNet121 99 1000 DenseNet121_Macro Avg 99 99 99 1000 DenseNet121_Weighted Avg 99 99 99 1000 Figure 7 illustrates how well DenseNet121 model performed throughout the 90% training and 10% validation phases for the purpose of recognizing and categorizing dis- eases that may aﬀect tomato leaf tissue. A training accuracy beginning at 72.00% and in- creasing to 99.98% was att ained as a consequence of these ﬁndings. According to the graphic representation, the accuracy at the validation phase started at 86.00% and reached 99.00%. The accuracy loss of the DenseNet121 model was 0.2 in validation. The perfor- mance of the DenseNet121 model with 80% training and 20% validation phases is pre- Agronomy 2023, 13, 1184 14 of 24 sented in Figure S1; it is noted that the mode achieved accuracy 98.45% and validation loss was decreased from 0.8 to 0.1. Agronomy 2023, 13, x FOR PEER REVIEW 16 of 26 Figure 7. Performance of DenseNet121 model; (a) DenseNet121 model accuracy and (b) Dense- Figure 7. Performance of DenseNet121 model; (a) DenseNet121 model accuracy and (b) DenseNet121 Net121 mode loss (90% training and 10% testing). mode loss (90% training and 10% testing). The proposed Dense-Net121 model and its evaluation yielded a confusion matrix, as shown in Figures 8 and S2. The evaluation parameters were (90%, 80% as training and 10%, 20% as testing). The confusion matrix displays the tTP), TN, FP, and FN values that were gathered for each class. The target_spot class has high environment misclassiﬁcation based on the multiple-graphics processing unit (MGPU) architecture. In terms of leaf dis- ease categorization, the suggested model att ained a validation accuracy of 99.64% with 90% training and 10% testing whereas the Dense-Net121 model was att ained accuracy 98.45% with 80% training and 20% testing. Figure Figure 8. 8. Confusion Confusion matrix of matrix of DenseNet121 DenseNet121 ( (90% 90% training a training and nd 10% te 10% testing). sting). 4.2.2. Results of the ResNet50V2 Model 4.2.2. Results of the ResNet50V2 Model The results of the ResNet50V2 model with 90% training and 10% testing are pre- The results of the ResNet50V2 model with 90% training and 10% testing are pre- sented in Table 6. The ResNet50V2 model was trained using eight batch sizes and the SGD sented in Table 6. The ResNet50V2 model was trained using eight batch sizes and the SGD optimizer for 100 epochs. To prevent overﬁtting and minimize training time, the early- optimizer for 100 epochs. To prevent overﬁtt ing and minimize training time, the early- stop method was employed, which ended the training after 20 epochs. The ResNet50V2 stop method was employed, which ended the training after 20 epochs. The ResNet50V2 model achieved accuracies of 99.49% in training and 95.60% in testing. The results of model achieved accuracies of 99.49% in training and 95.60% in testing. The results of Res- Net50V2 with 80% training and 20% testing is shown Table S2; the ResNet50V2 success- fully achieved 94.31%. Table 6. Results of ResNet50V2 model (90% training and 10% testing). Class Name Precision (%) Recall (%) F1-Score (%) Support Bacterial_Spot 99 100 100 100 Early_Blight 97 93 95 100 Late_Blight 88 100 93 100 Leaf_Mold 96 95 95 100 Septoria_Leaf_Spot 96 96 96 100 Spider_Mites_Two-Spotted_Spider_Mite 96 90 93 100 Target_Spot 98 87 92 100 Yellow_Leaf_Curl Virus 100 98 99 100 Mosaic_Virus 89 100 94 100 Healthy 00 97 98 100 Accuracy 96 1000 DenseNet121_Macro Avg 0.96 96 96 1000 DenseNet121_Weighted Avg 0.96 96 96 1000 Agronomy 2023, 13, 1184 15 of 24 ResNet50V2 with 80% training and 20% testing is shown Table S2; the ResNet50V2 success- fully achieved 94.31%. Table 6. Results of ResNet50V2 model (90% training and 10% testing). Class Name Precision (%) Recall (%) F1-Score (%) Support Bacterial_Spot 99 100 100 100 Early_Blight 97 93 95 100 Late_Blight 88 100 93 100 Leaf_Mold 96 95 95 100 Septoria_Leaf_Spot 96 96 96 100 Spider_Mites_Two-Spotted_Spider_Mite 96 90 93 100 Target_Spot 98 87 92 100 Agronomy 2023, 13, x FOR PEER REVIEW 17 of 26 Yellow_Leaf_Curl Virus 100 98 99 100 Mosaic_Virus 89 100 94 100 Healthy 00 97 98 100 Accuracy 96 1000 Figure 9 illustrates the eﬀectiveness of the ResNet50V2 model in identifying tomato DenseNet121_Macro Avg 0.96 96 96 1000 leaf diseases by dividing the dataset into 90% training and 10% testing. ResNet50V2′s val- DenseNet121_Weighted Avg 0.96 96 96 1000 idation accuracy is 96.00%, whereas its training accuracy was initially 78.00% and reached a maximum of 99.49%. At this point in the validation process, the ResNet50V2 model’s accuracy loss is less than 0.1. Figure 9 illustrates the effectiveness of the ResNet50V2 model in identifying tomato The eﬃciency of the ResNet50V2 model in determining the presence of tomato leaf leaf diseases by dividing the dataset into 90% training and 10% testing. ResNet50V2 s diseases with 80% training and 20% testing is seen in Figure S3. The accuracy of Res- validation accuracy is 96.00%, whereas its training accuracy was initially 78.00% and Net50V2 model in validation test is 94.30%. At this stage in the validation process, the reached a maximum of 99.49%. At this point in the validation process, the ResNet50V2 accuracy loss associated with the ResNet50V2 model is less than 0.2. model’s accuracy loss is less than 0.1. Figure 9. Performance of the ResNet50V2 model; (a) ResNet50V2model accuracy and (b) Res- Figure 9. Performance of the ResNet50V2 model; (a) ResNet50V2model accuracy and Net50V21 mode loss (90% training and 10% testing). (b) ResNet50V21 mode loss (90% training and 10% testing). The The Res efﬁciency Net50V of 2 mod the ResNet50V2 el’s confusion mat model rix i in s determining shown in Fig the ures pr10 and S4 b esence of tomato y dividing leaf diseases the dataset with in 80% 90%, tr 8a 0 i% ning as t and rain20% ing and testing 10%, is 2 seen 0% a in s t Figur esting. e S3. Thi The s maccuracy atrix dem of onst ResNet50V2 rates that model the model has in validation a strong posi test is 94.30%. tive imAt pact on toma this stage in to di thesea validation se detecti pr oocess, n and tha the t accuracy it misclaloss ssi- associated ﬁes 42 out of with every the10 ResNet50V2 00 samples. The suggest model is lessed than ResNet50V2 mode 0.2. l can predict the tested The ResNet50V2 model’s confusion matrix is shown in Figures 10 and S4 by dividing classes using the training dataset with 99.49% accuracy, as shown by the results repre- the dataset in 90%, 80% as training and 10%, 20% as testing. This matrix demonstrates that sented in the confusion matrix, 90% as training and 10% testing, whereas the misclassiﬁ- the model has a strong positive impact on tomato disease detection and that it misclassiﬁes cation of ResNet50V2 model with 80% training and 20% testing is 54 simples for all the 42 out of every 1000 samples. The suggested ResNet50V2 model can predict the tested classes. classes using the training dataset with 99.49% accuracy, as shown by the results represented This is the case because the results show that the model was able to predict the classes correctly. Examining the ﬁndings enables us to make this observation for ourselves. The confusion matrix that DenseNet121 produces demonstrates that the model has a strong positive impact on tomato disease detection, with only 10 misclassiﬁed samples out of a total of 1000 samples. Agronomy 2023, 13, 1184 16 of 24 Agronomy 2023, 13, x FOR PEER REVIEW 18 of 26 in the confusion matrix, 90% as training and 10% testing, whereas the misclassiﬁcation of ResNet50V2 model with 80% training and 20% testing is 54 simples for all the classes. Figure 10. The ResNet50V2 model’s confusion matrix (90% training and 10% testing). Figure 10. The ResNet50V2 model’s confusion matrix (90% training and 10% testing). This is the case because the results show that the model was able to predict the classes 4.2.3. VIT Model Results correctly. Examining the ﬁndings enables us to make this observation for ourselves. The The results achieved by the VIT model’s classiﬁcation performance for each class de- confusion matrix that DenseNet121 produces demonstrates that the model has a strong scribed with 9% training and 210% testing within the experimental dataset are shown in positive impact on tomato disease detection, with only 10 misclassiﬁed samples out of a Table 7. Because of this, we are able to deduce the performance of the suggested model, total of 1000 samples. which has the ability to categorize diseases with an accuracy that is more than 98.00%. Based on the data that are reﬂected in the table, it can be seen that the resulting recall 4.2.3. VIT Model Results metric value is high for each category that is speciﬁed in the dataset. The VIT model was The results achieved by the VIT model’s classiﬁcation performance for each class trained using eight batch sizes and the SGD optimizer for 100 epochs. The model achieved described with 9% training and 210% testing within the experimental dataset are shown in 98.00% accuracy on the training and testing data. The results of the VIT model, by splitt ing Table 7. Because of this, we are able to deduce the performance of the suggested model, the dataset into 80% training and 20% testing, are shown in Table S3. which has the ability to categorize diseases with an accuracy that is more than 98.00%. Based on the data that are reﬂected in the table, it can be seen that the resulting recall metric Table 7. Results of ViT model (90% training and 10% testing). value is high for each category that is speciﬁed in the dataset. The VIT model was trained using eight batch sizes and the SGD optimizer for 100 epochs. The model achieved 98.00% Class Name Precision (%) Recall (%) F1-Score (%) Support accuracy on the training and testing data. The results of the VIT model, by splitting the Bacterial_Spot 99 94 96 100 dataset into 80% training and 20% testing, are shown in Table S3. Early_Blight 97 99 98 100 Late_Blight 98 100 99 100 Leaf_Mold 99 96 97 100 Septoria_Leaf_Spot 99 99 99 100 Agronomy 2023, 13, 1184 17 of 24 Table 7. Results of ViT model (90% training and 10% testing). Class Name Precision (%) Recall (%) F1-Score (%) Support Agronomy 2023, 13, x FOR PEER REVIEW 19 of 26 Bacterial_Spot 99 94 96 100 Early_Blight 97 99 98 100 Late_Blight 98 100 99 100 Leaf_Mold 99 96 97 100 Spider_Mites_Two-Spotted_Spider_Mite 00 98 99 100 Septoria_Leaf_Spot 99 99 99 100 Target_Spot 95 99 97 100 Spider_Mites_Two-Spotted_Spider_Mite 00 98 99 100 Yellow_Leaf_Curl Virus 98 97 97 100 Target_Spot 95 99 97 100 Mosaic_Virus 100 100 100 100 Yellow_Leaf_Curl Virus 98 97 97 100 Mosaic_Virus 100 100 100 100 Healthy 95 98 97 100 Healthy 95 98 97 100 Accuracy Accuracy DenseNet121_Macro Avg 98 1000 DenseNet121_Macro Avg 98 1000 DenseNet121_Weighted Avg 98 98 98 1000 DenseNet121_Weighted Avg 98 98 98 1000 Bacterial_SpotBacterial_Spot 9 98 8 98 98 9898 10001000 Figure 11 shows a comparison of the training and validation accuracies by dividing Figure 11 shows a comparison of the training and validation accuracies by dividing the dataset in 90% training and 10% testing. Figure 11a demonstrates that an accuracy rate the dataset in 90% training and 10% testing. Figure 11a demonstrates that an accuracy rate of 98.00% may be obtained by sett ing the training stop at 100 epochs and keeping the rate of 98.00% may be obtained by setting the training stop at 100 epochs and keeping the rate of learning at 0.0001. Figure 11b compares validation and training losses. Consequently, of learning at 0.0001. Figure 11b compares validation and training losses. Consequently, given the study’s methodology, it is reasonable to predict that increasing the number of given the study’s methodology, it is reasonable to predict that increasing the number of iterations would lead to an increase in data accuracy. Conversely, if the training phase iterations would lead to an increase in data accuracy. Conversely, if the training phase duration increases, so will the total number of epochs. Figure S5 shows the accuracy and duration increases, so will the total number of epochs. Figure S5 shows the accuracy and loss performance of the VIT model using 80% training and 20% testing. loss performance of the VIT model using 80% training and 20% testing. Figure 11. VIT model performance; (a) VIT model accuracy and (b) VIT mode loss (90% training and Figure 11. VIT model performance; (a) VIT model accuracy and (b) VIT mode loss (90% training and 10% testing). 10% testing). As seen in Figure 12, the confusion matrix was devised to provide a visual represen- As seen in Figure 12, the confusion matrix was devised to provide a visual repre- tation of the classiﬁcation performance of the tomato leaf dataset obtained from the VIT sentation of the classiﬁcation performance of the tomato leaf dataset obtained from the model, or the number of images that were correctly categorized in a given number of iter- VIT model, or the number of images that were correctly categorized in a given number ations during validation and training, respectively. According to the VIT model’s confu- of iterations during validation and training, respectively. According to the VIT model’s sion matrix, the model has a strong positive impact on tomato disease detection, with only confusion matrix, the model has a strong positive impact on tomato disease detection, with 20 misclassiﬁed samples out of 1000 samples. Figure S6 shows the confusion matrix of VIT only 20 misclassiﬁed samples out of 1000 samples. Figure S6 shows the confusion matrix of model in 80% training and 20% testing. VIT model in 80% training and 20% testing. Agronomy 2023, 13, x FOR PEER REVIEW 20 of 26 Agronomy 2023, 13, 1184 18 of 24 Figure 12. Confusion matrix of ViT model (90% training and 10% testing). Figure 12. Confusion matrix of ViT model (90% training and 10% testing). 5. Discussion 5. Discussion Tomato is one of the most important crops worldwide and has high economic and Tomato is one of the most important crops worldwide and has high economic and nutritional value. However, it is often affected by various diseases, which can reduce both nutritional value. However, it is often aﬀected by various diseases, which can reduce both yield and quality. Early detection and accurate diagnosis of tomato diseases are crucial for yield and quality. Early detection and accurate diagnosis of tomato diseases are crucial effective disease management and prevention. for eﬀective disease management and prevention. This research further conﬁrms the effectiveness of proposed approaches to tomato This research further conﬁrms the eﬀectiveness of proposed approaches to tomato disease detection and provides valuable insight into different models’ performance. The disease detection and provides valuable insight into diﬀerent models’ performance. The results of this study have shown signiﬁcant accuracy for developing of automated systems results of this study have shown signiﬁcant accuracy for developing of automated systems that can detect and manage tomato plant diseases in the early stages, ultimately improving that can detect and manage tomato plant diseases in the early stages, ultimately improv- the efﬁciency and sustainability of tomato production. ing the eﬃciency and sustainability of tomato production. To diagnose tomato leaf diseases, this research examined multiple CNN models that To diagnose tomato leaf diseases, this research examined multiple CNN models that had been pretrained on the ImageNet dataset and then compared those models to the had been pretrained on the ImageNet dataset and then compared those models to the dataset. Three alternative CNN models—DenseNet121, ResNet50V2, and VIT—were dataset. Three alternative CNN models—DenseNet121, ResNet50V2, and VIT—were trained via transfer learning. Each model was educated and validated using the same trained via transfer learning. Each model was educated and validated using the same col- collection of tomato disease images, which included both infected and healthy examples. lection of tomato disease images, which included both infected and healthy examples. Ta- Table 8 shows the ﬁnal result of the proposed deep leaning model at training and testing ble 8 shows the ﬁnal result of the proposed deep leaning model at training and testing phase for detecting tomato disease. According to the ﬁndings, DenseNet121 successfully phase for detecting tomato disease. According to the ﬁndings, DenseNet121 successfully attained the highest possible training (99.88%) and test (99.00%) accuracies. In addition, its att ained the highest possible training (99.88%) and test (99.00%) accuracies. In addition, recall, accuracy, and F1 score were all above average, reaching 99.00%, 99.00%, and 98.99%, its recall, accuracy, and F1 score were all above average, reaching 99.00%, 99.00%, and respectively. ResNet50V2 obtained a training accuracy of 99.49% and a test accuracy 98.99%, respectively. ResNet50V2 obtained a training accuracy of 99.49% and a test accu- of 95.60%, both of which were lower than its predecessor. Its recall, accuracy, and F1 racy of 95.60%, both of which were lower than its predecessor. Its recall, accuracy, and F1 score were all lower than those of DenseNet121, reaching 95.60%, 94.80%, and 95.59%, score were all lower than those of DenseNet121, reaching 95.60%, 94.80%, and 95.59%, respectively. The accuracy of the VIT’s training and tests was also 98.00%. It had a recall of respectively. The accuracy of the VIT’s training and tests was also 98.00%. It had a recall 98.00%, a precision score of 98.00%, and an F1 score of 98.00%. The ﬁndings indicate that of 98.00%, a precision score of 98.00%, and an F1 score of 98.00%. The ﬁndings indicate DenseNet121 performed the best overall in identifying tomato diseases, followed by VIT that DenseNet121 performed the best overall in identifying tomato diseases, followed by and then ResNet50V2. The ROC plot of DenseNet121 is presented in Figure 13. VIT and then ResNet50V2. The ROC plot of DenseNet121 is presented in Figure 13. Agronomy 2023, 13, x FOR PEER REVIEW 21 of 26 Agronomy 2023, 13, 1184 19 of 24 Table 8. Summarized results of proposed deep learning in training and testing phases. Table 8. Summarized results of proposed deep learning in training and testing phases. Performance of Model Using 90% Training and 10% Testing of Dataset Performance of Model Using 90% Training and 10% Testing of Dataset Model Name Training Accuracy% Test Accuracy% Recall% Precision% F1 Score% Model Name Training Accuracy% Test Accuracy% Recall% Precision% F1 Score% DenseNet121 99.88 99 99 99 98.99 DenseNet121 99.88 99 99 99 98.99 ResNet50V2 99.49 95.60 95.60 95.8 95.59 ResNet50V2 99.49 95.60 95.60 95.8 95.59 ViT 98 98 98 98 97 ViT 98 98 98 98 97 Performance of Model Using 80% Training and 20% Testing of Dataset Performance of Model Using 80% Training and 20% Testing of Dataset Model Name Training Accuracy% Test Accuracy% Recall% Precision% F1 Score% Model Name Training Accuracy% Test Accuracy% Recall% Precision% F1 Score% DenseNet121 99.99 98.45 98.45 94.32 98.45 DenseNet121 99.99 98.45 98.45 94.32 98.45 ResNet50V2 99.99 94.43 94.31 98.46 94.31 ResNet50V2 99.99 94.43 94.31 98.46 94.31 ViT 99.27 94.90 94.90 94.97 94.89 ViT 99.27 94.90 94.90 94.97 94.89 Figure Figure 13. 13. ROC ROC curv curve e DenseNet121 DenseNet121 model model. . Additionally, a contrast was drawn between the level of difﬁculty provided by the Additionally, a contrast was drawn between the level of diﬃculty provided by the suggested model and that of a number of other models. The section devoted to accuracy suggested model and that of a number of other models. The section devoted to accuracy rates shows the individual successes of tomato leaf diseases in the datasets found in the rates shows the individual successes of tomato leaf diseases in the datasets found in the literature. The results of the proposed DenseNet121 model, which had the best rate of accu- literature. The results of the proposed DenseNet121 model, which had the best rate of racy (99.88%) in studies using a variety of methods other than the research recommended, accuracy (99.88%) in studies using a variety of methods other than the research recom- are shown in Table 9 and Figure 14. mended, are shown in Table 9 and Figure 14. Table 9. Comparison of results of DenseNet121 against different CNN models. Table 9. Comparison of results of DenseNet121 against diﬀerent CNN models. References Model Dataset Accuracy% References Model Dataset Accuracy% Ref. [62] Inception-v3 Same dataset 96.60 Ref. [62] Inception-v3 Same dataset 96.60 Ref. [63] ResNet-50 98.77 Ref. [63] ResNet-50 98.77 Ref. [64] MobileNet 88.4 Ref Ref.. [6 [65 4] ] Mob VGG16 ileNet 93.5 88.4 Ref. [66] CNN 98.2 Ref. [65] VGG16 93.5 Ref. [67] ResNet50 + SeNet 96.81 Ref. [66] CNN 98.2 Proposed Ref. [67] ResNet50 + SeNet 96.81 Propose system 99.64 DenseNet121 model Proposed Propose system 99.64 DenseNet121 model Agronomy 2023, 13, 1184 20 of 24 Agronomy 2023, 13, x FOR PEER REVIEW 22 of 26 Accuracy Inception-v3 ResNet-50 MobileNet VGG16 CNN ResNet50 + Proposed SeNet system Accuracy Figure 14. Comparison of results of DenseNet121 against diﬀerent CNN models in terms of accu- Figure 14. Comparison of results of DenseNet121 against different CNN models in terms of accuracy. racy. 6. Conclusions 6. Conclusions The purpose of this research was to analyze and apply many different deep-learning The purpose of this research was to analyze and apply many diﬀerent deep-learning models to the task of diagnosing diseases that affect tomato plants. The dataset from models to the task of diagnosing diseases that aﬀect tomato plants. The dataset from Kaggle, which is readily accessible, was selected since it contains information on ten Kaggle, which is readily accessible, was selected since it contains information on ten dif- different diseases that might affect tomato plant leaves. Three novel CNN architectures ferent diseases that might aﬀect tomato plant leaves. Three novel CNN architectures were were proposed for disease prediction and classiﬁcation in tomato plants using this dataset, proposed for disease prediction and classiﬁcation in tomato plants using this dataset, alongside the specialized deep learning architectures DenseNet121, ResNet50V2, and alongside the specialized deep learning architectures DenseNet121, ResNet50V2, and ViT ViT models. models. Deep learning has emerged as a powerful tool for detecting diseases in plants, in- cluding Deep le tomatoes. arning Using has emerge neurald as networks a pow to eranal ful tyze ool fo larr d gee datasets tecting di ofsea plant ses in images, plants deep , in- learning models can learn to identify patterns and features that are characteristic of spe- cluding tomatoes. Using neural networks to analyze large datasets of plant images, deep ciﬁc learning diseases. models This cancan learhelp n to ident farme ify rs pa andtt erns resear and chers features quickly thatand are cha accurately racteristdiagnose ic of spe- plant ciﬁc di diseases, seases. This allowing can help for farm mor er es tar angeted d resea tr rchers eatment quickl and y and prevent accu ion rate strategies. ly diagnose Based plant on the results of this study, it can be concluded that deep learning models, including diseases, allowing for more targeted treatment and prevention strategies. Based on the DenseNet121, ResNet50V2, and ViT, are effective for tomato disease detection. Among results of this study, it can be concluded that deep learning models, including Dense- these models, DenseNet121 showed the highest accuracy, with 99.88% on training data Net121, ResNet50V2, and ViT, are eﬀective for tomato disease detection. Among these and 99% on testing data. The ResNet50V2 and ViT models also showed high accuracy, models, DenseNet121 showed the highest accuracy, with 99.88% on training data and 99% but their performance was lower than DenseNet121. These ﬁndings suggest that deep on testing data. The ResNet50V2 and ViT models also showed high accuracy, but their learning models can provide accurate and efﬁcient solutions for tomato disease detection, performance was lower than DenseNet121. These ﬁndings suggest that deep learning which can ultimately beneﬁt the agriculture industry. The obtained results of the analysis models can provide accurate and eﬃcient solutions for tomato disease detection, which showed that the suggested model performed better than alternative models. The method can ultimately beneﬁt the agriculture industry. The obtained results of the analysis that has been presented for recognizing the diseases that affects tomatoes is an innovative showed that the suggested model performed bett er than alternative models. The method one. One of the limitations of this research is that the system has not been incorporated that has been presented for recognizing the diseases that aﬀects tomatoes is an innovative into a mobile application. However, it does provide a simple and low-cost method for one. One of the limitations of this research is that the system has not been incorporated diagnosing tomato leaf diseases by only requiring the user to take an image of the affected into a mobile application. However, it does provide a simple and low-cost method for plant’s leaf. In the future, we want to enhance the model by using more advanced forms of diagnosing tomato leaf diseases by only requiring the user to take an image of the aﬀected artiﬁcial intelligence that are supported by internet of things (IoT) technology. plant’s leaf. In the future, we want to enhance the model by using more advanced forms of artiﬁcial intelligence that are supported by internet of things (IoT) technology. Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/agronomy13051184/s1, Figure S1: Performance of DenseNet121 Supplementary Materials: The following supporting information can be downloaded at: model; (a) DenseNet121 model accuracy, (b) DenseNet121 mode loss with (80% training and 20% testing; www.mdpi.com/xxx/s1, Figure S1: Performance of DenseNet121 model; (a) DenseNet121 model ac- Figure S2: Confusion matrix of DenseNet121 with (80% training and 20 testing); Figure S3: Performance Agronomy 2023, 13, 1184 21 of 24 of the ResNet50V2 model; (a) ResNet50V2model accuracy and (b) ResNet50V21 mode loss with (80% training and 20% testing); Figure S4: The ResNet50V2 model’s confusion matrix with (80% training and 20% testing); Figure S5: VIT model performance; (a) VIT model accuracy and (b) VIT mode loss with (80% training and 20% testing); Figure S6: Confusion matrix of ViT model with (80% training and 20% testing); Table S1. Results of DenseNet121 model in (80% training and 20% testing); Table S2: Results of ResNet50V2 model with (80% training and 20% testing); Table S3: Results of ViT model with (80% training and 20% testing). Author Contributions: Conceptualization, M.S.A. and F.W.A.; methodology, M.S.A. and F.W.A.; software, M.S.A. and F.W.A.; validation, M.S.A. and F.W.A. formal analysis, M.S.A. and F.W.A. investigation, M.S.A. and F.W.A. resources, M.S.A. and F.W.A. data curation, M.S.A. and F.W.A. writing—original draft preparation, M.S.A. and F.W.A. writing—review and editing, M.S.A. and F.W.A.; visualization, M.S.A. and F.W.A. supervision, M.S.A. and F.W.A.; project administration, M.S.A. and F.W.A. funding acquisition, M.S.A. and F.W.A. All authors have read and agreed to the published version of the manuscript. Funding: The authors extend their appreciation to the Deputyship for Research & Innovation, Ministry of Education in Saudi Arabia for funding this research work through the project num- ber INST048. Data Availability Statement: https://www.kaggle.com/datasets/kaustubhb999/tomatoleaf (accessed on 20 January 2023). Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Schreinemachers, P.; Simmons, E.B.; Wopereis, M.C. Tapping the economic and nutritional power of vegetables. Glob. Food Secur. 2018, 16, 36–45. [CrossRef] 2. Stilwell, M. The Global Tomato Online News Processing in 2018. Available online: https://www.tomatonews.com/ (accessed on 15 February 2023). 3. Wang, R.; Lammers, M.; Tikunov, Y.; Bovy, A.G.; Angenent, G.C.; de Maagd, R.A. The rin, nor and Cnr spontaneous mutations inhibit tomato fruit ripening in additive and epistatic manners. Plant Sci. 2020, 294, 110436–110447. [CrossRef] [PubMed] 4. Ahmad, A.; Saraswat, D.; El Gamal, A. A survey on using deep learning techniques for plant disease diagnosis and recommenda- tions for development of appropriate tools. Smart Agric. Technol. 2023, 3, 100083. [CrossRef] 5. Yu, Y.; Samali, B.; Rashidi, M.; Mohammadi, M.; Nguyen, T.N.; Zhang, G. Vision-based concrete crack detection using a hybrid framework considering noise effect. J. Build. Eng. 2022, 61, 105246. [CrossRef] 6. Sahin, ¸ M.E.; Ulutas, ¸ H.; Yuce, E. A deep learning approach for detecting pneumonia in chest X-rays. Avrupa Bilim Teknol. Derg. 2021, 28, 562–567. [CrossRef] 7. Bock, C.H.; Parker, P.E.; Cook, A.Z.; Gottwald, T.R. Visual Rating and the Use of Image Analysis for Assessing Different Symptoms of Citrus Canker on Grapefruit Leaves. Plant Dis. 2008, 92, 530–541. [CrossRef] 8. Yu, Y.; Liang, S.; Samali, B.; Nguyen, T.N.; Zhai, C.; Li, J.; Xie, X. Torsional capacity evaluation of RC beams using an improved bird swarm algorithm optimised 2D convolutional neural network. Eng. Struct. 2022, 273, 115066. [CrossRef] 9. Spantideas, S.T.; Giannopoulos, A.E.; Kapsalis, N.C.; Capsalis, C.N. A deep learning method for modeling the magnetic signature of spacecraft equipment using multiple magnetic dipoles. IEEE Magn. Lett. 2021, 12, 1–5. [CrossRef] 10. Park, H.; Eun, J.S.; Kim, S.H. Image-based disease diagnosing and predicting of the crops through the deep learning mechanism. In Proceedings of the 2017 International Conference on Information and Communication Technology Convergence (ICTC), Jeju Island, Republic of Korea, 18–20 October 2017; pp. 129–131. 11. Sardogan, M.; Tuncer, A.; Ozen, Y. Plant leaf disease detection and classiﬁcation based on CNN with LVQ algorithm. In Proceedings of the 2018 3rd International Conference on Computer Science and Engineering (UBMK), Sarajevo, Bosnia and Herzegovina, 20–23 September 2018; pp. 382–385. 12. Singh, V.; Misra, A.K. Detection of plant leaf diseases using image segmentation and soft computing techniques. Inf. Process. Agric. 2017, 4, 41–49. [CrossRef] 13. Zhang, S.; Wu, X.; You, Z.; Zhang, L. Leaf image based cucumber disease recognition using sparse representation classiﬁcation. Comput. Electron. Agric. 2017, 134, 135–141. [CrossRef] 14. Ferentinos, K.P. Deep learning models for plant disease detection and diagnosis. Comput. Electron. Agric. 2018, 145, 311–318. [CrossRef] 15. Devaraj, A.; Rathan, K.; Jaahnavi, S.; Indira, K. Identiﬁcation of Plant Disease using Image Processing Technique. In Proceedings of the 2019 International Conference on Communication and Signal Processing (ICCSP), Chennai, India, 4–6 April 2019; pp. 749–753. Agronomy 2023, 13, 1184 22 of 24 16. Mugithe, P.K.; Mudunuri, R.V.; Rajasekar, B.; Karthikeyan, S. Image Processing Technique for Automatic Detection of Plant Diseases and Alerting System in Agricultural Farms. In Proceedings of the 2020 International Conference on Communication and Signal Processing (ICCSP), Chennai, India, 28–30 July 2020; pp. 1603–1607. 17. Phadikar, S.; Sil, J. Rice disease identiﬁcation using pattern recognition techniques. In Proceedings of the 2008 11th International Conference on Computer and Information Technology, Khulna, Bangladesh, 24–27 December 2008; pp. 420–423. 18. Sarayloo, Z.; Asemani, D. Designing a classiﬁer for automatic detection of fungal diseases in wheat plant: By pattern recognition techniques. In Proceedings of the 2015 23rd Iranian Conference on Electrical Engineering, Tehran, Iran, 10–14 May 2015; pp. 1193–1197. 19. Thangadurai, K.; Padmavathi, K. Computer Visionimage Enhancement for Plant Leaves Disease Detection. In Proceedings of the 2014 World Congress on Computing and Communication Technologies, Trichirappalli, India, 27 February–1 March 2014; pp. 173–175. 20. Yong, Z.; Tonghui, R.; Changming, L.; Chao, W.; Jiya, T. Research on Recognition Method of Common Corn Diseases Based on Computer Vision. In Proceedings of the 2019 11th International Conference on Intelligent Human-Machine Systems and Cybernetics (IHMSC), Hangzhou, China, 24–25 August 2019; Volume 1, pp. 328–331. 21. Khirade, S.D.; Patil, A.B. Plant Disease Detection Using Image Processing. In Proceedings of the 2015 International Conference on Computing Communication Control and Automation, Pune, India, 26–27 February 2015; pp. 768–771. 22. Kamilaris, A.; Prenafeta-Boldú, F.X. Deep learning in agriculture: A survey. Comput. Electron. Agric. 2018, 147, 70–90. [CrossRef] 23. Li, L.; Zhang, S.; Wang, B. Plant Disease Detection and Classiﬁcation by Deep Learning—A Review. IEEE Access 2021, 9, 56683–56698. [CrossRef] 24. Lee, S.H.; Chan, C.S.; Wilkin, P.; Remagnino, P. Deep-plant: Plant identiﬁcation with convolutional neural networks. In Proceedings of the 2015 IEEE International Conference on Image Processing (ICIP), Quebec City, QC, Canada, 27–30 September 2015; pp. 452–456. 25. Zhang, Y.; Song, C.; Zhang, D. Deep Learning-Based Object Detection Improvement for Tomato Disease. IEEE Access 2020, 8, 56607–56614. [CrossRef] 26. Sabrol, H.; Satish, K. Tomato plant disease classification in digital images using classification tree. In Proceedings of the International Conference on Communication and Signal Processing (ICCSP), Melmaruvathur, India, 6–8 April 2016; pp. 1242–1246. 27. Hasan, M.; Tanawala, B.; Patel, K.J. Deep learning precision farming: Tomato leaf disease detection by transfer learning. In Proceedings of the 2nd International Conference on Advanced Computing and Software Engineering (ICACSE), Sultanpur, Inida, 8–9 February 2019. 28. Adhikari, S.; Shrestha, B.; Baiju, B.; Kumar, S. Tomato plant diseases detection system using image processing. In Proceedings of the 1st KEC Conference on Engineering and Technology, Laliitpur, Nepal, 27 September 2018; Volume 1, pp. 81–86. 29. Salih, T.A. Deep Learning Convolution Neural Network to Detect and Classify Tomato Plant Leaf Diseases. Open Access Libr. J. 2020, 7, 12. [CrossRef] 30. Ishak, S.; Rahiman, M.H.; Kanaﬁah, S.N.; Saad, H. Leaf disease classiﬁcation using artiﬁcial neural network. J. Teknol. 2015, 77, 109–114. [CrossRef] 31. Sabrol, H.; Kumar, S. Fuzzy and neural network-based tomato plant disease classiﬁcation using natural outdoor images. Indian J. Sci. Technol. 2016, 9, 1–8. [CrossRef] 32. Rangarajan, A.K.; Purushothaman, R.; Ramesh, A. Tomato crop disease classiﬁcation using pre-trained deep learning algorithm. Procedia Comput. Sci. 2018, 133, 1040–1047. [CrossRef] 33. Coulibaly, S.; Kamsu-Foguem, B.; Kamissoko, D.; Traore, D. Deep neural networks with transfer learning in millet crop images. Comput. Ind. 2019, 108, 115–120. [CrossRef] 34. Sangeetha, R.; Rani, M. Tomato Leaf Disease Prediction Using Transfer Learning. In Proceedings of the International Advanced Computing Conference 2020, Panaji, India, 5–6 December 2020. 35. Mortazi, A.; Bagci, U. Automatically designing CNN architectures for medical image segmentation. In Proceedings of the International Workshop on Machine Learning in Medical Imaging, Granada, Spain, 16 September 2018; pp. 98–106. 36. Jiang, D.; Li, F.; Yang, Y.; Yu, S. A tomato leaf diseases classiﬁcation method based on deep learning. In Proceedings of the Chinese Control and Decision Conference (CCDC), Hefei, China, 22–24 August 2020; pp. 1446–1450. 37. Rashid, J.; Khan, I.; Ali, G.; Almotiri, S.H.; AlGhamdi, M.A.; Masood, K. Multi-Level Deep Learning Model for Potato Leaf Disease Recognition. Electronics 2021, 10, 2064. [CrossRef] 38. PlantVillege. Available online: https://www.kaggle.com/emmarex/plantdisease (accessed on 3 February 2023). 39. Lakshmanarao, A.; Babu, M.R.; Kiran, T.S.R. Plant Disease Prediction and classiﬁcation using Deep Learning ConvNets. In Proceedings of the 2021 International Conference on Artiﬁcial Intelligence and Machine Vision (AIMV), Gandhinagar, India, 24–26 September 2021; pp. 1–6. 40. Militante, S.V.; Gerardo, B.D.; Dionisio, N.V. Plant Leaf Detection and Disease Recognition using Deep Learning. In Proceedings of the 2019 IEEE Eurasia Conference on IOT, Communication and Engineering (ECICE), Yunlin, Taiwan, 3–6 October 2019; pp. 579–582. Agronomy 2023, 13, 1184 23 of 24 41. Marzougui, F.; Elleuch, M.; Kherallah, M. A Deep CNN Approach for Plant Disease Detection. In Proceedings of the 2020 21st International Arab Conference on Information Technology (ACIT), Giza, Egypt, 28–30 November 2020; pp. 1–6. 42. Ngugi, L.C.; Abdelwahab, M.; Abo-Zahhad, M. Tomato leaf segmentation algorithms for mobile phone applications using deep learning. Comput. Electron. Agric. 2020, 178, 105788. [CrossRef] 43. Rozaqi, A.J.; Sunyoto, A. Identiﬁcation of Disease in Potato Leaves Using Convolutional Neural Network (CNN) Algorithm. In Proceedings of the 2020 3rd International Conference on Information and Communications Technology (ICOIACT), Yogyakarta, Indonesia, 24–25 November 2020; pp. 72–76. 44. Mattihalli, C.; Gedefaye, E.; Endalamaw, F.; Necho, A. Real Time Automation of Agriculture Land, by automatically Detecting Plant Leaf Diseases and Auto Medicine. In Proceedings of the 2018 32nd International Conference on Advanced Information Networking and Applications Workshops (WAINA), Krakow, Poland, 16–18 May 2018; pp. 325–330. 45. Divyashri., P.; Pinto, L.A.; Mary, L.; Manasa., P.; Dass, S. The Real-Time Mobile Application for Identiﬁcation of Diseases in Coffee Leaves using the CNN Model. In Proceedings of the 2021 Second International Conference on Electronics and Sustainable Communication Systems (ICESC), Coimbatore, India, 4–6 August 2021; pp. 1694–1700. 46. Liu, J.; Wang, X. Early recognition of tomato gray leaf spot disease based on MobileNetv2-YOLOv3 model. Plant Methods 2020, 16, 83. [CrossRef] [PubMed] 47. Khasawneh, N.; Faouri, E.; Fraiwan, M. Automatic Detection of Tomato Diseases Using Deep Transfer Learning. Appl. Sci. 2022, 12, 8467. [CrossRef] 48. Mim, T.T.; Sheikh, M.H.; Shampa, R.A.; Reza, M.S.; Islam, M.S. Leaves Diseases Detection of Tomato Using Image Processing. In Proceedings of the 2019 8th International Conference System Modeling and Advancement in Research Trends (SMART), Moradabad, India, 22–23 November 2019; pp. 244–249. 49. Kumar, A.; Vani, M. Image Based Tomato Leaf Disease Detection. In Proceedings of the 2019 10th International Conference on Computing, Communication and Networking Technologies (ICCCNT), Kanpur, India, 6–8 July 2019; pp. 1–6. 50. Tm, P.; Pranathi, A.; SaiAshritha, K.; Chittaragi, N.B.; Koolagudi, S.G. Tomato Leaf Disease Detection Using Convolutional Neural Networks. In Proceedings of the 2018 Eleventh International Conference on Contemporary Computing (IC3), Noida, India, 2–4 August 2018; pp. 1–5. 51. Kang, F.; Li, J.; Wang, C.; Wang, F. A Lightweight Neural Network-Based Method for Identifying Early-Blight and Late-Blight Leaves of Potato. Appl. Sci. 2023, 13, 1487. [CrossRef] 52. Al-Gaashani, M.S.A.M.; Shang, F.; Muthanna, M.S.A.; Khayyat, M.; El-Latif, A.A.A. Tomato leaf disease classiﬁcation by exploiting transfer learning and feature concatenation. IET Image Process. 2022, 16, 913–925. [CrossRef] 53. Pathan, S.M.K.; Ali, M.F. Implementation of Faster R-CNN in Paddy Plant Disease Recognition System. In Proceedings of the 2019 3rd International Conference on Electrical, Computer & Telecommunication Engineering (ICECTE), Rajshahi, Bangladesh, 26–28 December 2019; pp. 189–192. 54. Zhou, G.; Zhang, W.; Chen, A.; He, M.; Ma, X. Rapid Detection of Rice Disease Based on FCM-KM and Faster R-CNN Fusion. IEEE Access 2019, 7, 143190–143206. [CrossRef] 55. Cardellicchio, A.; Solimani, F.; Dimauro, G.; Petrozza, A.; Summerer, S.; Cellini, F.; Renò, V. Detection of tomato plant phenotyping traits using YOLOv5-based single stage detectors. Comput. Electron. Agric. 2023, 207, 107757. [CrossRef] 56. Liu, G.; Nouaze, J.C.; Touko Mbouembe, P.L.; Kim, J.H. YOLO-Tomato: A Robust Algorithm for Tomato Detection Based on YOLOv3. Sensors 2020, 20, 2145. [CrossRef] 57. Liu, J.; Wang, X. Tomato Diseases and Pests Detection Based on Improved Yolo V3 Convolutional Neural Network. Front. Plant Sci. 2020, 11, 898. [CrossRef] 58. Aldhyani, T.H.H.; Nair, R.; Alzain, E.; Alkahtani, H.; Koundal, D. Deep Learning Model for the Detection of Real Time Breast Cancer Images Using Improved Dilation-Based Method. Diagnostics 2022, 12, 2505. [CrossRef] 59. Aldhyani, T.H.; Alshebami, A.S.A.; Alzahrani, M.Y. Soft Computing Model to Predict Chronic Diseases. J. Inf. Sci. Eng. 2020, 36, 365–376. 60. Al-Adhaileh, M.H.; Aldhyani, T.H.H. Artiﬁcial intelligence framework for modeling and predicting crop yield to enhance food security in Saudi Arabia. PeerJ Comput. Sci. 2022, 2022, e1104. [CrossRef] [PubMed] 61. Al-Adhaileh, M.H.; Verma, A.; Aldhyani, T.H.H.; Koundal, D. Potato Blight Detection Using Fine-Tuned CNN Architecture. Mathematics 2023, 11, 1516. [CrossRef] 62. Widiyanto, S.; Fitrianto, R.; Wardani, D.T. Implementation of Convolutional Neural Network Method for Classiﬁcation of Diseases in Tomato Leaves. In Proceedings of the 2019 Fourth International Conference on Informatics and Computing (ICIC), Semarang, Indonesia, 16–17 October 2019; pp. 1–5. 63. Mamun, M.A.A.; Karim, D.Z.; Pinku, S.N.; Bushra, T.A. TLNet: A Deep CNN model for Prediction of tomato Leaf Diseases. In Proceedings of the 2020 23rd International Conference on Computer and Information Technology (ICCIT), Dhaka, Bangladesh, 19–21 December 2020; pp. 1–6. 64. Elhassouny, A.; Smarandache, F. Smart mobile application to recognize tomato leaf diseases using Convolutional Neural Networks. In Proceedings of the 2019 International Conference of Computer Science and Renewable Energies (ICCSRE), Agadir, Morocco, 22–24 July 2019; pp. 1–4. 65. Agarwal, M.; Gupta, S.K.; Biswas, K.K. Development of Efﬁcient CNN model for Tomato crop disease identiﬁcation. Sustain. Comput. Inform. Syst. 2020, 28, 100407. [CrossRef] Agronomy 2023, 13, 1184 24 of 24 66. Lamba, M.; Gigras, Y.; Dhull, A. Classiﬁcation of plant diseases using machine and deep learning. Open Comput. Sci. 2021, 11, 491–508. [CrossRef] 67. Zhao, S.; Peng, Y.; Liu, J.; Wu, S. Tomato leaf disease diagnosis based on improved convolution neural network by attention module. Agriculture 2021, 11, 651. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Agronomy Multidisciplinary Digital Publishing Institute http://www.deepdyve.com/lp/multidisciplinary-digital-publishing-institute/transform-and-deep-learning-algorithms-for-the-early-detection-and-ZKUnLA6P9Y

Loading next page...

References (62)

M. Sahin, H. Ulutaş, E. Yüce (2021)
A deep learning approach for detecting pneumonia in chest X-rays
European Journal of Science and Technology
H. Sabrol, Satish Kumar (2016)
Fuzzy and Neural Network based Tomato Plant Disease Classification using Natural Outdoor Images
Indian journal of science and technology, 9
10.1109/ACIT50332.2020.9300072
10.1109/AIMV53313.2021.9670918
10.1109/SMART46866.2019.9117437
Solemane Coulibaly, B. Kamsu-Foguem, Dantouma Kamissoko, Daouda Traoré (2019)
Deep neural networks with transfer learning in millet crop images
Comput. Ind., 108
Yang Yu, B. Samali, M. Rashidi, Masoud Mohammadi, T. Nguyen, Guang Zhang (2022)
Vision-based concrete crack detection using a hybrid framework considering noise effect
Journal of Building Engineering
Yang Zhang, Chenglong Song, Dongwen Zhang (2020)
Deep Learning-Based Object Detection Improvement for Tomato Disease
IEEE Access, 8
N. Khasawneh, Esraa Faouri, M. Fraiwan (2022)
Automatic Detection of Tomato Diseases Using Deep Transfer Learning
Applied Sciences
T. Salih, A. Ali, M. Ahmed (2020)
Deep Learning Convolution Neural Network to Detect and Classify Tomato Plant Leaf Diseases
OALib
10.1109/IranianCEE.2015.7146396
10.1007/978-3-030-00919-9_12
V. Singh, A. Misra (2017)
Detection of plant leaf diseases using image segmentation and soft computing techniques
Information Processing in Agriculture, 4
10.1109/UBMK.2018.8566635
M. Agarwal, S. Gupta, K. Biswas (2020)
Development of Efficient CNN model for Tomato crop disease identification
Sustain. Comput. Informatics Syst., 28
10.1109/WAINA.2018.00106
A. Ahmad, D. Saraswat, A. Gamal (2022)
A Survey on Using Deep Learning Techniques for Plant Disease Diagnosis and Recommendations for Development of Appropriate Tools
Smart Agricultural Technology
10.1109/ECICE47484.2019.8942686
S. Spantideas, A. Giannopoulos, N. Kapsalis, C. Capsalis (2021)
A Deep Learning Method for Modeling the Magnetic Signature of Spacecraft Equipment Using Multiple Magnetic Dipoles
IEEE Magnetics Letters, 12
Monika Lamba, Yogita Gigras, A. Dhull (2021)
Classification of plant diseases using machine and deep learning
Open Computer Science, 11
Jun Liu, Xuewei Wang (2020)
Early recognition of tomato gray leaf spot disease based on MobileNetv2-YOLOv3 model
Plant Methods, 16
Clive Bock, P. Parker, A. Cook, Tim Gottwald (2008)
Visual Rating and the Use of Image Analysis for Assessing Different Symptoms of Citrus Canker on Grapefruit Leaves.
Plant disease, 92 4
M. Al-Adhaileh, Amit Verma, Theyazn Aldhyani, D. Koundal (2023)
Potato Blight Detection Using Fine-Tuned CNN Architecture
Mathematics
A. Kamilaris, F. Prenafeta-Boldú (2018)
Deep learning in agriculture: A survey
ArXiv, abs/1807.11809
Shanwen Zhang, Xiaowei Wu, Zhu-Hong You, Liqing Zhang (2017)
Leaf image based cucumber disease recognition using sparse representation classification
Comput. Electron. Agric., 134
Lawrence Ngugi, M. Abdelwahab, M. Abo-Zahhad (2020)
Tomato leaf segmentation algorithms for mobile phone applications using deep learning
Comput. Electron. Agric., 178
10.1109/ICCUBEA.2015.153
10.1109/ICIP.2015.7350839
Guoxiong Zhou, Wenzhuo Zhang, Aibin Chen, Mingfang He, Xue Ma (2019)
Rapid Detection of Rice Disease Based on FCM-KM and Faster R-CNN Fusion
IEEE Access, 7
M. Al-Adhaileh, Theyazn Aldhyani (2022)
Artificial intelligence framework for modeling and predicting crop yield to enhance food security in Saudi Arabia
PeerJ Computer Science, 8
K. Ferentinos (2018)
Deep learning models for plant disease detection and diagnosis
Comput. Electron. Agric., 145
10.1109/ICCITECHN.2008.4803079
Mosin Hasan, B. Tanawala, Krina Patel (2019)
Deep Learning Precision Farming: Tomato Leaf Disease Detection by Transfer Learning
Computational Materials Science eJournal
10.1109/ICCSRE.2019.8807737
Yang Yu, S. Liang, B. Samali, T. Nguyen, Chenxi Zhai, Jianchun Li, Xingyang Xie (2022)
Torsional capacity evaluation of RC beams using an improved bird swarm algorithm optimised 2D convolutional neural network
Engineering Structures
10.1109/ICCCNT45670.2019.8944692
10.1109/ICCIT51783.2020.9392664
Lili Li, Shujuan Zhang, Bin Wang (2021)
Plant Disease Detection and Classification by Deep Learning—A Review
IEEE Access, 9
10.1109/ICTC.2017.8190957
10.1109/CCDC49329.2020.9164457
Shengyi Zhao, Yun Peng, Jizhan Liu, Shuo Wu (2021)
Tomato Leaf Disease Diagnosis Based on Improved Convolution Neural Network by Attention Module
Agriculture
Pepijn Schreinemachersa, Emmy Simmonsb, Marco Wopereisc (2017)
Tapping the economic and nutritional power of vegetables
Global Food Security, 16
Theyazn Aldhyani, R. Nair, Elham Alzain, Hasan Alkahtani, D. Koundal (2022)
Deep Learning Model for the Detection of Real Time Breast Cancer Images Using Improved Dilation-Based Method
Diagnostics, 12
10.1109/ICIC47613.2019.8985909
A. Cardellicchio, Firozeh Solimani, G. Dimauro, A. Petrozza, S. Summerer, F. Cellini, V. Renó (2023)
Detection of tomato plant phenotyping traits using YOLOv5-based single stage detectors
Comput. Electron. Agric., 207
10.1109/ICOIACT50329.2020.9332037
Theyazn Aldhyani, Ali ِAlshebami, M. Alzahrani (2020)
Soft Computing Model to Predict Chronic Diseases
J. Inf. Sci. Eng., 36
Jun Liu, Xuewei Wang (2020)
Tomato Diseases and Pests Detection Based on Improved Yolo V3 Convolutional Neural Network
Frontiers in Plant Science, 11
Feilong Kang, Jia Li, Chunguang Wang, Fuxiang Wang (2023)
A Lightweight Neural Network-Based Method for Identifying Early-Blight and Late-Blight Leaves of Potato
Applied Sciences
S. Ishak, M. Rahiman, S. Kanafiah, H. Saad (2015)
LEAF DISEASE CLASSIFICATION USING ARTIFICIAL NEURAL NETWORK
, 77
10.1109/IC3.2018.8530532
Rufang Wang, Michiel Lammers, Y. Tikunov, A. Bovy, G. Angenent, R. Maagd (2020)
The rin, nor and Cnr spontaneous mutations inhibit tomato fruit ripening in additive and epistatic manners.
Plant science : an international journal of experimental plant biology, 294
10.1109/ICCSP48568.2020.9182065
10.1109/ICCSP.2016.7754351
10.1109/WCCCT.2014.39
A. Rangarajan, Raja Purushothaman, Aniirudh Ramesh (2018)
Tomato crop disease classification using pre-trained deep learning algorithm
Procedia Computer Science, 133
Javed Rashid, I. Khan, Ghulam Ali, Sultan Almotiri, Mohammed Alghamdi, Khalid Masood (2021)
Multi-Level Deep Learning Model for Potato Leaf Disease Recognition
Electronics, 10
10.1109/IHMSC.2019.00082
Mehdhar Al-gaashani, Fengjun Shang, Mohammed Muthanna, Mashael Khayyat, A. El-latif (2022)
Tomato leaf disease classification by exploiting transfer learning and feature concatenation
IET Image Process., 16
Yuce (2021)
A deep learning approach for detecting pneumonia in chest X-rays
Avrupa Bilim Teknol. Derg., 28
Guoxu Liu, Joseph Nouaze, Philippe Mbouembe, J. Kim (2020)
YOLO-Tomato: A Robust Algorithm for Tomato Detection Based on YOLOv3
Sensors (Basel, Switzerland), 20
10.1109/ICCSP.2019.8698056

Publisher: Multidisciplinary Digital Publishing Institute
Copyright: © 1996-2023 MDPI (Basel, Switzerland) unless otherwise stated Disclaimer Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. Terms and Conditions Privacy Policy
ISSN: 2073-4395
DOI: 10.3390/agronomy13051184
Publisher site: See Article on Publisher Site

Abstract

agronomy Article Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease Mohammed Saeed Alzahrani and Fawaz Waselallah Alsaade * College of Computer Science and Information Technology, King Faisal University, Al-Ahsa P.O. Box 4000, Saudi Arabia * Correspondence: falsaade@kfu.edu.sa Abstract: Plant diseases pose the greatest threat to food supply integrity, and it is a signiﬁcant challenge to identify plant diseases in their earliest stages to reduce the potential for associated economic damage. Through the use of computer vision, a cutting-edge artiﬁcial intelligence is offered as a solution to this problem so that tomato leaf diseases may be classiﬁed. The proposed network is expected to provide excellent results. Transfer learning is also used to make the model efﬁcient and cost effective. Since tomato diseases may signiﬁcantly inﬂuence crop output and quality, early identiﬁcation and diagnosis of these diseases are essential for successful treatment. Deep learning has shown a great deal of promise in plant disease identiﬁcation, providing excellent accuracy and efﬁciency. In this investigation, we compared the performance of three different deep learning models—DenseNet169, ResNet50V2, and a transform model, namely ViT, with regard to diagnosing diseases affecting tomatoes. Both diseased and healthy tomato samples were included in the dataset of photos of tomato diseases used for training and testing the models. The DenseNet121 model had the best results, with a training accuracy of (99.88%) and a testing accuracy of (99.00%). This gave it the greatest overall accuracy. Both the ResNet50V2 and VIT models attained high levels of accuracy, with testing accuracies of (95.60% and 98.00%), respectively. Our results demonstrate deep learning’s potential for accurate and efﬁcient tomato disease detection, which could aid in early disease management and ultimately improve crop yield and quality. The experimental ﬁndings show that the suggested ensemble models stand out due to the short amount of time required for Citation: Alzahrani, M.S.; Alsaade, training and testing as well as their exceptional classiﬁcation performances. Because of this study, F.W. Transform and Deep Learning professionals will be able to facilitate the early diagnosis of plant diseases in a straightforward and Algorithms for the Early Detection expedient way, thereby preventing the emergence of new infections. and Recognition of Tomato Leaf Disease. Agronomy 2023, 13, 1184. Keywords: artiﬁcial intelligence; deep leaning; transform learning; food security; plant diseases https://doi.org/10.3390/ agronomy13051184 Academic Editors: Louis Kouadio and Luís Manuel Navas Gracia 1. Introduction Received: 7 March 2023 Plants are crucial to our survival because they give us sustenance and shield us from Revised: 14 April 2023 radiation. Without plants, it is difﬁcult to imagine there ever being life on Earth; they Accepted: 18 April 2023 not only supply food for all terrestrial creatures but also shield the ozone layer, which Published: 22 April 2023 blocks harmful ultraviolet light. Tomatoes are globally grown vegetables because they are nutrient rich and safe for human consumption [1]. Around 160,000,000 metric tons of tomatoes are consumed annually across the globe [2]. Many people believe that tomato sales could help rural communities earn much-needed income, which would signiﬁcantly Copyright: © 2023 by the authors. impact poverty levels [3]. Tomatoes have a high concentration of nutrients and are grown Licensee MDPI, Basel, Switzerland. all over the world; as a result, they have a major impact on the agricultural economy when This article is an open access article cultivated and harvested. Tomatoes are beneﬁcial in preventing diseases, such as gingival distributed under the terms and bleeding, hypertension, and hepatitis, due to their pharmacological properties [1], and conditions of the Creative Commons their anti-cancer properties are well documented. Tomatoes are increasingly in demand Attribution (CC BY) license (https:// because of their growing popularity. Statistically, almost 80.00% of agricultural output [2] creativecommons.org/licenses/by/ is attributed to small farmers. However, diseases and pests wipe out approximately half 4.0/). Agronomy 2023, 13, 1184. https://doi.org/10.3390/agronomy13051184 https://www.mdpi.com/journal/agronomy Agronomy 2023, 13, 1184 2 of 24 of these farmers’ harvests every year. Research into ﬁeld crop disease detection is crucial because the diseases and insects that parasitize tomatoes have a profound impact on their growth. According to a 2020 study compiled by the Food and Agricultural Organization Corporate Statistical Database (FAOSTAT), 186,821 million tons of tomatoes were produced globally [1]. Agriculture is the fundamental contributor to economic expansion as well as the foundation upon which human civilization was built. The many different kinds of plant diseases that farmers have to protect their crops against have proven to be a signiﬁcant obstacle for agricultural production. Understanding how to prevent plant diseases and implementing preventative measures are the two most important factors in crop yield optimization. Early detection of plant diseases is essential if one wants to achieve maximum agricultural output while at the same time preserving ﬁnancial resources and reducing the amount of crop loss. The fact that everything is handled by computers also makes it simple to put into action. On the one hand, accurate diagnosis and categorization of diseases within a reasonable amount of time is highly important for preserving both the quality and quantity of tomatoes. Environmental conditions may play a role in the development of a wide range of plant diseases. The disease triangle is a conceptual model that represents the relationship between three crucial factors: the environment, the host, and the infectious agent. The disease triangle was developed to explain how diseases spread. This type was created in the 1950s and has been widely used since its introduction. It follows that disease cannot occur, since it is impossible for it to take place if any one of these three requirements is absent from the triangle. Many abiotic factors, including air movement, temperature, humidity, pH, and watering, could potentially have signiﬁcant effects on plants. Fungus, bacteria, and viruses are some organisms that may attack a plant. The infectious agent is an organism that causes the plant to become diseased. The term “host” refers to a plant that is infected with a disease. Disease development is the consequence of all the risk factors converging simultaneously [3]. In general, infections are deﬁned by symptoms that manifest by working their way up through the plant, and as a result, the majority of diseases have rapid transmission rates once they have infected a plant. Pathogenic fungi, bacteria, and viruses, as well as poor climatic conditions, are some of the many potential causes of plant diseases. Because of the presence of these diseases, a plant’s fundamental functions, such as photosynthesis, pollination, fertilization, and germination, may be disrupted. This emphasizes the critical need to accurately diagnose diseases as early as medical science currently allows. Technology has advanced to the point that it is possible to use tools to diagnose whether a plant is diseased, and if so, what kind of disease it has, rather than relying only on the judgment of a human expert. Results from procedures such as object recognition and classiﬁcation, as well as image processing and artiﬁcial intelligence algorithms, are becoming increasingly high quality as the quality of photos obtained by technological equipment continues to improve. Machine learning (ML) and deep learning (DL) have proved much more effective than traditional methods of optimization and prediction. First, unlike traditional techniques that require humans to extract characteristics and are limited by the amount of data, new systems may learn automatically from large quantities of data. Second, ML and DL models may generalize well to data that have not yet been seen, which is a signiﬁcant improvement on prior methods. In contrast to more traditional methods, machine learning and deep learning models may pick up on nonlinear and intricate con- nections in the data. Hence, ML excels at handling situations with many moving parts, especially those with complex interactions. Nowadays, AI (Artiﬁcial intelligence) is exten- sively used in many ﬁelds, including communication, building, magnetism, physics, and biology [4–8]. It is crucial to accurately identify plant diseases and classify them in a timely manner in this approach [9]. AI has progressed to the point that it can now automatically identify plant diseases from raw images [10,11]. To date, there has been much research on determining the causes of plant diseases. Most studies used preexisting datasets, models, and libraries to conduct their experiments. Agronomy 2023, 13, 1184 3 of 24 To automate the processes of identifying and classifying plant leaf diseases, Singh and Misra [12] devised an algorithm for the picture segmentation technique. Using a genetic algorithm, they achieved an overall success rate of 97.60% across ﬁve different diseases while trying to identify them. Zhang et al. [13] analyzed a database of cucumber leaf samples to identify diseases. In this study, the researchers segmented diseased leaves using k-means clustering to obtain shape and color information that could be utilized for disease diagnosis. When classifying these damaged leaves using the sparse representation method, an 86.00% accuracy rate was attained. Convolutional neural network (CNN) models, a kind of AI, can be used for disease detection and diagnosis in plants [14]. Model training was performed using a dataset that included a total of 87,848 images. The collection also includes 58 unique plant–disease combinations, which are applied to 25 different plant species. A 99.53% success rate was calculated from the data. Image processing [15,16], pattern recognition [17,18], and computer vision [19,20] have all seen rapid development and use in agriculture in recent years, with a particular emphasis on automating disease- and pest-detection processes. These tasks often cause problems due to the complexity and time commitment involved in preprocessing and constructing picture features for traditional computer vision models. The accuracy with which the feature extraction processes and with which the learning algorithm have been built also affects the efﬁcacy of these systems [21–23]. Deep learning is gaining momentum in sickness detection as a result of increased processing power, increased storage capacity, and the availability of large datasets. This method has recently been applied to the detection of plant diseases; a problem that has proven difﬁcult to solve. As a subset of machine learning, deep learning is a term with a speciﬁc meaning. CNNs are among the most popular algorithms used in deep learning for tasks, including picture classiﬁcation, object recognition, and semantic segmentation [24,25]. CNNs are helpful for discovering patterns in images, objects, and sceneries because they learn to classify based on data from the image, so the user is no longer required to laboriously isolate the desired features within the image [26,27]. This article will assess and explain a selection of the many different approaches to deep learning that are currently in use. Although a number of studies have been conducted to investigate the impact of disease detection on tomato crops, the existing model still requires improvement. Consequently, we proposed a CNN model with two convolution layers, two max pooling layers, a disease in tomato plants by analyzing data from the hidden layer, and a ﬂattening layer for tomato plants. Farmers can now solve problems independently without seeking advice from agriculture industry specialists. This entails identifying the many diseases that might affect crops. Our model was developed to assist in early-stage diagnoses of plant diseases. This will increase the overall productivity of agricultural activities, and as a result, food availability. The need for an automated technique for diagnosing diseases that might impact tomato plants is the key driver of this project. The following are some of this study’s contributions that ﬁll some of the gaps in existing research: Evaluating a wide variety of characteristics, including a crop’s production, yield capac- ity, grain quality, and nutrient retention, leads to an accurate plant disease diagnosis. Providing advice on constructing new CNN models and creating new ensemble structures using the proposed CNNs. Using an advanced transform learning lie to detect tomato leaf disease is critical to saving food security. Increasing the accuracy percentage of the proposed model in comparison to earlier studies in the ﬁeld that were reported in academic journals. Increasing categorization reliability and reducing the amount of time required for training and examinations. This paper presents an architecture for early disease identiﬁcation and classiﬁcation in tomato leaves that is based on three different deep learning models—DenseNet169, ResNet50V2, and transform ViT model and data augmentation. To properly forecast the Agronomy 2023, 13, 1184 4 of 24 kind of diseases that will impact the tomato leaves, this study aims to build a trustworthy framework for screening photographs of tomato leaves for indicators of disease. 2. Study Background Researchers from a wide variety of institutions have developed automated diseases detection systems using cutting-edge technologies, such as machine learning and neural network designs like Inception V3 net, VGG 16 net, and Squeeze Net. They use highly accurate procedures to diagnose plant diseases in tomato leaf tissues. In order to detect and categorize tomato diseases, a pretrained network model has an accuracy of 94.00% to 95.00% [28,29]. Using a dataset of 300 images and the Tree Classiﬁcation Model and Segmentation, six types of tomato leaf disease were identiﬁed and classiﬁed [30]. It has been proposed [31] that we can classify leaf-affecting plant diseases with 93.75% accuracy using a speciﬁc approach. Plant leaf diseases may be accurately identiﬁed and categorized using image processing software and a classiﬁcation scheme [32]. A smartphone with an 8-megapixel camera was used to photograph a variety of conditions and then divide the resulting data into two groups: healthy and sick. The image processing procedure was composed of three distinct actions: boosting contrast, image segmentation, and distinctive feature locating. An artiﬁcial neural network with several layers and a feed-forward neural net was used to conduct classiﬁcation tasks, and the results of these networks are compared. Compared to the ﬁndings obtained by the multilayer perceptron (MLP) network and the radial basis function (RBF) network, they are much more favorable. Over the course of the study, the image of the plant blade was dissected into healthy and sick sections. However, this did not allow pinpointing the cause of the problem. In order to diagnose leaf diseases, researchers used color space analysis, color time, a histogram, and color coherence as a classiﬁer, which achieved 87.2% in terms of accuracy. Researchers have used AlexNet and VGG 19 models with a frame size of 13,262 to detect diseases that are hurting tomato harvests. The model was utilized to achieve 97.49% accuracy [33]. To obtain a 95.00% detection rate for viruses that damage dairy crops [34], we used transfer learning and a CNN model. With the goal of identifying and categorizing the conditions of tomato plants’ leaf surfaces, an AlexNet-based deep learning mechanism using neural network-trained transfer learning achieved an accuracy of 95.75% [35,36]. Resnet-50 was created to identify 1000 unique diseases that may harm tomato leaves by presenting 3000 images, each of which was tagged with a disease name, such as “lesion blight”, “late blight”, or “yellow curl leaf”. The ﬁrst convolution layer ’s kernel size was increased to 11 11, and the comparing network activation function was switched to Leaky- ReLU. The model has been reﬁned over many iterations, and its performance in classifying diseases has improved to an accuracy of 98.30% and a precision of 98.00% [37]. Simpliﬁed eight-layer CNN models have been proposed for disease detection and classiﬁcation in tomato leaves [38]. This research used the PlantVillage dataset [39], a compilation of data on various agricultural products. By applying deep learning to the tomato leaf dataset, the author focused on disease diagnoses to improve performance. In recent years, CNNs have emerged as a reliable tool for diagnosing plant diseases [40,41]. Some research [42,43] has concentrated on improving feature detection quality by removing obstacles caused by inconsistencies in illumination and background homogeneity in high-stakes scenarios. Feature detection improvements have been the subject of other investigations that have sought to implement them by detecting complicated contexts. Few authors have developed real-time models to hasten plant disease detection [44,45]. Other authors’ work has also led to the early identiﬁcation of plant diseases via model development [46,47]. In Ref. [48], the authors investigate digital photos of tomato leaves to determine the presence of various diseases. The authors implement a classiﬁcation model based on CNN and AI-derived algorithms that is 96.55% accurate in recognizing ﬁve different diseases. In several studies, deep neural network models have been used to identify diseases in tomato leaves. In Ref. [49], for example, the authors compare four alternative models (LeNet, VGG16, ResNet, and Xception) and conclude that the VGG16 model Agronomy 2023, 13, 1184 5 of 24 achieves the greatest performance (99.25% accuracy) when used to categorize nine distinct disorders. The effectiveness of deep neural network models for diagnosing diseases in tomato leaves has been studied in other research. According to Ref. [50], an identical issue was solved with 95.00% or higher accuracy using the AlexNet, GoogleNet, and LeNet models. Agarwal et al. [51] built CNN architecture and compared it to other ML models (such as random forest and decision trees) and DL models for classifying data into 10 groups (VGG16, Incep-tionv3, and MobileNet). The result was a 99.20% accuracy boost. Many research efforts have focused on improving classiﬁcation accuracy by combining random forests, support vector machines, and multinomial logistic regression, which are just a few examples of the classiﬁcation networks that may be employed with the obtained leaf characteristics [52]. With the help of MobileNetv2 and NASNetMobile, we were able to successfully extract leaf features. Researchers have shown that classiﬁcation accuracy may be greatly improved by combining these two techniques. Many studies have successfully diagnosed plant diseases using algorithms such as Mask R-CNN [53]. Computing costs and model sizes have been decreased via the use of several techniques namely K-nearest neighbors and Gabor ﬁlters and K-nearest neighbors (KNN). Both methods have been used to attempt to reduce the time and resources needed to run deep learning calculations. To reduce computational costs, the authors of Ref. [54] used a SqueezeNet architecture with only 33 ﬁlters. YOLO-Tomato, which is based on YOLOv3 and was utilized by the authors of Refs. [55–57] to improve tomato identiﬁcation, has been presented as a solution for dealing with these issues. YOLOv3 has been designed with a thick architecture in order to make the reuse of features easier and to assist in learning a model that is both more accurate and more compact. 3. Materials and Methods Due to tomato leaves’ complicated designs and the wide variety of diseases that affect tomatoes, the disease identiﬁcation process can be difﬁcult. Deep learning has emerged as a strong technology in recent years as support for the computer-assisted detection of diseases affecting tomatoes. This technique uses deep neural networks, which can learn complicated patterns from extensive volumes of picture data. The recommended approach for detecting tomato diseases via deep learning includes several essential stages. To begin, a tomato leaf picture dataset is gathered. This dataset will comprise examples of various diseased and healthy leaves. After preprocessing, the photos are utilized to teach a deep neural network, such as a CNN, to recognize intricate patterns within images. After the model has been trained, it can categorize newly acquired pictures as belonging to either the healthy class or one of the disease classes. Adjustments to the model’s parameters and the inclusion of new data may both enhance the model’s performance and make it more accurate. In conclusion, model performance may be tested using a test set of pictures to determine its accuracy and generalization capacity. The fundamental architecture of the suggested system for the classiﬁcation and detection of tomato plant leaf diseases is presented in Figure 1. 3.1. Dataset This study used the Tomato Leaf Diseases Dataset, consisting of 11,000 images of tomato leaves affected by 10 distinct diseases. Each class, including tomato mosaic virus, target spot, bacterial spot, tomato yellow leaf curl virus, late blight, leaf mold, early blight, spider mites (two-spotted spider mite), tomato healthy, and Septoria leaf spot, contains 1100 images. The dataset is publicly accessible on Kaggle. Table 1 displays the tomato leaf features, and the sample numbers of each class, and a sample of the tomato leaf diseases is presented in Figure 2. The dataset is available here https://www.kaggle.com/datasets/ kaustubhb999/tomatoleaf (accessed on 20 January 2023). Agronomy 2023, 13, 1184 6 of 24 Agronomy 2023, 13, x FOR PEER REVIEW 6 of 26 Figure Figure 1. 1. The The structure of the structure of the suggested suggested approach. approach. Table 1. Tomato leaf features. 3.1. Dataset This study used the Tomato Leaf Diseases Dataset, consisting of 11,000 images of Class Number of Samples Description tomato leaves aﬀected by 10 distinct diseases. Each class, including tomato mosaic virus, This is a viral disease that can cause signiﬁcant damage to tomato plants, target spot, bacterial spot, tomato yellow leaf curl virus, late blight, leaf mold, early blight, leading to reduced growth and yield. Symptoms include mosaic-like patterns spider mites (two-spott ed spider mite), tomato healthy, and Septoria leaf spot, contains of light and dark green on leaves, stunted growth, and deformed fruit. The Mosaic virus 1100 1100 images. The dataset is publicly accessible on Kaggle. Table 1 displays the tomato leaf disease is transmitted by contact with infected plant material and can be managed using resistant tomato varieties and cultural practices, such features, and the sample numbers of each class, and a sample of the tomato leaf diseases as sanitation. is presented in Figure 2. The dataset is available here htt ps://www.kaggle.com/da- tasets/kaustubhb999/tomatoleaf (accessed on 20 January 2023). This disease is caused by the fungus Corynespora cassiicola and can cause circular, sunken lesions with concentric rings on leaves and stems. It can lead Target spot 1100 Table 1. Tomato leafto features. defoliation and reduced fruit yield, but it can be managed using fungicides and cultural practices, such as crop rotation and sanitation. Class Number of Samples Description This is a common disease that can cause severe damage to tomato plants, This is a viral disease that can cause signiﬁcant damage to tomato leading to reduced yields and poor fruit quality. Symptoms include brown, plants, leading to reduced growth and yield. Symptoms include mo- sunken lesions on leaves and fruit, which eventually turn black and crusty. Bacterial spot 1100 The disease is caused by the bacterium Xanthomonas campestris pv. vesicatoria saic-like patt erns of light and dark green on leaves, stunted growth, and Mosaic virus 1100 and can be controlled using copper-based fungicides and cultural practices, deformed fruit. The disease is transmitt ed by contact with infected such as crop rotation and sanitation. plant material and can be managed using resistant tomato varieties and This is a viral disease that can cause signiﬁcant damage to tomato plants, cultural practices, such as sanitation. leading to stunted growth and reduced fruit yield. Symptoms include leaf This disease is caused by the fungus Corynespora cassiicola and can Yellow leaf curl virus 1100 yellowing and curling as well as distorted fruit. The disease is transmitted by cause circular, sunken lesions with concentric rings on leaves and the whiteﬂy Bemisia tabaci, and it can be managed using insecticides, resistant Target spot 1100 stems. It can lead to defoliation and reduced fruit yield, but it can be tomato varieties, and cultural practices, such as crop rotation and sanitation. managed using fungicides and cultural practices, such as crop rotation This disease is caused by the fungus Phytophthora infestans and can cause rapid and sanitation. and devastating damage to tomato plants. Symptoms include water-soaked Late blight 1100 lesions on leaves and stems, which quickly turn brown and necrotic. The This is a common disease that can cause severe damage to tomato disease can be managed using fungicides and cultural practices, such as crop plants, leading to reduced yields and poor fruit quality. Symptoms in- rotation, sanitation, and removal of infected plant material. clude brown, sunken lesions on leaves and fruit, which eventually turn Bacterial spot 1100 This disease is caused by the fungus Passalora fulva and can cause signiﬁcant black and crusty. The disease is caused by the bacterium Xanthomonas damage to tomato leaves, reducing plant growth and yield. Symptoms include campestris pv. vesicatoria and can be controlled using copper-based fun- Leaf mold 1100 yellowing and brown lesions on the plant’s lower leaves. The disease can be gicides and cultural practices, such as crop rotation and sanitation. managed using fungicides and cultural practices, such as crop rotation and sanitation. Agronomy 2023, 13, 1184 7 of 24 Table 1. Cont. Class Number of Samples Description This disease is caused by the fungus Alternaria solani and is characterized by concentric rings of dark brown spots on the plant’s lower leaves. It can cause Early blight 1100 defoliation and reduced fruit yield but can be managed using fungicides and cultural practices, such as crop rotation and pruning. These are tiny arachnids that can cause signiﬁcant damage to tomato plants by feeding on the undersides of leaves, causing yellowing and stunted growth. Spider mites 1100 They can be managed using predatory mites, insecticidal soaps, and cultural practices, such as crop rotation and sanitation. This class contains images of healthy tomato leaves, which can be used as Tomato healthy 1100 references for comparison with diseased leaves. This disease is caused by the fungus Septoria lycopersici and is characterized by small, circular lesions with dark brown centers and yellow halos on the lower Septoria leaf spot 1100 leaves of the plant. It can cause signiﬁcant defoliation and yield loss, but can Agronomy 2023, 13, x FOR PEER REVIEW be managed using fungicides and cultural practices, such as crop rotation 8 of 26 and sanitation. Figure 2. Demonstrated dataset samples. Figure 2. Demonstrated dataset samples. 3.2. Data Preprocessing 3.2. Data Preprocessing The data preprocessing step is essential because it improves both the data quality and The data preprocessing step is essential because it improves both the data quality and overall performance of the classiﬁcation methods used for the picture categorization process. overall performance of the classiﬁcation methods used for the picture categorization pro- The primary goal of data preprocessing is to prepare images for the deep learning model cess. The primary goal of data preprocessing is to prepare images for the deep learning by removing any discrepancies, noise, or outliers that could negatively impact the model’s model by removing any discrepancies, noise, or outliers that could negatively impact the accuracy. This is accomplished by organizing data in a way that makes it easier for the model’s accuracy. This is accomplished by organizing data in a way that makes it easier model to analyze. In deep learning models for image classiﬁcation, several common data for the model to analyze. In deep learning models for image classiﬁcation, several com- preprocessing approaches are applied, such as image resizing and normalization. Image mon data preprocessing approaches are applied, such as image resizing and normaliza- tion. Image resizing involves adjusting the size of the images to ensure that they are of a standard size for the model’s training. This helps reduce model complexity and ensures that it is trained on images of the same size. Images normalization entails adjusting the brightness and contrast levels of each picture so that they are uniform across all of them. This is done to guarantee that the model is trained on photos that have comparable qual- ities. Applying data preprocessing techniques results in a more consistent and high-qual- ity dataset, which can improve the model’s accuracy. These techniques can be imple- mented using various image-processing libraries, such as PIL and OpenCV. 3.3. Transfer Learning Algorithms Transfer learning approaches allow a model to be trained on one task and then ap- plied to a diﬀerent but related task. This is particularly useful when there is a lack of la- beled data for a speciﬁc task or when the task is similar to one that has already been solved. It can save a lot of time and resources by leveraging the knowledge and features learned from a pretrained model. Additionally, transfer learning can improve model performance by reducing the risk of overﬁtt ing and allowing the model to be bett er generalized to a new dataset. Both the quantity of labeled data available for the new task and the degree of similarity between the pretrained model and the new task should be taken into account Agronomy 2023, 13, 1184 8 of 24 resizing involves adjusting the size of the images to ensure that they are of a standard size for the model’s training. This helps reduce model complexity and ensures that it is trained on images of the same size. Images normalization entails adjusting the brightness and contrast levels of each picture so that they are uniform across all of them. This is done to guarantee that the model is trained on photos that have comparable qualities. Applying data preprocessing techniques results in a more consistent and high-quality dataset, which can improve the model’s accuracy. These techniques can be implemented using various image-processing libraries, such as PIL and OpenCV. 3.3. Transfer Learning Algorithms Transfer learning approaches allow a model to be trained on one task and then applied to a different but related task. This is particularly useful when there is a lack of labeled data for a speciﬁc task or when the task is similar to one that has already been solved. It can save a lot of time and resources by leveraging the knowledge and features learned from a pretrained model. Additionally, transfer learning can improve model performance by reducing the risk of overﬁtting and allowing the model to be better generalized to a new dataset. Both the quantity of labeled data available for the new task and the degree of similarity between the pretrained model and the new task should be taken into account before deciding to use transfer learning. It may be preferable to retrain the model from scratch if the new job is quite different from the one the pretrained model was built for, or if there is a huge quantity of labeled data available. Model performance may be enhanced via transfer learning, and the need for labeled data can be reduced. It is applicable to many different models, including CNNs, and may be implemented using either feature-based or ﬁne-tuning approaches. 3.3.1. DenseNet121 Model DenseNet121 is a popular CNN architecture that Huang et al. introduced in 2017 as part of the DenseNet family. The network is based on the concept of dense connections between layers, which allow information to ﬂow more efﬁciently through the network and improve its accuracy. The DenseNet121 architecture consists of multiple dense blocks, each containing a set of convolutional layers that are densely connected to all subsequent layers [58–61]. In this research, we employ a pretrained DenseNet121 model for disease identiﬁcation in tomatoes. It is common practice to apply transfer learning, and this was done to initialize the pretrained model using weights learned from the huge ImageNet dataset. A new set of layers, consisting of a global average pooling layer, two fully connected layers with 512 and 256 neurons, and two batch normalization layers, was added after the top layer of the pretrained model was eliminated and replaced. Once the second layer was completely linked, the ReLU activation function was applied to a new layer that served as an activation layer. Finally, a softmax activation layer was added to the output layer, which comprised of 10 neurons that corresponded to the 10 different types of tomato diseases. Figure 3 illustrates this point. This architecture was chosen because DenseNet models have shown strong perfor- mance in various image classiﬁcation tasks and have a compact architecture suitable for transfer learning. The global average pooling layer was used to reduce the number of pa- rameters in the model, and batch normalization layers were added to improve the stability and convergence of the training process. The important parameters of DenseNet121 model are presented in Table 2. Agronomy 2023, 13, 1184 9 of 24 Table 2. Describe DenseNet121 model parameters. DenseNet121 Parameters Value Batch_Size 8 Optimizer SGD Learning Rate (lr) 0.0001 Momentum 0.9 Early Stopping (es) Yes Model Checkpoint (mc) Yes Agronomy 2023, 13, x FOR PEER REVIEW 10 of 26 Number of Epochs 100 Figure 3. DenseNet121 model for tomato disease detection. Figure 3. DenseNet121 model for tomato disease detection. 3.3.2. ResNet50V2 Model 3.3.2. ResNet50V2 Model ResNet50V2 is a convolutional neural network architecture that has been used for ResNet50V2 is a convolutional neural network architecture that has been used for computer vision for analysis and classiﬁcation images, including disease classiﬁcation in computer vision for analysis and classiﬁcation images, including disease classiﬁcation in plants. The pretrained model was initialized with weights learned from the large image plants. The pretrained model was initialized with weights learned from the large image pr processi ocessing ng data dataset, set, which which is is a a common pra common practice ctice i in n tra transfer nsfer l learning. earning. The top The top la layer yer of of the the pretrained model was removed and replaced with a new set of layers. The classiﬁer that pretrained model was removed and replaced with a new set of layers. The classiﬁer that we we ad added ded on on top top of the pr of the pretrained etrained ResNet50V2 ResNet50V2 mode modell is de is designed signed t to o ad adapt apt the n the network etwork to to a new im a new image age clclassiﬁcation assiﬁcation task. The p task. The retraine pretrained d model model output is output passed through a is passed through Glob- a alAveragePooling2D layer that reduces feature maps’ spatial dimensions by averaging GlobalAveragePooling2D layer that reduces feature maps’ spatial dimensions by averaging them them alon along g t the he height an height and d width width axe axes. s. This results This results in in a a ﬁxed-length ﬁxed-length fe featur ature e v vector ector tthat hat encodes encodes the the most most important features important features of of the the input im input image. age. The The feature v feature vector ector is proce is processed ssed by two fu by two fully lly connect connected ed lay layers ers w with ith 51 512 2 and 2 and 256 56 n neu- eu- rrons, respectiv ons, respectively ely, wh , which ich learn to map learn to map the the fe featur ature v e vector ector to a hig to a high-level h-level representation representation that captures the discriminative information for the new image classiﬁcation task. Batch that captures the discriminative information for the new image classiﬁcation task. Batch normalization normalization lay layers ers ar are e added added after after e each ach full fully y conn connected ected lay layer er to to stabilize stabilize the the trainin training g process and accelerate convergence. After the second fully connected layer, nonlinearity process and accelerate convergence. After the second fully connected layer, nonlinearity is introduced into the network by adding an activation layer using the ReLU activation is introduced into the network by adding an activation layer using the ReLU activation function. function. This al This allows lows the network to the network to acquir acquire more e more nuanced repre nuanced representations. sentations. Finally, the second fully connected layer ’s output is passed through a fully connected Finally, the second fully connected layer’s output is passed through a fully connected layer with 10 neurons and a softmax activation function, which outputs a probability layer with 10 neurons and a softmax activation function, which outputs a probability dis- distribution over the 10 classes in the dataset. This layer learns to map the high-level tribution over the 10 classes in the dataset. This layer learns to map the high-level repre- representation to the ﬁnal output of the network, which is the predicted class label for the sentation to the ﬁnal output of the network, which is the predicted class label for the input image. The structure of ResNet50V2 model to detect tomato diseases is presented in Fig- ure 4. The parameters of developing the ResNet50V2 model for detecting tomato disease are presented in Table 3. Agronomy 2023, 13, 1184 10 of 24 Agronomy 2023, 13, x FOR PEER REVIEW 11 of 26 input image. The structure of ResNet50V2 model to detect tomato diseases is presented in Figure 4. The parameters of developing the ResNet50V2 model for detecting tomato disease are presented in Table 3. Figure 4. DenseNet121 model for tomato disease detection. Figure 4. DenseNet121 model for tomato disease detection. T Table able 3. 3. Describe Describe ResNet50V2 model parameters. ResNet50V2 model parameters. Parameters Values Parameters Values #Optimizer SGD #Optimizer SGD #Learning Rate (lr ) 0.0001 #Learning Rate (lr) 0.0001 #Momentum 0.9 #Momentum 0.9 #Early Stopping (es) Yes #Early Stopping (es) Yes #Model Checkpoint (mc) Yes #Model Checkpoint (mc) Yes #Number of Epochs 100 #Number of Epochs 100 3.3.3. Vision Transformer 3.3.3. Vision Transformer Vision transformer (VIT) is a new architecture for image classiﬁcation that has gained Vision transformer (VIT) is a new architecture for image classiﬁcation that has gained signiﬁcant attention in the deep learning community. Unlike traditional CNNs, VITs use signiﬁcant att ention in the deep learning community. Unlike traditional CNNs, VITs use transformer-based architecture, which has been highly successful in natural language transformer-based architecture, which has been highly successful in natural language pro- processing tasks. cessing tasks. VIT architecture consists of several key components. First, the input image is divided VIT architecture consists of several key components. First, the input image is divided into a sequence of ﬁxed-sized patches. Each patch is then ﬂattened into a 1D vector and into a sequence of ﬁxed-sized patches. Each patch is then ﬂatt ened into a 1D vector and passed through an embedding layer, which maps the patch to a higher-dimensional feature passed through an embedding layer, which maps the patch to a higher-dimensional fea- space. These patch embeddings are then fed into a transformer encoder, which applies ture space. These patch embeddings are then fed into a transformer encoder, which ap- a series of self-attention mechanisms to learn the contextual relationships between the plies a series of self-att ention mechanisms to learn the contextual relationships between patches, as shown in Figure 5. the patches, as shown in Figure 5. The following steps outline the process we used to train our data and input it into the VIT model for classiﬁcation. The input image is passed through a “patches” layer, which divides the image into a grid of non-overlapping 6 6 patches. This turns the 2D image into a 3D tensor of shape (batch size, patch size patch size number of channels) and the number of color channels (typically 3 for RGB images). The patch tensor is then fed through a “patch encoder” layer, which applies a learned linear transformation (via a dense layer) to each patch and adds a learnable position embedding to each patch. This step helps the model capture spatial relationships between patches and makes it aware of their relative positions within the image. Agronomy 2023, 13, x FOR PEER REVIEW 12 of 26 Agronomy 2023, 13, 1184 11 of 24 Figure 5. Image after transformer encoder. Figure 5. Image after transformer encoder. This layer produces the model’s predictions for each input image. During training, The following steps outline the process we used to train our data and input it into the the model is optimized using the AdamW optimizer, a variant of the Adam optimizer VIT model for classiﬁcation. The input image is passed through a “patches” layer, which that includes weight decay regularization. The loss function used is sparse categorical divides the image into a grid of non-overlapping 6 × 6 patches. This turns the 2D image Agronomy 2023, 13, x FOR PEER REVIEW 13 of 26 cross-entropy, which is a common choice for multiclass classiﬁcation problems. Figure 6 into a 3D tensor of shape (batch size, patch size × patch size × number of channels) and shows the structure of ViT transform model used to detect tomato disease. The signiﬁcant the number of color channels (typically 3 for RGB images). The patch tensor is then fed parameters of ViT model are shown in Table 4. through a “patch encoder” layer, which applies a learned linear transformation (via a dense layer) to each patch and adds a learnable position embedding to each patch. This step helps the model capture spatial relationships between patches and makes it aware of their relative positions within the image. This layer produces the model’s predictions for each input image. During training, the model is optimized using the AdamW optimizer, a variant of the Adam optimizer that includes weight decay regularization. The loss function used is sparse categorical cross- entropy, which is a common choice for multiclass classiﬁcation problems. Figure 6 shows the structure of ViT transform model used to detect tomato disease. The signiﬁcant pa- rameters of ViT model are shown in Table 4. Figure 6. ViT model for tomato disease detection. Figure 6. ViT model for tomato disease detection. Table 4. Describe ViT model parameters. Parameter Value Learning_Rate 0.001 Weight_Decay 0.0001 Batch_Size 128 Number_Epochs 50 Image_Size 72 Patch_Size 6 Number_Patches 144 Projection_Dimension 64 Number_Heads 4 Transformer_Units (128, 64) 4. Experiment The eﬀectiveness of DenseNet121, ResNet50V2, and VIT, three deep learning classi- ﬁcation algorithms, was assessed within the context of this research as they relate to the diagnosis of tomato diseases. The models were trained, tested, and validated using a da- taset that included photographs of both healthy and sick tomatoes. The conﬁguration for training a model was done on a laptop equipped with an 8th generation Core i7 CPU and a GPU 1070 utilized and 8GBs RAM. The TensorFlow and transformer architecture library was used for developing the models. Both the training and testing sets were created from the dataset utilized for this investigation. These two sets were then compared. This divi- sion’s mission was to verify that the deep learning models generated for this research could correctly categorize new photographs. To do this, every category in the dataset was broken up into 1000 photographs that would be used for training, and 100 images that would be used for testing. The primary goal of this division was to supply the deep learn- ing models with a diverse set of photographs from which they could learn, with the end goal of evaluating the models’ capacity to generalize their ﬁndings to images that they had not seen before. Agronomy 2023, 13, 1184 12 of 24 Table 4. Describe ViT model parameters. Parameter Value Learning_Rate 0.001 Weight_Decay 0.0001 Batch_Size 128 Number_Epochs 50 Image_Size 72 Patch_Size 6 Number_Patches 144 Projection_Dimension 64 Number_Heads 4 Transformer_Units (128, 64) 4. Experiment The effectiveness of DenseNet121, ResNet50V2, and VIT, three deep learning clas- siﬁcation algorithms, was assessed within the context of this research as they relate to the diagnosis of tomato diseases. The models were trained, tested, and validated using a dataset that included photographs of both healthy and sick tomatoes. The conﬁguration for training a model was done on a laptop equipped with an 8th generation Core i7 CPU and a GPU 1070 utilized and 8GBs RAM. The TensorFlow and transformer architecture library was used for developing the models. Both the training and testing sets were created from the dataset utilized for this investigation. These two sets were then compared. This division’s mission was to verify that the deep learning models generated for this research could correctly categorize new photographs. To do this, every category in the dataset was broken up into 1000 photographs that would be used for training, and 100 images that would be used for testing. The primary goal of this division was to supply the deep learning models with a diverse set of photographs from which they could learn, with the end goal of evaluating the models’ capacity to generalize their ﬁndings to images that they had not seen before. 4.1. Performance Measurement To measure the success of a deep learning model, its accuracy is often measured against a gold standard. It is a measure of the model’s predictive efﬁcacy that may be calculated by dividing the number of accurate forecasts by the total number of predictions. TP + TN Accuracy = 100% (1) FP + FN + TP + TN It is standard practice to utilize precision as a measure of a deep learning model’s success, particularly when the job at hand is one of categorization. This metric assesses the reliability of the model’s positive predictions by dividing the number of correct forecasts by the total number of correct predictions. It can be collated by Formula (2). TP Precision = 100 % (2) TP + FP Deep learning models make extensive use of the assessment statistic known as recall, particularly when dealing with categorization issues. It provides a numerical representation of the fraction of the dataset’s positive instances that correspond to accurate positive predictions generated by the model. In plainer language, it calculates the number of true positive predictions as a fraction of the total number of positive cases included in the dataset. TP Recall = 100 (3) TP + FN The F1 score is a commonly used evaluation metric in deep learning models, especially for classiﬁcation problems. It is a way to balance the trade-off between precision and recall Agronomy 2023, 13, 1184 13 of 24 by taking the harmonic mean of the two metrics. This makes it a useful metric when dealing with imbalanced datasets or when trying to identify rare classes. precision Sensitivity F1 score = 2 100 (4) precision + Sensitivity where: FP: false positive TP: True positive FN: false negative TN: true negative. 4.2. Results 4.2.1. Results of the DenseNet121 Model Table 5 shows the DenseNet121 model’s high performance in 90% training and 10 testing. The DenseNet121 model was trained using eight batch sizes and the SGD optimizer for 100 epochs. To prevent overﬁtting and minimize training time, the early-stop method was employed, which ended the training after 34 epochs. DenseNet121 scored high in accuracy (99.88%) in the training phase, whereas its testing accuracy was 0.99%. Table S1 shows the result of DenseNet121 with 90% training and 10 testing; it is observed that the accuracy of the DenseNet121model is (98%). Table 5. Results of DenseNet121 model (90% training and 10% testing). Class Name Precision% Recall% F1-Score% Support Bacterial_Spot 96 100 98 100 Early_Blight 99 100 00 100 Late_Blight 100 100 100 100 Leaf_Mold 100 100 100 100 Septoria_Leaf_Spot 97 100 99 100 Spider_Mites_Two-Spotted_Spider_Mite 99 99 99 100 Target_Spot 100 93 96 100 Yellow_Leaf_Curl Virus 100 98 99 100 Mosaic_Virus 99 100 100 100 Healthy 00 100 100 100 Accuracy of DenseNet121 99 1000 DenseNet121_Macro Avg 99 99 99 1000 DenseNet121_Weighted Avg 99 99 99 1000 Figure 7 illustrates how well DenseNet121 model performed throughout the 90% training and 10% validation phases for the purpose of recognizing and categorizing diseases that may affect tomato leaf tissue. A training accuracy beginning at 72.00% and increasing to 99.98% was attained as a consequence of these ﬁndings. According to the graphic representation, the accuracy at the validation phase started at 86.00% and reached 99.00%. The accuracy loss of the DenseNet121 model was 0.2 in validation. The performance of the DenseNet121 model with 80% training and 20% validation phases is presented in Figure S1; it is noted that the mode achieved accuracy 98.45% and validation loss was decreased from 0.8 to 0.1. The proposed Dense-Net121 model and its evaluation yielded a confusion matrix, as shown in Figures 8 and S2. The evaluation parameters were (90%, 80% as training and 10%, 20% as testing). The confusion matrix displays the tTP), TN, FP, and FN values that were gathered for each class. The target_spot class has high environment misclassiﬁcation based on the multiple-graphics processing unit (MGPU) architecture. In terms of leaf disease categorization, the suggested model attained a validation accuracy of 99.64% with 90% training and 10% testing whereas the Dense-Net121 model was attained accuracy 98.45% with 80% training and 20% testing. Agronomy 2023, 13, x FOR PEER REVIEW 15 of 26 Septoria_Leaf_Spot 97 100 99 100 Spider_Mites_Two-Spotted_Spider_Mite 99 99 99 100 Target_Spot 100 93 96 100 Yellow_Leaf_Curl Virus 100 98 99 100 Mosaic_Virus 99 100 100 100 Healthy 00 100 100 100 Accuracy of DenseNet121 99 1000 DenseNet121_Macro Avg 99 99 99 1000 DenseNet121_Weighted Avg 99 99 99 1000 Figure 7 illustrates how well DenseNet121 model performed throughout the 90% training and 10% validation phases for the purpose of recognizing and categorizing dis- eases that may aﬀect tomato leaf tissue. A training accuracy beginning at 72.00% and in- creasing to 99.98% was att ained as a consequence of these ﬁndings. According to the graphic representation, the accuracy at the validation phase started at 86.00% and reached 99.00%. The accuracy loss of the DenseNet121 model was 0.2 in validation. The perfor- mance of the DenseNet121 model with 80% training and 20% validation phases is pre- Agronomy 2023, 13, 1184 14 of 24 sented in Figure S1; it is noted that the mode achieved accuracy 98.45% and validation loss was decreased from 0.8 to 0.1. Agronomy 2023, 13, x FOR PEER REVIEW 16 of 26 Figure 7. Performance of DenseNet121 model; (a) DenseNet121 model accuracy and (b) Dense- Figure 7. Performance of DenseNet121 model; (a) DenseNet121 model accuracy and (b) DenseNet121 Net121 mode loss (90% training and 10% testing). mode loss (90% training and 10% testing). The proposed Dense-Net121 model and its evaluation yielded a confusion matrix, as shown in Figures 8 and S2. The evaluation parameters were (90%, 80% as training and 10%, 20% as testing). The confusion matrix displays the tTP), TN, FP, and FN values that were gathered for each class. The target_spot class has high environment misclassiﬁcation based on the multiple-graphics processing unit (MGPU) architecture. In terms of leaf dis- ease categorization, the suggested model att ained a validation accuracy of 99.64% with 90% training and 10% testing whereas the Dense-Net121 model was att ained accuracy 98.45% with 80% training and 20% testing. Figure Figure 8. 8. Confusion Confusion matrix of matrix of DenseNet121 DenseNet121 ( (90% 90% training a training and nd 10% te 10% testing). sting). 4.2.2. Results of the ResNet50V2 Model 4.2.2. Results of the ResNet50V2 Model The results of the ResNet50V2 model with 90% training and 10% testing are pre- The results of the ResNet50V2 model with 90% training and 10% testing are pre- sented in Table 6. The ResNet50V2 model was trained using eight batch sizes and the SGD sented in Table 6. The ResNet50V2 model was trained using eight batch sizes and the SGD optimizer for 100 epochs. To prevent overﬁtting and minimize training time, the early- optimizer for 100 epochs. To prevent overﬁtt ing and minimize training time, the early- stop method was employed, which ended the training after 20 epochs. The ResNet50V2 stop method was employed, which ended the training after 20 epochs. The ResNet50V2 model achieved accuracies of 99.49% in training and 95.60% in testing. The results of model achieved accuracies of 99.49% in training and 95.60% in testing. The results of Res- Net50V2 with 80% training and 20% testing is shown Table S2; the ResNet50V2 success- fully achieved 94.31%. Table 6. Results of ResNet50V2 model (90% training and 10% testing). Class Name Precision (%) Recall (%) F1-Score (%) Support Bacterial_Spot 99 100 100 100 Early_Blight 97 93 95 100 Late_Blight 88 100 93 100 Leaf_Mold 96 95 95 100 Septoria_Leaf_Spot 96 96 96 100 Spider_Mites_Two-Spotted_Spider_Mite 96 90 93 100 Target_Spot 98 87 92 100 Yellow_Leaf_Curl Virus 100 98 99 100 Mosaic_Virus 89 100 94 100 Healthy 00 97 98 100 Accuracy 96 1000 DenseNet121_Macro Avg 0.96 96 96 1000 DenseNet121_Weighted Avg 0.96 96 96 1000 Agronomy 2023, 13, 1184 15 of 24 ResNet50V2 with 80% training and 20% testing is shown Table S2; the ResNet50V2 success- fully achieved 94.31%. Table 6. Results of ResNet50V2 model (90% training and 10% testing). Class Name Precision (%) Recall (%) F1-Score (%) Support Bacterial_Spot 99 100 100 100 Early_Blight 97 93 95 100 Late_Blight 88 100 93 100 Leaf_Mold 96 95 95 100 Septoria_Leaf_Spot 96 96 96 100 Spider_Mites_Two-Spotted_Spider_Mite 96 90 93 100 Target_Spot 98 87 92 100 Agronomy 2023, 13, x FOR PEER REVIEW 17 of 26 Yellow_Leaf_Curl Virus 100 98 99 100 Mosaic_Virus 89 100 94 100 Healthy 00 97 98 100 Accuracy 96 1000 Figure 9 illustrates the eﬀectiveness of the ResNet50V2 model in identifying tomato DenseNet121_Macro Avg 0.96 96 96 1000 leaf diseases by dividing the dataset into 90% training and 10% testing. ResNet50V2′s val- DenseNet121_Weighted Avg 0.96 96 96 1000 idation accuracy is 96.00%, whereas its training accuracy was initially 78.00% and reached a maximum of 99.49%. At this point in the validation process, the ResNet50V2 model’s accuracy loss is less than 0.1. Figure 9 illustrates the effectiveness of the ResNet50V2 model in identifying tomato The eﬃciency of the ResNet50V2 model in determining the presence of tomato leaf leaf diseases by dividing the dataset into 90% training and 10% testing. ResNet50V2 s diseases with 80% training and 20% testing is seen in Figure S3. The accuracy of Res- validation accuracy is 96.00%, whereas its training accuracy was initially 78.00% and Net50V2 model in validation test is 94.30%. At this stage in the validation process, the reached a maximum of 99.49%. At this point in the validation process, the ResNet50V2 accuracy loss associated with the ResNet50V2 model is less than 0.2. model’s accuracy loss is less than 0.1. Figure 9. Performance of the ResNet50V2 model; (a) ResNet50V2model accuracy and (b) Res- Figure 9. Performance of the ResNet50V2 model; (a) ResNet50V2model accuracy and Net50V21 mode loss (90% training and 10% testing). (b) ResNet50V21 mode loss (90% training and 10% testing). The The Res efﬁciency Net50V of 2 mod the ResNet50V2 el’s confusion mat model rix i in s determining shown in Fig the ures pr10 and S4 b esence of tomato y dividing leaf diseases the dataset with in 80% 90%, tr 8a 0 i% ning as t and rain20% ing and testing 10%, is 2 seen 0% a in s t Figur esting. e S3. Thi The s maccuracy atrix dem of onst ResNet50V2 rates that model the model has in validation a strong posi test is 94.30%. tive imAt pact on toma this stage in to di thesea validation se detecti pr oocess, n and tha the t accuracy it misclaloss ssi- associated ﬁes 42 out of with every the10 ResNet50V2 00 samples. The suggest model is lessed than ResNet50V2 mode 0.2. l can predict the tested The ResNet50V2 model’s confusion matrix is shown in Figures 10 and S4 by dividing classes using the training dataset with 99.49% accuracy, as shown by the results repre- the dataset in 90%, 80% as training and 10%, 20% as testing. This matrix demonstrates that sented in the confusion matrix, 90% as training and 10% testing, whereas the misclassiﬁ- the model has a strong positive impact on tomato disease detection and that it misclassiﬁes cation of ResNet50V2 model with 80% training and 20% testing is 54 simples for all the 42 out of every 1000 samples. The suggested ResNet50V2 model can predict the tested classes. classes using the training dataset with 99.49% accuracy, as shown by the results represented This is the case because the results show that the model was able to predict the classes correctly. Examining the ﬁndings enables us to make this observation for ourselves. The confusion matrix that DenseNet121 produces demonstrates that the model has a strong positive impact on tomato disease detection, with only 10 misclassiﬁed samples out of a total of 1000 samples. Agronomy 2023, 13, 1184 16 of 24 Agronomy 2023, 13, x FOR PEER REVIEW 18 of 26 in the confusion matrix, 90% as training and 10% testing, whereas the misclassiﬁcation of ResNet50V2 model with 80% training and 20% testing is 54 simples for all the classes. Figure 10. The ResNet50V2 model’s confusion matrix (90% training and 10% testing). Figure 10. The ResNet50V2 model’s confusion matrix (90% training and 10% testing). This is the case because the results show that the model was able to predict the classes 4.2.3. VIT Model Results correctly. Examining the ﬁndings enables us to make this observation for ourselves. The The results achieved by the VIT model’s classiﬁcation performance for each class de- confusion matrix that DenseNet121 produces demonstrates that the model has a strong scribed with 9% training and 210% testing within the experimental dataset are shown in positive impact on tomato disease detection, with only 10 misclassiﬁed samples out of a Table 7. Because of this, we are able to deduce the performance of the suggested model, total of 1000 samples. which has the ability to categorize diseases with an accuracy that is more than 98.00%. Based on the data that are reﬂected in the table, it can be seen that the resulting recall 4.2.3. VIT Model Results metric value is high for each category that is speciﬁed in the dataset. The VIT model was The results achieved by the VIT model’s classiﬁcation performance for each class trained using eight batch sizes and the SGD optimizer for 100 epochs. The model achieved described with 9% training and 210% testing within the experimental dataset are shown in 98.00% accuracy on the training and testing data. The results of the VIT model, by splitt ing Table 7. Because of this, we are able to deduce the performance of the suggested model, the dataset into 80% training and 20% testing, are shown in Table S3. which has the ability to categorize diseases with an accuracy that is more than 98.00%. Based on the data that are reﬂected in the table, it can be seen that the resulting recall metric Table 7. Results of ViT model (90% training and 10% testing). value is high for each category that is speciﬁed in the dataset. The VIT model was trained using eight batch sizes and the SGD optimizer for 100 epochs. The model achieved 98.00% Class Name Precision (%) Recall (%) F1-Score (%) Support accuracy on the training and testing data. The results of the VIT model, by splitting the Bacterial_Spot 99 94 96 100 dataset into 80% training and 20% testing, are shown in Table S3. Early_Blight 97 99 98 100 Late_Blight 98 100 99 100 Leaf_Mold 99 96 97 100 Septoria_Leaf_Spot 99 99 99 100 Agronomy 2023, 13, 1184 17 of 24 Table 7. Results of ViT model (90% training and 10% testing). Class Name Precision (%) Recall (%) F1-Score (%) Support Agronomy 2023, 13, x FOR PEER REVIEW 19 of 26 Bacterial_Spot 99 94 96 100 Early_Blight 97 99 98 100 Late_Blight 98 100 99 100 Leaf_Mold 99 96 97 100 Spider_Mites_Two-Spotted_Spider_Mite 00 98 99 100 Septoria_Leaf_Spot 99 99 99 100 Target_Spot 95 99 97 100 Spider_Mites_Two-Spotted_Spider_Mite 00 98 99 100 Yellow_Leaf_Curl Virus 98 97 97 100 Target_Spot 95 99 97 100 Mosaic_Virus 100 100 100 100 Yellow_Leaf_Curl Virus 98 97 97 100 Mosaic_Virus 100 100 100 100 Healthy 95 98 97 100 Healthy 95 98 97 100 Accuracy Accuracy DenseNet121_Macro Avg 98 1000 DenseNet121_Macro Avg 98 1000 DenseNet121_Weighted Avg 98 98 98 1000 DenseNet121_Weighted Avg 98 98 98 1000 Bacterial_SpotBacterial_Spot 9 98 8 98 98 9898 10001000 Figure 11 shows a comparison of the training and validation accuracies by dividing Figure 11 shows a comparison of the training and validation accuracies by dividing the dataset in 90% training and 10% testing. Figure 11a demonstrates that an accuracy rate the dataset in 90% training and 10% testing. Figure 11a demonstrates that an accuracy rate of 98.00% may be obtained by sett ing the training stop at 100 epochs and keeping the rate of 98.00% may be obtained by setting the training stop at 100 epochs and keeping the rate of learning at 0.0001. Figure 11b compares validation and training losses. Consequently, of learning at 0.0001. Figure 11b compares validation and training losses. Consequently, given the study’s methodology, it is reasonable to predict that increasing the number of given the study’s methodology, it is reasonable to predict that increasing the number of iterations would lead to an increase in data accuracy. Conversely, if the training phase iterations would lead to an increase in data accuracy. Conversely, if the training phase duration increases, so will the total number of epochs. Figure S5 shows the accuracy and duration increases, so will the total number of epochs. Figure S5 shows the accuracy and loss performance of the VIT model using 80% training and 20% testing. loss performance of the VIT model using 80% training and 20% testing. Figure 11. VIT model performance; (a) VIT model accuracy and (b) VIT mode loss (90% training and Figure 11. VIT model performance; (a) VIT model accuracy and (b) VIT mode loss (90% training and 10% testing). 10% testing). As seen in Figure 12, the confusion matrix was devised to provide a visual represen- As seen in Figure 12, the confusion matrix was devised to provide a visual repre- tation of the classiﬁcation performance of the tomato leaf dataset obtained from the VIT sentation of the classiﬁcation performance of the tomato leaf dataset obtained from the model, or the number of images that were correctly categorized in a given number of iter- VIT model, or the number of images that were correctly categorized in a given number ations during validation and training, respectively. According to the VIT model’s confu- of iterations during validation and training, respectively. According to the VIT model’s sion matrix, the model has a strong positive impact on tomato disease detection, with only confusion matrix, the model has a strong positive impact on tomato disease detection, with 20 misclassiﬁed samples out of 1000 samples. Figure S6 shows the confusion matrix of VIT only 20 misclassiﬁed samples out of 1000 samples. Figure S6 shows the confusion matrix of model in 80% training and 20% testing. VIT model in 80% training and 20% testing. Agronomy 2023, 13, x FOR PEER REVIEW 20 of 26 Agronomy 2023, 13, 1184 18 of 24 Figure 12. Confusion matrix of ViT model (90% training and 10% testing). Figure 12. Confusion matrix of ViT model (90% training and 10% testing). 5. Discussion 5. Discussion Tomato is one of the most important crops worldwide and has high economic and Tomato is one of the most important crops worldwide and has high economic and nutritional value. However, it is often affected by various diseases, which can reduce both nutritional value. However, it is often aﬀected by various diseases, which can reduce both yield and quality. Early detection and accurate diagnosis of tomato diseases are crucial for yield and quality. Early detection and accurate diagnosis of tomato diseases are crucial effective disease management and prevention. for eﬀective disease management and prevention. This research further conﬁrms the effectiveness of proposed approaches to tomato This research further conﬁrms the eﬀectiveness of proposed approaches to tomato disease detection and provides valuable insight into different models’ performance. The disease detection and provides valuable insight into diﬀerent models’ performance. The results of this study have shown signiﬁcant accuracy for developing of automated systems results of this study have shown signiﬁcant accuracy for developing of automated systems that can detect and manage tomato plant diseases in the early stages, ultimately improving that can detect and manage tomato plant diseases in the early stages, ultimately improv- the efﬁciency and sustainability of tomato production. ing the eﬃciency and sustainability of tomato production. To diagnose tomato leaf diseases, this research examined multiple CNN models that To diagnose tomato leaf diseases, this research examined multiple CNN models that had been pretrained on the ImageNet dataset and then compared those models to the had been pretrained on the ImageNet dataset and then compared those models to the dataset. Three alternative CNN models—DenseNet121, ResNet50V2, and VIT—were dataset. Three alternative CNN models—DenseNet121, ResNet50V2, and VIT—were trained via transfer learning. Each model was educated and validated using the same trained via transfer learning. Each model was educated and validated using the same col- collection of tomato disease images, which included both infected and healthy examples. lection of tomato disease images, which included both infected and healthy examples. Ta- Table 8 shows the ﬁnal result of the proposed deep leaning model at training and testing ble 8 shows the ﬁnal result of the proposed deep leaning model at training and testing phase for detecting tomato disease. According to the ﬁndings, DenseNet121 successfully phase for detecting tomato disease. According to the ﬁndings, DenseNet121 successfully attained the highest possible training (99.88%) and test (99.00%) accuracies. In addition, its att ained the highest possible training (99.88%) and test (99.00%) accuracies. In addition, recall, accuracy, and F1 score were all above average, reaching 99.00%, 99.00%, and 98.99%, its recall, accuracy, and F1 score were all above average, reaching 99.00%, 99.00%, and respectively. ResNet50V2 obtained a training accuracy of 99.49% and a test accuracy 98.99%, respectively. ResNet50V2 obtained a training accuracy of 99.49% and a test accu- of 95.60%, both of which were lower than its predecessor. Its recall, accuracy, and F1 racy of 95.60%, both of which were lower than its predecessor. Its recall, accuracy, and F1 score were all lower than those of DenseNet121, reaching 95.60%, 94.80%, and 95.59%, score were all lower than those of DenseNet121, reaching 95.60%, 94.80%, and 95.59%, respectively. The accuracy of the VIT’s training and tests was also 98.00%. It had a recall of respectively. The accuracy of the VIT’s training and tests was also 98.00%. It had a recall 98.00%, a precision score of 98.00%, and an F1 score of 98.00%. The ﬁndings indicate that of 98.00%, a precision score of 98.00%, and an F1 score of 98.00%. The ﬁndings indicate DenseNet121 performed the best overall in identifying tomato diseases, followed by VIT that DenseNet121 performed the best overall in identifying tomato diseases, followed by and then ResNet50V2. The ROC plot of DenseNet121 is presented in Figure 13. VIT and then ResNet50V2. The ROC plot of DenseNet121 is presented in Figure 13. Agronomy 2023, 13, x FOR PEER REVIEW 21 of 26 Agronomy 2023, 13, 1184 19 of 24 Table 8. Summarized results of proposed deep learning in training and testing phases. Table 8. Summarized results of proposed deep learning in training and testing phases. Performance of Model Using 90% Training and 10% Testing of Dataset Performance of Model Using 90% Training and 10% Testing of Dataset Model Name Training Accuracy% Test Accuracy% Recall% Precision% F1 Score% Model Name Training Accuracy% Test Accuracy% Recall% Precision% F1 Score% DenseNet121 99.88 99 99 99 98.99 DenseNet121 99.88 99 99 99 98.99 ResNet50V2 99.49 95.60 95.60 95.8 95.59 ResNet50V2 99.49 95.60 95.60 95.8 95.59 ViT 98 98 98 98 97 ViT 98 98 98 98 97 Performance of Model Using 80% Training and 20% Testing of Dataset Performance of Model Using 80% Training and 20% Testing of Dataset Model Name Training Accuracy% Test Accuracy% Recall% Precision% F1 Score% Model Name Training Accuracy% Test Accuracy% Recall% Precision% F1 Score% DenseNet121 99.99 98.45 98.45 94.32 98.45 DenseNet121 99.99 98.45 98.45 94.32 98.45 ResNet50V2 99.99 94.43 94.31 98.46 94.31 ResNet50V2 99.99 94.43 94.31 98.46 94.31 ViT 99.27 94.90 94.90 94.97 94.89 ViT 99.27 94.90 94.90 94.97 94.89 Figure Figure 13. 13. ROC ROC curv curve e DenseNet121 DenseNet121 model model. . Additionally, a contrast was drawn between the level of difﬁculty provided by the Additionally, a contrast was drawn between the level of diﬃculty provided by the suggested model and that of a number of other models. The section devoted to accuracy suggested model and that of a number of other models. The section devoted to accuracy rates shows the individual successes of tomato leaf diseases in the datasets found in the rates shows the individual successes of tomato leaf diseases in the datasets found in the literature. The results of the proposed DenseNet121 model, which had the best rate of accu- literature. The results of the proposed DenseNet121 model, which had the best rate of racy (99.88%) in studies using a variety of methods other than the research recommended, accuracy (99.88%) in studies using a variety of methods other than the research recom- are shown in Table 9 and Figure 14. mended, are shown in Table 9 and Figure 14. Table 9. Comparison of results of DenseNet121 against different CNN models. Table 9. Comparison of results of DenseNet121 against diﬀerent CNN models. References Model Dataset Accuracy% References Model Dataset Accuracy% Ref. [62] Inception-v3 Same dataset 96.60 Ref. [62] Inception-v3 Same dataset 96.60 Ref. [63] ResNet-50 98.77 Ref. [63] ResNet-50 98.77 Ref. [64] MobileNet 88.4 Ref Ref.. [6 [65 4] ] Mob VGG16 ileNet 93.5 88.4 Ref. [66] CNN 98.2 Ref. [65] VGG16 93.5 Ref. [67] ResNet50 + SeNet 96.81 Ref. [66] CNN 98.2 Proposed Ref. [67] ResNet50 + SeNet 96.81 Propose system 99.64 DenseNet121 model Proposed Propose system 99.64 DenseNet121 model Agronomy 2023, 13, 1184 20 of 24 Agronomy 2023, 13, x FOR PEER REVIEW 22 of 26 Accuracy Inception-v3 ResNet-50 MobileNet VGG16 CNN ResNet50 + Proposed SeNet system Accuracy Figure 14. Comparison of results of DenseNet121 against diﬀerent CNN models in terms of accu- Figure 14. Comparison of results of DenseNet121 against different CNN models in terms of accuracy. racy. 6. Conclusions 6. Conclusions The purpose of this research was to analyze and apply many different deep-learning The purpose of this research was to analyze and apply many diﬀerent deep-learning models to the task of diagnosing diseases that affect tomato plants. The dataset from models to the task of diagnosing diseases that aﬀect tomato plants. The dataset from Kaggle, which is readily accessible, was selected since it contains information on ten Kaggle, which is readily accessible, was selected since it contains information on ten dif- different diseases that might affect tomato plant leaves. Three novel CNN architectures ferent diseases that might aﬀect tomato plant leaves. Three novel CNN architectures were were proposed for disease prediction and classiﬁcation in tomato plants using this dataset, proposed for disease prediction and classiﬁcation in tomato plants using this dataset, alongside the specialized deep learning architectures DenseNet121, ResNet50V2, and alongside the specialized deep learning architectures DenseNet121, ResNet50V2, and ViT ViT models. models. Deep learning has emerged as a powerful tool for detecting diseases in plants, in- cluding Deep le tomatoes. arning Using has emerge neurald as networks a pow to eranal ful tyze ool fo larr d gee datasets tecting di ofsea plant ses in images, plants deep , in- learning models can learn to identify patterns and features that are characteristic of spe- cluding tomatoes. Using neural networks to analyze large datasets of plant images, deep ciﬁc learning diseases. models This cancan learhelp n to ident farme ify rs pa andtt erns resear and chers features quickly thatand are cha accurately racteristdiagnose ic of spe- plant ciﬁc di diseases, seases. This allowing can help for farm mor er es tar angeted d resea tr rchers eatment quickl and y and prevent accu ion rate strategies. ly diagnose Based plant on the results of this study, it can be concluded that deep learning models, including diseases, allowing for more targeted treatment and prevention strategies. Based on the DenseNet121, ResNet50V2, and ViT, are effective for tomato disease detection. Among results of this study, it can be concluded that deep learning models, including Dense- these models, DenseNet121 showed the highest accuracy, with 99.88% on training data Net121, ResNet50V2, and ViT, are eﬀective for tomato disease detection. Among these and 99% on testing data. The ResNet50V2 and ViT models also showed high accuracy, models, DenseNet121 showed the highest accuracy, with 99.88% on training data and 99% but their performance was lower than DenseNet121. These ﬁndings suggest that deep on testing data. The ResNet50V2 and ViT models also showed high accuracy, but their learning models can provide accurate and efﬁcient solutions for tomato disease detection, performance was lower than DenseNet121. These ﬁndings suggest that deep learning which can ultimately beneﬁt the agriculture industry. The obtained results of the analysis models can provide accurate and eﬃcient solutions for tomato disease detection, which showed that the suggested model performed better than alternative models. The method can ultimately beneﬁt the agriculture industry. The obtained results of the analysis that has been presented for recognizing the diseases that affects tomatoes is an innovative showed that the suggested model performed bett er than alternative models. The method one. One of the limitations of this research is that the system has not been incorporated that has been presented for recognizing the diseases that aﬀects tomatoes is an innovative into a mobile application. However, it does provide a simple and low-cost method for one. One of the limitations of this research is that the system has not been incorporated diagnosing tomato leaf diseases by only requiring the user to take an image of the affected into a mobile application. However, it does provide a simple and low-cost method for plant’s leaf. In the future, we want to enhance the model by using more advanced forms of diagnosing tomato leaf diseases by only requiring the user to take an image of the aﬀected artiﬁcial intelligence that are supported by internet of things (IoT) technology. plant’s leaf. In the future, we want to enhance the model by using more advanced forms of artiﬁcial intelligence that are supported by internet of things (IoT) technology. Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/agronomy13051184/s1, Figure S1: Performance of DenseNet121 Supplementary Materials: The following supporting information can be downloaded at: model; (a) DenseNet121 model accuracy, (b) DenseNet121 mode loss with (80% training and 20% testing; www.mdpi.com/xxx/s1, Figure S1: Performance of DenseNet121 model; (a) DenseNet121 model ac- Figure S2: Confusion matrix of DenseNet121 with (80% training and 20 testing); Figure S3: Performance Agronomy 2023, 13, 1184 21 of 24 of the ResNet50V2 model; (a) ResNet50V2model accuracy and (b) ResNet50V21 mode loss with (80% training and 20% testing); Figure S4: The ResNet50V2 model’s confusion matrix with (80% training and 20% testing); Figure S5: VIT model performance; (a) VIT model accuracy and (b) VIT mode loss with (80% training and 20% testing); Figure S6: Confusion matrix of ViT model with (80% training and 20% testing); Table S1. Results of DenseNet121 model in (80% training and 20% testing); Table S2: Results of ResNet50V2 model with (80% training and 20% testing); Table S3: Results of ViT model with (80% training and 20% testing). Author Contributions: Conceptualization, M.S.A. and F.W.A.; methodology, M.S.A. and F.W.A.; software, M.S.A. and F.W.A.; validation, M.S.A. and F.W.A. formal analysis, M.S.A. and F.W.A. investigation, M.S.A. and F.W.A. resources, M.S.A. and F.W.A. data curation, M.S.A. and F.W.A. writing—original draft preparation, M.S.A. and F.W.A. writing—review and editing, M.S.A. and F.W.A.; visualization, M.S.A. and F.W.A. supervision, M.S.A. and F.W.A.; project administration, M.S.A. and F.W.A. funding acquisition, M.S.A. and F.W.A. All authors have read and agreed to the published version of the manuscript. Funding: The authors extend their appreciation to the Deputyship for Research & Innovation, Ministry of Education in Saudi Arabia for funding this research work through the project num- ber INST048. Data Availability Statement: https://www.kaggle.com/datasets/kaustubhb999/tomatoleaf (accessed on 20 January 2023). Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Schreinemachers, P.; Simmons, E.B.; Wopereis, M.C. Tapping the economic and nutritional power of vegetables. Glob. Food Secur. 2018, 16, 36–45. [CrossRef] 2. Stilwell, M. The Global Tomato Online News Processing in 2018. Available online: https://www.tomatonews.com/ (accessed on 15 February 2023). 3. Wang, R.; Lammers, M.; Tikunov, Y.; Bovy, A.G.; Angenent, G.C.; de Maagd, R.A. The rin, nor and Cnr spontaneous mutations inhibit tomato fruit ripening in additive and epistatic manners. Plant Sci. 2020, 294, 110436–110447. [CrossRef] [PubMed] 4. Ahmad, A.; Saraswat, D.; El Gamal, A. A survey on using deep learning techniques for plant disease diagnosis and recommenda- tions for development of appropriate tools. Smart Agric. Technol. 2023, 3, 100083. [CrossRef] 5. Yu, Y.; Samali, B.; Rashidi, M.; Mohammadi, M.; Nguyen, T.N.; Zhang, G. Vision-based concrete crack detection using a hybrid framework considering noise effect. J. Build. Eng. 2022, 61, 105246. [CrossRef] 6. Sahin, ¸ M.E.; Ulutas, ¸ H.; Yuce, E. A deep learning approach for detecting pneumonia in chest X-rays. Avrupa Bilim Teknol. Derg. 2021, 28, 562–567. [CrossRef] 7. Bock, C.H.; Parker, P.E.; Cook, A.Z.; Gottwald, T.R. Visual Rating and the Use of Image Analysis for Assessing Different Symptoms of Citrus Canker on Grapefruit Leaves. Plant Dis. 2008, 92, 530–541. [CrossRef] 8. Yu, Y.; Liang, S.; Samali, B.; Nguyen, T.N.; Zhai, C.; Li, J.; Xie, X. Torsional capacity evaluation of RC beams using an improved bird swarm algorithm optimised 2D convolutional neural network. Eng. Struct. 2022, 273, 115066. [CrossRef] 9. Spantideas, S.T.; Giannopoulos, A.E.; Kapsalis, N.C.; Capsalis, C.N. A deep learning method for modeling the magnetic signature of spacecraft equipment using multiple magnetic dipoles. IEEE Magn. Lett. 2021, 12, 1–5. [CrossRef] 10. Park, H.; Eun, J.S.; Kim, S.H. Image-based disease diagnosing and predicting of the crops through the deep learning mechanism. In Proceedings of the 2017 International Conference on Information and Communication Technology Convergence (ICTC), Jeju Island, Republic of Korea, 18–20 October 2017; pp. 129–131. 11. Sardogan, M.; Tuncer, A.; Ozen, Y. Plant leaf disease detection and classiﬁcation based on CNN with LVQ algorithm. In Proceedings of the 2018 3rd International Conference on Computer Science and Engineering (UBMK), Sarajevo, Bosnia and Herzegovina, 20–23 September 2018; pp. 382–385. 12. Singh, V.; Misra, A.K. Detection of plant leaf diseases using image segmentation and soft computing techniques. Inf. Process. Agric. 2017, 4, 41–49. [CrossRef] 13. Zhang, S.; Wu, X.; You, Z.; Zhang, L. Leaf image based cucumber disease recognition using sparse representation classiﬁcation. Comput. Electron. Agric. 2017, 134, 135–141. [CrossRef] 14. Ferentinos, K.P. Deep learning models for plant disease detection and diagnosis. Comput. Electron. Agric. 2018, 145, 311–318. [CrossRef] 15. Devaraj, A.; Rathan, K.; Jaahnavi, S.; Indira, K. Identiﬁcation of Plant Disease using Image Processing Technique. In Proceedings of the 2019 International Conference on Communication and Signal Processing (ICCSP), Chennai, India, 4–6 April 2019; pp. 749–753. Agronomy 2023, 13, 1184 22 of 24 16. Mugithe, P.K.; Mudunuri, R.V.; Rajasekar, B.; Karthikeyan, S. Image Processing Technique for Automatic Detection of Plant Diseases and Alerting System in Agricultural Farms. In Proceedings of the 2020 International Conference on Communication and Signal Processing (ICCSP), Chennai, India, 28–30 July 2020; pp. 1603–1607. 17. Phadikar, S.; Sil, J. Rice disease identiﬁcation using pattern recognition techniques. In Proceedings of the 2008 11th International Conference on Computer and Information Technology, Khulna, Bangladesh, 24–27 December 2008; pp. 420–423. 18. Sarayloo, Z.; Asemani, D. Designing a classiﬁer for automatic detection of fungal diseases in wheat plant: By pattern recognition techniques. In Proceedings of the 2015 23rd Iranian Conference on Electrical Engineering, Tehran, Iran, 10–14 May 2015; pp. 1193–1197. 19. Thangadurai, K.; Padmavathi, K. Computer Visionimage Enhancement for Plant Leaves Disease Detection. In Proceedings of the 2014 World Congress on Computing and Communication Technologies, Trichirappalli, India, 27 February–1 March 2014; pp. 173–175. 20. Yong, Z.; Tonghui, R.; Changming, L.; Chao, W.; Jiya, T. Research on Recognition Method of Common Corn Diseases Based on Computer Vision. In Proceedings of the 2019 11th International Conference on Intelligent Human-Machine Systems and Cybernetics (IHMSC), Hangzhou, China, 24–25 August 2019; Volume 1, pp. 328–331. 21. Khirade, S.D.; Patil, A.B. Plant Disease Detection Using Image Processing. In Proceedings of the 2015 International Conference on Computing Communication Control and Automation, Pune, India, 26–27 February 2015; pp. 768–771. 22. Kamilaris, A.; Prenafeta-Boldú, F.X. Deep learning in agriculture: A survey. Comput. Electron. Agric. 2018, 147, 70–90. [CrossRef] 23. Li, L.; Zhang, S.; Wang, B. Plant Disease Detection and Classiﬁcation by Deep Learning—A Review. IEEE Access 2021, 9, 56683–56698. [CrossRef] 24. Lee, S.H.; Chan, C.S.; Wilkin, P.; Remagnino, P. Deep-plant: Plant identiﬁcation with convolutional neural networks. In Proceedings of the 2015 IEEE International Conference on Image Processing (ICIP), Quebec City, QC, Canada, 27–30 September 2015; pp. 452–456. 25. Zhang, Y.; Song, C.; Zhang, D. Deep Learning-Based Object Detection Improvement for Tomato Disease. IEEE Access 2020, 8, 56607–56614. [CrossRef] 26. Sabrol, H.; Satish, K. Tomato plant disease classification in digital images using classification tree. In Proceedings of the International Conference on Communication and Signal Processing (ICCSP), Melmaruvathur, India, 6–8 April 2016; pp. 1242–1246. 27. Hasan, M.; Tanawala, B.; Patel, K.J. Deep learning precision farming: Tomato leaf disease detection by transfer learning. In Proceedings of the 2nd International Conference on Advanced Computing and Software Engineering (ICACSE), Sultanpur, Inida, 8–9 February 2019. 28. Adhikari, S.; Shrestha, B.; Baiju, B.; Kumar, S. Tomato plant diseases detection system using image processing. In Proceedings of the 1st KEC Conference on Engineering and Technology, Laliitpur, Nepal, 27 September 2018; Volume 1, pp. 81–86. 29. Salih, T.A. Deep Learning Convolution Neural Network to Detect and Classify Tomato Plant Leaf Diseases. Open Access Libr. J. 2020, 7, 12. [CrossRef] 30. Ishak, S.; Rahiman, M.H.; Kanaﬁah, S.N.; Saad, H. Leaf disease classiﬁcation using artiﬁcial neural network. J. Teknol. 2015, 77, 109–114. [CrossRef] 31. Sabrol, H.; Kumar, S. Fuzzy and neural network-based tomato plant disease classiﬁcation using natural outdoor images. Indian J. Sci. Technol. 2016, 9, 1–8. [CrossRef] 32. Rangarajan, A.K.; Purushothaman, R.; Ramesh, A. Tomato crop disease classiﬁcation using pre-trained deep learning algorithm. Procedia Comput. Sci. 2018, 133, 1040–1047. [CrossRef] 33. Coulibaly, S.; Kamsu-Foguem, B.; Kamissoko, D.; Traore, D. Deep neural networks with transfer learning in millet crop images. Comput. Ind. 2019, 108, 115–120. [CrossRef] 34. Sangeetha, R.; Rani, M. Tomato Leaf Disease Prediction Using Transfer Learning. In Proceedings of the International Advanced Computing Conference 2020, Panaji, India, 5–6 December 2020. 35. Mortazi, A.; Bagci, U. Automatically designing CNN architectures for medical image segmentation. In Proceedings of the International Workshop on Machine Learning in Medical Imaging, Granada, Spain, 16 September 2018; pp. 98–106. 36. Jiang, D.; Li, F.; Yang, Y.; Yu, S. A tomato leaf diseases classiﬁcation method based on deep learning. In Proceedings of the Chinese Control and Decision Conference (CCDC), Hefei, China, 22–24 August 2020; pp. 1446–1450. 37. Rashid, J.; Khan, I.; Ali, G.; Almotiri, S.H.; AlGhamdi, M.A.; Masood, K. Multi-Level Deep Learning Model for Potato Leaf Disease Recognition. Electronics 2021, 10, 2064. [CrossRef] 38. PlantVillege. Available online: https://www.kaggle.com/emmarex/plantdisease (accessed on 3 February 2023). 39. Lakshmanarao, A.; Babu, M.R.; Kiran, T.S.R. Plant Disease Prediction and classiﬁcation using Deep Learning ConvNets. In Proceedings of the 2021 International Conference on Artiﬁcial Intelligence and Machine Vision (AIMV), Gandhinagar, India, 24–26 September 2021; pp. 1–6. 40. Militante, S.V.; Gerardo, B.D.; Dionisio, N.V. Plant Leaf Detection and Disease Recognition using Deep Learning. In Proceedings of the 2019 IEEE Eurasia Conference on IOT, Communication and Engineering (ECICE), Yunlin, Taiwan, 3–6 October 2019; pp. 579–582. Agronomy 2023, 13, 1184 23 of 24 41. Marzougui, F.; Elleuch, M.; Kherallah, M. A Deep CNN Approach for Plant Disease Detection. In Proceedings of the 2020 21st International Arab Conference on Information Technology (ACIT), Giza, Egypt, 28–30 November 2020; pp. 1–6. 42. Ngugi, L.C.; Abdelwahab, M.; Abo-Zahhad, M. Tomato leaf segmentation algorithms for mobile phone applications using deep learning. Comput. Electron. Agric. 2020, 178, 105788. [CrossRef] 43. Rozaqi, A.J.; Sunyoto, A. Identiﬁcation of Disease in Potato Leaves Using Convolutional Neural Network (CNN) Algorithm. In Proceedings of the 2020 3rd International Conference on Information and Communications Technology (ICOIACT), Yogyakarta, Indonesia, 24–25 November 2020; pp. 72–76. 44. Mattihalli, C.; Gedefaye, E.; Endalamaw, F.; Necho, A. Real Time Automation of Agriculture Land, by automatically Detecting Plant Leaf Diseases and Auto Medicine. In Proceedings of the 2018 32nd International Conference on Advanced Information Networking and Applications Workshops (WAINA), Krakow, Poland, 16–18 May 2018; pp. 325–330. 45. Divyashri., P.; Pinto, L.A.; Mary, L.; Manasa., P.; Dass, S. The Real-Time Mobile Application for Identiﬁcation of Diseases in Coffee Leaves using the CNN Model. In Proceedings of the 2021 Second International Conference on Electronics and Sustainable Communication Systems (ICESC), Coimbatore, India, 4–6 August 2021; pp. 1694–1700. 46. Liu, J.; Wang, X. Early recognition of tomato gray leaf spot disease based on MobileNetv2-YOLOv3 model. Plant Methods 2020, 16, 83. [CrossRef] [PubMed] 47. Khasawneh, N.; Faouri, E.; Fraiwan, M. Automatic Detection of Tomato Diseases Using Deep Transfer Learning. Appl. Sci. 2022, 12, 8467. [CrossRef] 48. Mim, T.T.; Sheikh, M.H.; Shampa, R.A.; Reza, M.S.; Islam, M.S. Leaves Diseases Detection of Tomato Using Image Processing. In Proceedings of the 2019 8th International Conference System Modeling and Advancement in Research Trends (SMART), Moradabad, India, 22–23 November 2019; pp. 244–249. 49. Kumar, A.; Vani, M. Image Based Tomato Leaf Disease Detection. In Proceedings of the 2019 10th International Conference on Computing, Communication and Networking Technologies (ICCCNT), Kanpur, India, 6–8 July 2019; pp. 1–6. 50. Tm, P.; Pranathi, A.; SaiAshritha, K.; Chittaragi, N.B.; Koolagudi, S.G. Tomato Leaf Disease Detection Using Convolutional Neural Networks. In Proceedings of the 2018 Eleventh International Conference on Contemporary Computing (IC3), Noida, India, 2–4 August 2018; pp. 1–5. 51. Kang, F.; Li, J.; Wang, C.; Wang, F. A Lightweight Neural Network-Based Method for Identifying Early-Blight and Late-Blight Leaves of Potato. Appl. Sci. 2023, 13, 1487. [CrossRef] 52. Al-Gaashani, M.S.A.M.; Shang, F.; Muthanna, M.S.A.; Khayyat, M.; El-Latif, A.A.A. Tomato leaf disease classiﬁcation by exploiting transfer learning and feature concatenation. IET Image Process. 2022, 16, 913–925. [CrossRef] 53. Pathan, S.M.K.; Ali, M.F. Implementation of Faster R-CNN in Paddy Plant Disease Recognition System. In Proceedings of the 2019 3rd International Conference on Electrical, Computer & Telecommunication Engineering (ICECTE), Rajshahi, Bangladesh, 26–28 December 2019; pp. 189–192. 54. Zhou, G.; Zhang, W.; Chen, A.; He, M.; Ma, X. Rapid Detection of Rice Disease Based on FCM-KM and Faster R-CNN Fusion. IEEE Access 2019, 7, 143190–143206. [CrossRef] 55. Cardellicchio, A.; Solimani, F.; Dimauro, G.; Petrozza, A.; Summerer, S.; Cellini, F.; Renò, V. Detection of tomato plant phenotyping traits using YOLOv5-based single stage detectors. Comput. Electron. Agric. 2023, 207, 107757. [CrossRef] 56. Liu, G.; Nouaze, J.C.; Touko Mbouembe, P.L.; Kim, J.H. YOLO-Tomato: A Robust Algorithm for Tomato Detection Based on YOLOv3. Sensors 2020, 20, 2145. [CrossRef] 57. Liu, J.; Wang, X. Tomato Diseases and Pests Detection Based on Improved Yolo V3 Convolutional Neural Network. Front. Plant Sci. 2020, 11, 898. [CrossRef] 58. Aldhyani, T.H.H.; Nair, R.; Alzain, E.; Alkahtani, H.; Koundal, D. Deep Learning Model for the Detection of Real Time Breast Cancer Images Using Improved Dilation-Based Method. Diagnostics 2022, 12, 2505. [CrossRef] 59. Aldhyani, T.H.; Alshebami, A.S.A.; Alzahrani, M.Y. Soft Computing Model to Predict Chronic Diseases. J. Inf. Sci. Eng. 2020, 36, 365–376. 60. Al-Adhaileh, M.H.; Aldhyani, T.H.H. Artiﬁcial intelligence framework for modeling and predicting crop yield to enhance food security in Saudi Arabia. PeerJ Comput. Sci. 2022, 2022, e1104. [CrossRef] [PubMed] 61. Al-Adhaileh, M.H.; Verma, A.; Aldhyani, T.H.H.; Koundal, D. Potato Blight Detection Using Fine-Tuned CNN Architecture. Mathematics 2023, 11, 1516. [CrossRef] 62. Widiyanto, S.; Fitrianto, R.; Wardani, D.T. Implementation of Convolutional Neural Network Method for Classiﬁcation of Diseases in Tomato Leaves. In Proceedings of the 2019 Fourth International Conference on Informatics and Computing (ICIC), Semarang, Indonesia, 16–17 October 2019; pp. 1–5. 63. Mamun, M.A.A.; Karim, D.Z.; Pinku, S.N.; Bushra, T.A. TLNet: A Deep CNN model for Prediction of tomato Leaf Diseases. In Proceedings of the 2020 23rd International Conference on Computer and Information Technology (ICCIT), Dhaka, Bangladesh, 19–21 December 2020; pp. 1–6. 64. Elhassouny, A.; Smarandache, F. Smart mobile application to recognize tomato leaf diseases using Convolutional Neural Networks. In Proceedings of the 2019 International Conference of Computer Science and Renewable Energies (ICCSRE), Agadir, Morocco, 22–24 July 2019; pp. 1–4. 65. Agarwal, M.; Gupta, S.K.; Biswas, K.K. Development of Efﬁcient CNN model for Tomato crop disease identiﬁcation. Sustain. Comput. Inform. Syst. 2020, 28, 100407. [CrossRef] Agronomy 2023, 13, 1184 24 of 24 66. Lamba, M.; Gigras, Y.; Dhull, A. Classiﬁcation of plant diseases using machine and deep learning. Open Comput. Sci. 2021, 11, 491–508. [CrossRef] 67. Zhao, S.; Peng, Y.; Liu, J.; Wu, S. Tomato leaf disease diagnosis based on improved convolution neural network by attention module. Agriculture 2021, 11, 651. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Journal

Agronomy – Multidisciplinary Digital Publishing Institute

Published: Apr 22, 2023

Keywords: artificial intelligence; deep leaning; transform learning; food security; plant diseases

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease

Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease

Transform and Deep Learning Algorithms for the Early Detection and Recognition of Tomato Leaf Disease

References (62)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies