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

Learn More →

Development and evaluation of a facile mesh-to-surface tool for customised wheelchair cushions

Development and evaluation of a facile mesh-to-surface tool for customised wheelchair cushions Background Custom orthoses are becoming more commonly prescribed for upper and lower limbs. They require some form of shape-capture of the body parts they will be in contact with, which generates an STL file that design- ers prepare for manufacturing. For larger devices such as custom-contoured wheelchair cushions, the STL created during shape-capture can contain hundreds of thousands of tessellations, making them difficult to alter and prepare for manufacturing using mesh-editing software. This study covers the development and testing of a mesh-to-surface workflow in a parametric computer-aided design software using its visual programming language such that STL files of custom wheelchair cushions can be efficiently converted into a parametric single surface. Methods A volunteer in the clinical space with expertise in computer-aided design aided was interviewed to under- stand and document the current workflow for creating a single surface from an STL file of a custom wheelchair cush- ion. To understand the user needs of typical clinical workers with little computer-aided design experience, potential end-users of the process were tasked with completing the workflow and providing feedback during the experience. This feedback was used to automate part of the computer-aided design process using a visual programming tool, cre- ating a new semi-automated workflow for mesh-to-surface translation. Both the original and semi-automated process were then evaluated by nine volunteers with varying levels of computer-aided design experience. Results The semi-automated process showed a 37% reduction in the total number of steps required to convert an STL model to a parametric surface. Regardless of previous computer-aided design experience, volunteers completed the semi-automated workflow 31% faster on average than the manual workflow. Conclusions The creation of a semi-automated process for creating a single parametric surface of a custom wheel- chair cushion from an STL mesh makes mesh-to-surface conversion more efficient and more user-friendly to all, regardless of computer-aided design experience levels. The steps followed in this study may guide others in the development of their own mesh-to-surface tools in the wheelchair sector, as well as those creating other large cus- tom prosthetic devices. Keywords Wheelchair, Cushion, Grasshopper, Scan to solid, Mesh editing Introduction Wheelchairs enable autonomy in many individuals across *Correspondence: Aisling Ní Annaidh the world but sitting in a wheelchair for long periods can aisling.niannaidh@ucd.ie lead to pain, discomfort, and injuries such as pressure School of Mechanical and Materials Engineering, University College ulcers and spinal deformities. To prevent or correct such Dublin, Belfield, Dublin, Ireland SeatTech Posture and Mobility Services, Enable Ireland, Dublin, Ireland issues, customized wheelchair seating is manufactured for individual needs. Custom contoured seating (CCS) © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Nace et al. 3D Printing in Medicine (2023) 9:3 Page 2 of 12 refers to customized seating that is based on an impres- While the software is part of a cleared medical device, to sion taken of a wheelchair user’s body and matches date all such clearances have been for anatomic models exactly to that shape. Custom wheelchair seating systems and anatomic guides for parts 3D printed from CT, MR, require shape capture of a surface moulded to the shape and ultrasound images [6]. Such a software tool is not of the wheelchair user to design an effective personal - necessarily usable for other products due to this system- ised postural aid and pressure relieving intervention [1]. based method, but it suggests the need for further devel- Methods to capture shapes include contact digitisers and opment 3D printing software packages and a broadening non-contact scanning, including high-end time-of-flight of their potential approved use cases to match the wide or triangulation laser scanners as well as more affordable range of products that can be made using 3D printing. structured-light scanners [2]. To widen the range of potential products, we can look Patient specific 3D printing typically begins with CT to CAD software packages which enable users to design and MR images that are then segmented and converted and prepare models for manufacturing using direct edit- to printable files [3]. However, there are a large portfolio ing or parametric modelling instead of mesh-only edit- of devices customized to a patient for which the Stand- ing software. Some CAD software offer mesh editing ard Tessellation File (STL) file is created via alternative tools similar to the mesh-specific software, and more and strategies. These scanning techniques include laser trigo - more CAD software producers are attempting to provide nometric triangulation-based scanners and structured a means to convert mesh objects to non-uniform rational light scanners which both generate the digital recreation basis splines (NURBS) objects that are editable within a of an object or surface as a point cloud or a mesh object. CAD environment. The NURBS object produced from Mesh objects are made up of tessellations that, when such one-click conversions have as many surfaces as imported into a computer-aided design (CAD) program, the mesh object had tessellations; for 3D scans, this can are not editable because they are recognised as either a mean an object with tens of thousands of small surfaces single entity or as thousands of small, individual surfaces. that have to be manipulated when preparing the model Because of this, if a 3D scan needs to be edited before for manufacturing, requiring large amounts of computer manufacturing – to make a solid cushion or back support power [7]. A NURBS object with hundreds or thousands with the scanned surface, for example – the scan must be of surfaces is not a user-friendly object for those unfamil- altered in a mesh-specific software or ‘converted’ into a iar with CAD, such as the clinical teams prescribing and CAD model and edited in a CAD environment. Mesh- producing custom prosthetics, orthoses, and wheelchair specific software, like Meshmixer (Autodesk) or Mesh - seating. Lab (ISTI-CNR), offer sculpting as well as refinement and This leads to a need for other ways to convert meshes smoothing tools to alter the shape of a mesh object itera- to CAD models. Recent research has attempted to fill tively. Fernandez-Vicente et al. [4] developed an inexpen- this void, specifically for custom prostheses and orthoses sive, efficient method for scanning, designing, and 3D designed from body scans. Blaya et  al. [8], Fernandez- printing a thumb orthosis using only Meshmixer as the Vicente et  al. [4], Baronio et  al. [9], Palousek et  al. [10], digital model creation tool. The new design and fabrica - and Paterson et  al. [11] developed five different digital tion method suggested digitisation and 3D printing can workflows for designing custom wrist orthoses or splints cut costs and production time for orthosis producers. for3D printing. Works by Blaya et  al. [8], Fernandez- However, this process was not tested with clinical spe- Vicente et  al. [4], and Paterson et  al. [11] suggest that cialists, who may be less familiar with digital design tools digital mesh models generated from 3D scans of wrists and require more time and training to create an appro- generate fewer tessellations than those of moulded cush- priate orthosis. Furthermore, mesh editing programs ions due to the size of wrist scans. This difference may like Meshmixer do not offer parametric editing tools make using mesh editing tools a less taxing task for wrist which would be used to alter measurable dimensions of orthosis model generation than for CCS. However, these a wheelchair cushion, such as the height or depth of the investigations did not evaluate the efficiency or user- cushion for a better fit in a wheelchair base. Thus, for friendliness of the digital workflows, nor did they evalu - designing medical devices like wheelchair cushions that ate the digital workflows with clinical experts to gauge require more than the scanned shape in the model crea- the viability of the processes in a clinical setting. tion process [1], they do not meet user needs for digital Baronio et  al. [9] developed a digital workflow that alterations prior to manufacturing. used both mesh editing software and NURBS conver- There are 3D printing software packages cleared by sion for CAD editing. This study deemed CAD editing the United States Food and Drug Administration (FDA) necessary in their digital workflow due to the volumet - for specific intended uses as part of a system that must ric modelling needed to create an effective hand ortho - include a specific 3D printer, material, and anatomy [5]. sis that was comfortable, light, and well-designed for its Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 3 of 12 function. The authors discussed in their study how more 6) and is documented in a step-by-step format [see Addi- user-friendly digital editing tools are needed in the clini- tional file  1]. Briefly, three key actions were identified in cal environment to use scanning and 3D printing to make the process: orienting the scan in the modelling space; custom orthoses, otherwise a CAD specialist is needed defining the STL surface with a set of curves; and using on the clinical team, implying that this may be a barrier the curves to create a new CAD surface in the shape of to adoption for clinical teams. In their 2014 study on the the scan. These actions are shown in process order in design and 3D printing of custom hand-wrist orthoses in Fig. 1. a clinical setting, Palousek et al. [10] similarly concluded User needs trial that the digital workflow would require a CAD specialist Initial development work focussed on identifying user or the automation of the workflow tasks that required the needs of the clinical team. For this reason, the original CAD specialist. manual step-by-step process was provided to two poten- In 2018, Li and Tanaka [12] created a process that auto- tial clinical end-users to complete as a tutorial. The two mated some of the CAD modelling tasks required in the clinical users (volunteers C1 and C2) had varying levels digital design of custom wrist orthoses using Rhinoc- of CAD experience and worked in the wheelchair seat- eros 3D and its visual programming plugin, Grasshop- ing industry at the time that they completed the tutorials. per (Rhinoceros 3D, McNeel & Associates). This study Each volunteer gave live feedback during monitored trials found the new workflow required 2 to 3 min of computer of the manual process to capture user needs of the clini- work compared to 20 min to 3 h of CAD modelling in the cal team. The conversation during each test was open, to traditional digital splint process [13]. The authors show allow for any constructive feedback and each volunteer evidence in their study that the automated tasks created was asked a mixture of opened-ended questions [see using Grasshopper could be used in the design of wear- Additional file 3]. able orthotic devices and prostheses. In a separate study, Volunteer C1, a beginner CAD user, took approxi- Li and Tanaka [12] trialled their new process with five mately 21  min to complete the guide and had difficulty nursing students unfamiliar with CAD software. Their locating some tools and with using the ‘gumball’ tool. study indicated that an automated process with a sim- Specifically volunteer C1 could not locate the ‘line tool’ plified software interface enabled the nursing students and the ‘surface from the network or curves’ tool. Other to learn how to correctly use the process to design hand issues highlighted included (i) making the gridlines that orthoses from 3D scans in less than one workday. would be projected onto the STL scan and (ii) selecting In this manuscript, a post-processing method to con- multiple lines in the modelling workspace. For (i), the vert scans of custom back support and seat cushion volunteer did not know how many lines should be pre- surfaces into 3D CAD models was developed using Rhi- sent; for (ii) the volunteer could not select multiple lines. noceros 3D’s Grasshopper plugin and trialled by poten- tial clinical users of the process. Feedback from the trials was used to prototype a more user-friendly interface to the CAD environment and to simplify the conversion process further to make it more accessible to those with little to no CAD experience. The process used herein was based on Li and Tanaka’s work on the digital design of wrist orthoses in clinical settings [12]. The manuscript is divided into three main tasks: the documentation and assessment of the manual method for CCS design; the development of an automated workflow; and the evalua - tion of the proposed automated workflow. Methods Capture of existing manual workflow The first step in the development of an improved scan- to-print workflow was the capture and documentation of the existing process implemented by an expert clini- cal CAD user (volunteer C0). The researchers recorded the manual CAD method implemented by volunteer C0 Fig. 1 The three key actions in the workflow of converting an STL of in the prescription and manufacture of CCS systems. a custom cushion scan into a CAD surface in Rhinoceros software The original method employed Rhinoceros 3D (Version Nace et al. 3D Printing in Medicine (2023) 9:3 Page 4 of 12 Upon completing the trial, volunteer C1 described the own set of skills to learn beyond the traditional CAD process as too long and suggested that the process should environment. However, Grasshopper has methods of cre- be as automated as possible to simplify the process. They ating separate Graphic User Interfaces (GUIs) with which stated the process should take no longer than 15  min to users of the routine can interact instead of the visual pro- complete. gramming environment of Grasshopper. To coincide with Volunteer C2, who had no previous experience with the user-need of “ease of use” in this project, the authors any CAD tools, took approximately 38  min to complete decided to use Grasshopper’s tools to create a simple user the guide. Volunteer C2 struggled to understand the dif- interface for the automated workflow. ferent viewpoints in the Rhinoceros software and had To create the user interface within Grasshopper, the issues creating the gridlines. The volunteer described the modelling workflow first had to be translated into Grass - process as “time-consuming” and stated that the ideal hopper’s environment. The steps taken in the manual process should take less than 30 min. They suggested that workflow were translated into Grasshopper’s visual pro - a dialogue box on screen identifying the next steps would gramming language. Figure 2 shows the entire Grasshop- be helpful. per program, broken down into action blocks 1 through The key user needs identified from feedback during 7. Blocks 1 through 6 are directly related to making a the trial are listed in Table  1. These (clinical) user needs CAD-editable model from the STL scan and block 7 cre- were used to tailor the proposed automated process ates the user interface that allows the user to set parame- for the anticipated future users. Analysis of these user ters of the model in blocks 1 through 6 without changing needs led to the determination of the following design the Grasshopper file. The user interface is shown in Fig.  3. requirements: The function of blocks 1 through 6 of the scan-to-CAD Grasshopper code, and how the user interface employs 1. Automate as many steps as possible, including the them, are further described in the following sections. A orientation and creation of the surface step and the step-by-step guide for using the Grasshopper scan-to- gridlines step. CAD process with the custom user interface tool in Fig. 3 2. Customise a toolbar or user interface to remove can be found in Additional file 2. tools in Rhino’s interface that are not required and decrease overall time requirement. Block 1: Orienting the scan using Grasshopper The orientation of the STL scan in a 3D workspace dic - tates how it is translated into a 3D CAD model. The Grasshopper program requires the user to orient the Semi‑automated workflow development model such that the surface that conforms to the user This section details the development of a semi-auto - can be seen in the XY-plane and that the front of the seat mated workflow that utilizes Grasshopper in Rhinoc - cushion or the bottom of the back support runs parallel eros 3D with the goal of satisfying the defined user needs to the Y-axis, as shown in Fig. 4. gathered from the trial with clinical users. Grasshopper Block 1 enables the user to orient the STL cushion scan is a plug-in for Rhinoceros that enables the automation in the Rhinoceros modelling space regardless of the scan’s of many modelling actions through coded routines or the original orientation, one of the three key actions listed in plug-in’s own visual programming language. This visual Fig. 1. The programming block enables the user to rotate programming environment is not ideal for inexperienced the scan about the X-, Y-, and Z-axes separately in such CAD users to interact with when modelling, as it has its Table 1 A tabulated list of user needs and their description for the custom cushion scan-to-CAD object process, discovered through the trial by potential end users of the original Rhino scan-to-CAD process User Need Description The process is efficient Time to complete process does not exceed 15 min There are no unnecessary steps in the process The process is easy to use The process is easy to learn for all users regardless of their CAD skill level The user interface is easy to understand The user interface does not cause confusion The user interface accommodates users with a range of skill levels The controls are easy to use and locate Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 5 of 12 Fig. 2 Picture of the Grasshopper program that automates the conversion of a STL cushion or back support scan into a CAD-editable surface. There are seven action blocks in the Grasshopper program numbered in the picture, with descriptions summarising what each block does a way that any rotation is accounted for the other axes if part of the polyline is wider than the cushion surface, without the user needing to take action to ensure this. it will not project onto the surface and the Grasshop- Each axis of rotation is defined in Grasshopper as the per program will fail to move on to the next steps in the unit vector of its axis. The input of each rotation is con - program. The toggle should be switched to ON once nected to the user interface with sliding bars, enabling the position of the plane and polylines have been set as full rotation about each axis by the end user. needed. Block 2: Define and project edge curve Block 3: Split edge curve Once the correct orientation is set by the user, the scan Once the outline has been defined and projected onto the needs to be “baked” into the Rhinoceros modelling STL model, this outline must be divided parallel to the workspace, so that it can interact with both Rhino and X- and Y-axes to prepare the outline for final steps of the Grasshopper tools. Baking any model into Rhino also Grasshopper program. Block 3 divides the outline from allows for saving and exporting of the model. The user Block 2 into three segments using lines created by the must bake the oriented STL scan by toggling the switch user in the Rhino workspace – one line in the X-direction in the “Launch surface analysis” tab to ON. As soon as and one line in the Y-direction. This is necessary to define this occurs, block 2 of the Grasshopper program begins the surface in the Rhino modelling workspace, instead of its work. in tessellations used in the STL mesh format. Block 2 of the Grasshopper program is set to capture Block 3 is connected to the user interface tabs 5 and 6, the naked edges of the STL scan that has been baked into shown respectively in use in Figs. 6 and 7. Tab 5 asks the Rhino, turn the edges into a set of joined curves, and con- user to split the outline once by drawing a vertical line vert the curves into a polyline. This polyline is projected in the Top viewport using Rhino’s Line tool. The vertical up to a plane above the model surface and copied. The line must intersect the outline placed above the cushion; copy of the polyline can be stretched or shrunk in the once this happens, red x’s appear along the outline show- plane. The polyline copy is used to define a new outline of ing nodes along the outline curve. Tab 6 of the user inter- the CAD cushion surface, as shown in Fig. 5. face functions similarly to tab 5 but for a horizontal line The user interface has 3 interactive tools in this step: in the Top viewport that splits the outline without red x’s two numerical sliders and a toggle switch. The first in two. This horizontal line must intersect the outline, as slider is used to move the plane up or down so that it shown in Fig. 7. is above the cushion surface, while the second slider The newly split outline is now made up of 3 curves: one stretches or shrinks the polyline copy that will define in the general X-direction and two in a perpendicular the new outer edges of the future CAD surface. The user direction. These curves will be the first used in the rest of must set the second polyline such that it resides inside the Grasshopper program to build a contour map of the the visible boundaries of the top of the cushion surface; CAD surface from the STL cushion. Nace et al. 3D Printing in Medicine (2023) 9:3 Page 6 of 12 modelling workspace will make contours from the STL scan. Once the contour direction has been defined by the user, the slider in tab 7 of the user interface is used to define the number of contour lines made and the distance between them in centimetres. The number of contours and space between them are used in block 6 of the Grass- hopper scan-to-CAD program. Block 6 uses a Rhino soft- ware tool Surface from a network of curves that creates a surface using a network of intersecting curves. The tool requires curves in the network to intersect each other not more than once, and to have all curves in one direction intersecting all curves in the other direc- tion at least once. If these requirements are not met by the network of curves, a surface will not be created. To make the selection of curves useable in the Grasshopper program without requiring the user creating the surface to start the process over or conduct heavy computational editing, the Grasshopper scan-to-CAD program uses the slider tool in tab 7, shown in Fig. 8, to enable live defini - tion of the network of curves used in the Rhino tool; the contours made in block 5 are the curves used in block 6, along with the curves made in blocks 3 and 4 of the pro- gram. The user of the custom program is able to see when a surface can be created from the network of curves when a red surface overlaps the imported STL mesh, visible in Fig. 8. There is not one correct set of curves that will cre - ate a surface; thus the slider enables choice for the user. Fig. 3 The user interface created using Grasshopper. This user Once a surface can be made, the user can bake the CAD interface allows users to set some modelling parameters to convert surface into the Rhino modelling workspace by using a an STL cushion scan to a CAD-editable model through 8 steps toggle in tab 8 of the user interface (Fig. 3). Block 7: Creating a user interface for the process Block 4: Create contours in the Y‑direction The user interface was the last step in creating the auto - The next step in the Grasshopper program consists of mated process. Once the general steps that needed to be blocks 4 and 5 from Fig.  2, as they need to be concur- taken to create a CAD-editable surface from an STL scan rent to work correctly. Block 4 takes the three lines in the were programmed into the Grasshopper environment, Y-direction –the horizontal line drawn by the user and the Human UI Grasshopper plug-in was used to make an the two lines created from the split by the user-drawn interactive user interface that leads users through the pro- line—and rebuilds them so that they have an equal num- cess. Human UI makes it so that parametric input tools in ber of segments and control points, or nodes. The control Grasshopper, such as sliders and Boolean toggle switches, points are then projected down onto the STL scan from can instead be placed in a separate window or tab from the three horizontal lines, and these projected points are the Grasshopper environment and the parameters put in used to create NURBS curves along the shape of the scan, the user interface’s window are sent to the Grasshopper running in the Y-direction. This portion of the Grasshop - program live to update the model in the workspace. per process does not require input from the user to run. Workflow evaluation Blocks 5 and 6: Create contours and a surface from contours Once both a manual CAD process and the semi-auto- Block 5 creates NURBS curves running in the X-direc- mated process – referred to as the Grasshopper process tion that conform to the contours of the STL scan using from here on – were created, a tutorial was written for Rhino’s Contour tool. This block correlates to tab 7 in the the Grasshopper process [Additional file  2] similar to that user interface in which the user is asked to use Rhino’s written for the original CAD process [Additional file  1]. Point and Line tools to define in which direction the Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 7 of 12 Fig. 4 The correct orientation setting to begin the Grasshopper scan-to-CAD process. The Grasshopper program requires the user to orient the model such that the surface that conforms to the user can be seen in the XY-plane and that the front of the seat cushion or the bottom of the back support runs parallel to the Y-axis Fig. 5 A depiction of how the polyline defining the outer edge of the cushion surface should be set using the Grasshopper scan-to-CAD program and its user interface Nine volunteers were then recruited and tasked with tri- file 1]. The new Grasshopper scan-to-CAD program alling both processes by completing the tutorials for each requires 12 steps to complete, a 37% reduction in the process and recording the time it took for each volunteer number of total steps to convert an STL scan to a CAD to complete each process. Each volunteer self-reported model. The types of steps required in each process are their CAD experience level which were categorized using shown in Fig. 9, where the step type refers to what part this study’s experience level organisation method and are of the software the user interacts with in the step. The listed in Table 2 below. figure shows that the number of steps in which the user must work in the modelling workspace, highlighted in orange in Fig.  9, is reduced by 50% in the Grasshopper Results process compared to the original process. This reduc - A total of 19 steps to convert an STL mesh model of a tion suggests an easier process for users unfamiliar with CCS cushion or back support into a CAD model were CAD tools. documented in the original process [see Additonal Nace et al. 3D Printing in Medicine (2023) 9:3 Page 8 of 12 Fig. 6 Using the Grasshopper scan-to-CAD program at step 5, where a vertical (X-direction) line is drawn by the user in the Rhino modelling workspace at the front end of a seat cushion or the bottom end of a back support. The red x’s seen projected from the hovering outline to the STL cushion show that the line intersects the outline as needed by the Grasshopper program Fig. 7 Using the Grasshopper scan-to-CAD program at step 6, where a horizonal (Y-direction) line is drawn by the user in the Rhino modelling workspace across the center of the STL cushion. The red x’s seen projected from the outline to the STL cushion show that the horizontal line intersects the outline as needed for the Grasshopper program Results comparing the time taken for each volunteer volunteers (N = 9). This change equates to a 31% decrease to complete the two modelling processes are shown in in the time needed to complete the modelling process, Fig. 10 below. The figure shows a decrease in the time to regardless of CAD skill level. Additionally, within each complete trials as the CAD experience level of the user CAD experience group there was a decrease in the time increased, regardless of the modelling process. Although to complete the modelling process when following the this result is expected, it confirms that the modelling pro - Grasshopper method as compared to the original model- cess and tutorials are logical to engineering users. Results ling process. These results suggest that using the Grass - also indicated that the Grasshopper process decreased hopper method in the creation of custom wheelchair the time to convert an STL scan to a CAD surface from cushions would reduce the time needed to complete the an overall average of 29 min to an average of 20 min for all digital portion of the manufacturing process. Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 9 of 12 Fig. 8 Tab 7 in the Grasshopper scan-to-CAD program user interface, showing the slider that adjusts the number of contour lines defining the cushion surface. A surface created from the contours is shown in the red areas overlapping the yellow mesh surface Table 2 Description of the different categories of CAD software experience. Representative numbers for each CAD level were assigned to each category for easy reference in the study CAD Experience Level Experience Level Description Experience Level Volunteers Category Representative Number at this Level New user No experience with any CAD software 0 V1, V2, V3 Beginner user Learning one or more CAD software packages or last used CAD soft- 1 V4, V5, V6 ware more than 2 years before trial Intermediate user Uses a CAD software between twice a year and once a month 2 V7, V8, V9 Discussion these volunteers still indicate that the tutorials for both Results from both clinical users and the volunteer com- processes are sufficient to complete the modelling pro - parison trials of this study suggest that a manual process cess, though the Grasshopper method requires less time is valid though not the optimal method, especially for to complete regardless of CAD skill level. inexperienced and beginner CAD users. Feedback from The quantitative results support the positive feedback medium- to high-skill CAD users suggest that the Grass- on the Grasshopper process from clinical users. The hopper process is more accessible and easier to use than decrease in time to convert the STL to a CAD surface is the manual conversion process using only Rhino software similar to the percent decrease in the number of steps, and its modelling workspace. The scan-to-CAD process suggesting that the custom user interface in the Grass- cannot be completely automated due to the variance in hopper scan-to-CAD process is less confusing or less the shape of input scans, thus all users new to the process overwhelming than just interacting with the Rhino inter- will find tutorials on how to use the Grasshopper pro - face and its CAD tools. As with any skill, it is expected cess useful. Results from Li and Tanaka’s work on custom that the average time to complete the Grasshopper pro- wrist splint design suggest that users of all skill levels can cess would decrease further as users become more famil- learn a custom scan preparation process with sufficient iar with the method. The time decrease could enable training [12]. Those implementing such custom methods faster production per seating system and increased pro- into their practice will benefit from training their users, duction capacity for any clinical team producing custom and training only the use of the custom process is shorter seating, enabling further outreach to wheelchair users. than that required to be an expert Rhino or other CAD The user interface could be further optimised, software user. Training was not performed with the sec- for example by creating a plugin for Rhino using ond cohort of volunteers in this study; the results from C/C + + with its own unique and streamlined GUI, Nace et al. 3D Printing in Medicine (2023) 9:3 Page 10 of 12 Fig. 9 A comparison of the types of steps required in the original scan-to-CAD process in Rhino and the new process using Grasshopper. The type of each step is classified by what part of the software the user interacts with in the step Fig. 10 Time to complete each process trial, grouped and averaged (N = 3 per group) by CAD experience level of the volunteers, with standard deviation bars shown Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 11 of 12 Abbreviations though that is outside the scope of this study. This study CCS Custom-contoured seating was designed to demonstrate whether it is feasible for STL Standard tessellation language clinical teams without CAD experts or skilled pro- CAD Computer-aided design NURBS Non-uniform rational basis splines grammers to prepare STL scans for 3D printing or fur- GUI Graphic user interface ther CAD manipulation. The semi-automated process outlined here has been trialled with non-expert users Supplementary Information of CAD software and has recorded an average time to The online version contains supplementary material available at https:// doi. independent completion of 20.2  min, thereby demon- org/ 10. 1186/ s41205- 022- 00165-5. strating feasibility. The process outlined here to create a custom user interface and scan-to-CAD conversion Additional file 1. process can be used by other researchers, manufactur- Additional file 2. ers, and clinical teams as a baseline process to prepare Additional file 3. custom orthoses and seating for different manufactur - ing processes. Detailing the Grasshopper process here Acknowledgements allows for customisation and alteration of the method The authors would like to thank Bart Van der Velde for initial guidance on the potential use of the Rhinoceros 3D software in wheelchair seating design and as needed for other applications. As the Rhino software manufacturing. and Grasshopper plug-in are updated and new tools are created, the Grasshopper scan-to-CAD process can be Authors’ contributions All authors organized the protocol for the studies. SN conducted the studies changed to suit the desires and needs of users. and collected and analysed data from the studies. SN was a main contributor in the writing of the manuscript. All authors read, edited, and approved the final manuscript. Conclusion The goal of the modelling process in this study was to Funding convert an STL format scan of a custom moulded seat This research is supported by the Irish Research Council and Enable Ireland under the Irish Research Council’s Employment Based Postgraduate Pro- or back support into a CAD object, editable not only in gramme (Grant No.: irc72e6e373cd8ee4981332f32b5e9773be). Additionally, Rhino but in other CAD software packages if desired. the authors wish to acknowledge I-Form, funded by Science Foundation The manual workflow accomplishes this task but had Ireland (SFI) Grant Number 16/RC/3872. room for improvement for use in a clinical setting with Availability of data and materials users who may not be familiar with CAD software The datasets generated from the study are available upon reasonable request tools. As more clinical teams turn to different forms in an anonymized, quantitative form to protect the privacy of participants. Step-by-step guides used by participants in the study are supplemented as of scanning to make custom wheelchair seating, hav- Appendices to the article. The Rhinoceros 3D plug-in tool developed through ing tools to fit the needs of the team and the wheelchair the work in this study is available upon reasonable request. user is key to increasing use of high-end technology in clinical settings and orthosis manufacturing. Publish- Declarations ing details of processes such as the semi-automated Ethics approval and consent to participate process described herein further enables the use of new Need for ethics approval was waived and informed consent for participation technology by non-profits and low-budget teams. in the software trials were obtained from all participants. It also allows other teams the ability to custom- Consent for publication ise a digital workflow for processing scans for other Not applicable. 3D printed medical product design applications. Any process that requires an STL model to be parametri- Competing interests The authors declare that they have no competing interests. cally edited after a scan, such as upper and lower limb prosthetic design or moulds for custom dental devices, could take the building blocks of the Grasshopper tool Received: 1 November 2022 Accepted: 13 December 2022 created and tested in this paper and restructure them to suit the needs of the product. Devices in the medi- cal industry are already benefiting from moving to 3D References printing as a manufacturing method [14, 15]. To make 1. Nace S, Tiernan J, Ni Annaidh A. Manufacturing custom-contoured custom 3D printed products more widely available to wheelchair seating: a state of the art review. Prosthet Orthot Int. patients, creating and supplying software tools that 2019;43(4):382–95. 2. Tasker LH, Shapcott NG, Holland PM. The use and validation of a laser ease the pain of device design without majorly increas- scanner for computer aided design and manufacturing of wheelchair ing costs or design difficulty is key. seating. J Med Eng Technol. 2011;35(6–7):377–85. Nace et al. 3D Printing in Medicine (2023) 9:3 Page 12 of 12 3. Mitsouras D, Liacouras P, Imanzadeh A, Giannopoulos A, Cai T, Kum- amaru K, et al. Medical 3D Printing for the Radiologist. Radiographics. 2015;35:1965–88. 4. Fernandez-Vicente M, Escario Chust A, Conejero A, Fernandez M. Low cost digital fabrication approach for thumb orthoses. Rapid Prototyp J. 2017;23:6. 5. Rybicki F. The impact of regulation, reimbursement, and research on the value of 3D printing and other 3D procedures in medicine. 3D Print Med. 2022;8(1):6. 6. Mitsouras D, Liacouras P, Wake N, Rybicki F. RadioGraphics Update: Medi- cal 3D Printing for the Radiologist. Radiographics. 2020;40(4):E21–3. 7. Smith R. Development of a 3D printed, custom contoured back piece for wheelchair users. Dublin: University College Dublin; 2016. 8. Blaya F, D’amato RO, Pedro PS, Juanes JA, Lopez-Silva JA, Lagándara JG. Study, design and prototyping of arm splint with additive manufacturing process. ACM Int Conf Proc Series. 2017;2:Part F1322-7. 9. Baronio G, Harran S, Signoroni A. A critical analysis of a hand orthosis reverse engineering and 3D printing process. Appl Bionics Biomech. 2016;9(2016):1–7. 10. Palousek D, Rosicky J, Koutny D, Stoklásek P, Navrat T. Pilot study of the wrist orthosis design process. Rapid Prototyp J. 2014;20(1):27–32. 11. Paterson AM, Bibb RJ, Campbell RI. Evaluation of a digitised splinting approach with multiple-material functionality using Additive Manufacturing technologies. 23rd Annual International Solid Freeform Fabrication Sympo- sium, Austin, Texas, United States. 2012. 12. Li J, Tanaka H. Feasibility study applying a parametric model as the design generator for 3D–printed orthosis for fracture immobilization. 3D Print Med. 2018;4(1):1–15. 13. Li J, Tanaka H. Rapid customization system for 3D-printed splint using programmable modeling technique – a practical approach. 3D Print Med. 2018;4(1):5. 14. Kadakia RJ, Wixted CM, Kelly CN, Hanselman AE, Adams SB. From patient to procedure: the process of creating a custom 3D-printed medical device for foot and ankle pathology. Foot Ankle Spec. 2021;14(3):271–80. 15. Burnard JL, Parr WCH, Choy WJ, Walsh WR, Mobbs RJ. 3D-printed spine surgery implants: a systematic review of the efficacy and clinical safety profile of patient-specific and off-the-shelf devices. Eur Spine J. 2020;29(6):1248–60. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png 3D Printing in Medicine Springer Journals

Development and evaluation of a facile mesh-to-surface tool for customised wheelchair cushions

Loading next page...
 
/lp/springer-journals/development-and-evaluation-of-a-facile-mesh-to-surface-tool-for-kbUEmeGZlI

References (18)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2023
eISSN
2365-6271
DOI
10.1186/s41205-022-00165-5
Publisher site
See Article on Publisher Site

Abstract

Background Custom orthoses are becoming more commonly prescribed for upper and lower limbs. They require some form of shape-capture of the body parts they will be in contact with, which generates an STL file that design- ers prepare for manufacturing. For larger devices such as custom-contoured wheelchair cushions, the STL created during shape-capture can contain hundreds of thousands of tessellations, making them difficult to alter and prepare for manufacturing using mesh-editing software. This study covers the development and testing of a mesh-to-surface workflow in a parametric computer-aided design software using its visual programming language such that STL files of custom wheelchair cushions can be efficiently converted into a parametric single surface. Methods A volunteer in the clinical space with expertise in computer-aided design aided was interviewed to under- stand and document the current workflow for creating a single surface from an STL file of a custom wheelchair cush- ion. To understand the user needs of typical clinical workers with little computer-aided design experience, potential end-users of the process were tasked with completing the workflow and providing feedback during the experience. This feedback was used to automate part of the computer-aided design process using a visual programming tool, cre- ating a new semi-automated workflow for mesh-to-surface translation. Both the original and semi-automated process were then evaluated by nine volunteers with varying levels of computer-aided design experience. Results The semi-automated process showed a 37% reduction in the total number of steps required to convert an STL model to a parametric surface. Regardless of previous computer-aided design experience, volunteers completed the semi-automated workflow 31% faster on average than the manual workflow. Conclusions The creation of a semi-automated process for creating a single parametric surface of a custom wheel- chair cushion from an STL mesh makes mesh-to-surface conversion more efficient and more user-friendly to all, regardless of computer-aided design experience levels. The steps followed in this study may guide others in the development of their own mesh-to-surface tools in the wheelchair sector, as well as those creating other large cus- tom prosthetic devices. Keywords Wheelchair, Cushion, Grasshopper, Scan to solid, Mesh editing Introduction Wheelchairs enable autonomy in many individuals across *Correspondence: Aisling Ní Annaidh the world but sitting in a wheelchair for long periods can aisling.niannaidh@ucd.ie lead to pain, discomfort, and injuries such as pressure School of Mechanical and Materials Engineering, University College ulcers and spinal deformities. To prevent or correct such Dublin, Belfield, Dublin, Ireland SeatTech Posture and Mobility Services, Enable Ireland, Dublin, Ireland issues, customized wheelchair seating is manufactured for individual needs. Custom contoured seating (CCS) © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Nace et al. 3D Printing in Medicine (2023) 9:3 Page 2 of 12 refers to customized seating that is based on an impres- While the software is part of a cleared medical device, to sion taken of a wheelchair user’s body and matches date all such clearances have been for anatomic models exactly to that shape. Custom wheelchair seating systems and anatomic guides for parts 3D printed from CT, MR, require shape capture of a surface moulded to the shape and ultrasound images [6]. Such a software tool is not of the wheelchair user to design an effective personal - necessarily usable for other products due to this system- ised postural aid and pressure relieving intervention [1]. based method, but it suggests the need for further devel- Methods to capture shapes include contact digitisers and opment 3D printing software packages and a broadening non-contact scanning, including high-end time-of-flight of their potential approved use cases to match the wide or triangulation laser scanners as well as more affordable range of products that can be made using 3D printing. structured-light scanners [2]. To widen the range of potential products, we can look Patient specific 3D printing typically begins with CT to CAD software packages which enable users to design and MR images that are then segmented and converted and prepare models for manufacturing using direct edit- to printable files [3]. However, there are a large portfolio ing or parametric modelling instead of mesh-only edit- of devices customized to a patient for which the Stand- ing software. Some CAD software offer mesh editing ard Tessellation File (STL) file is created via alternative tools similar to the mesh-specific software, and more and strategies. These scanning techniques include laser trigo - more CAD software producers are attempting to provide nometric triangulation-based scanners and structured a means to convert mesh objects to non-uniform rational light scanners which both generate the digital recreation basis splines (NURBS) objects that are editable within a of an object or surface as a point cloud or a mesh object. CAD environment. The NURBS object produced from Mesh objects are made up of tessellations that, when such one-click conversions have as many surfaces as imported into a computer-aided design (CAD) program, the mesh object had tessellations; for 3D scans, this can are not editable because they are recognised as either a mean an object with tens of thousands of small surfaces single entity or as thousands of small, individual surfaces. that have to be manipulated when preparing the model Because of this, if a 3D scan needs to be edited before for manufacturing, requiring large amounts of computer manufacturing – to make a solid cushion or back support power [7]. A NURBS object with hundreds or thousands with the scanned surface, for example – the scan must be of surfaces is not a user-friendly object for those unfamil- altered in a mesh-specific software or ‘converted’ into a iar with CAD, such as the clinical teams prescribing and CAD model and edited in a CAD environment. Mesh- producing custom prosthetics, orthoses, and wheelchair specific software, like Meshmixer (Autodesk) or Mesh - seating. Lab (ISTI-CNR), offer sculpting as well as refinement and This leads to a need for other ways to convert meshes smoothing tools to alter the shape of a mesh object itera- to CAD models. Recent research has attempted to fill tively. Fernandez-Vicente et al. [4] developed an inexpen- this void, specifically for custom prostheses and orthoses sive, efficient method for scanning, designing, and 3D designed from body scans. Blaya et  al. [8], Fernandez- printing a thumb orthosis using only Meshmixer as the Vicente et  al. [4], Baronio et  al. [9], Palousek et  al. [10], digital model creation tool. The new design and fabrica - and Paterson et  al. [11] developed five different digital tion method suggested digitisation and 3D printing can workflows for designing custom wrist orthoses or splints cut costs and production time for orthosis producers. for3D printing. Works by Blaya et  al. [8], Fernandez- However, this process was not tested with clinical spe- Vicente et  al. [4], and Paterson et  al. [11] suggest that cialists, who may be less familiar with digital design tools digital mesh models generated from 3D scans of wrists and require more time and training to create an appro- generate fewer tessellations than those of moulded cush- priate orthosis. Furthermore, mesh editing programs ions due to the size of wrist scans. This difference may like Meshmixer do not offer parametric editing tools make using mesh editing tools a less taxing task for wrist which would be used to alter measurable dimensions of orthosis model generation than for CCS. However, these a wheelchair cushion, such as the height or depth of the investigations did not evaluate the efficiency or user- cushion for a better fit in a wheelchair base. Thus, for friendliness of the digital workflows, nor did they evalu - designing medical devices like wheelchair cushions that ate the digital workflows with clinical experts to gauge require more than the scanned shape in the model crea- the viability of the processes in a clinical setting. tion process [1], they do not meet user needs for digital Baronio et  al. [9] developed a digital workflow that alterations prior to manufacturing. used both mesh editing software and NURBS conver- There are 3D printing software packages cleared by sion for CAD editing. This study deemed CAD editing the United States Food and Drug Administration (FDA) necessary in their digital workflow due to the volumet - for specific intended uses as part of a system that must ric modelling needed to create an effective hand ortho - include a specific 3D printer, material, and anatomy [5]. sis that was comfortable, light, and well-designed for its Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 3 of 12 function. The authors discussed in their study how more 6) and is documented in a step-by-step format [see Addi- user-friendly digital editing tools are needed in the clini- tional file  1]. Briefly, three key actions were identified in cal environment to use scanning and 3D printing to make the process: orienting the scan in the modelling space; custom orthoses, otherwise a CAD specialist is needed defining the STL surface with a set of curves; and using on the clinical team, implying that this may be a barrier the curves to create a new CAD surface in the shape of to adoption for clinical teams. In their 2014 study on the the scan. These actions are shown in process order in design and 3D printing of custom hand-wrist orthoses in Fig. 1. a clinical setting, Palousek et al. [10] similarly concluded User needs trial that the digital workflow would require a CAD specialist Initial development work focussed on identifying user or the automation of the workflow tasks that required the needs of the clinical team. For this reason, the original CAD specialist. manual step-by-step process was provided to two poten- In 2018, Li and Tanaka [12] created a process that auto- tial clinical end-users to complete as a tutorial. The two mated some of the CAD modelling tasks required in the clinical users (volunteers C1 and C2) had varying levels digital design of custom wrist orthoses using Rhinoc- of CAD experience and worked in the wheelchair seat- eros 3D and its visual programming plugin, Grasshop- ing industry at the time that they completed the tutorials. per (Rhinoceros 3D, McNeel & Associates). This study Each volunteer gave live feedback during monitored trials found the new workflow required 2 to 3 min of computer of the manual process to capture user needs of the clini- work compared to 20 min to 3 h of CAD modelling in the cal team. The conversation during each test was open, to traditional digital splint process [13]. The authors show allow for any constructive feedback and each volunteer evidence in their study that the automated tasks created was asked a mixture of opened-ended questions [see using Grasshopper could be used in the design of wear- Additional file 3]. able orthotic devices and prostheses. In a separate study, Volunteer C1, a beginner CAD user, took approxi- Li and Tanaka [12] trialled their new process with five mately 21  min to complete the guide and had difficulty nursing students unfamiliar with CAD software. Their locating some tools and with using the ‘gumball’ tool. study indicated that an automated process with a sim- Specifically volunteer C1 could not locate the ‘line tool’ plified software interface enabled the nursing students and the ‘surface from the network or curves’ tool. Other to learn how to correctly use the process to design hand issues highlighted included (i) making the gridlines that orthoses from 3D scans in less than one workday. would be projected onto the STL scan and (ii) selecting In this manuscript, a post-processing method to con- multiple lines in the modelling workspace. For (i), the vert scans of custom back support and seat cushion volunteer did not know how many lines should be pre- surfaces into 3D CAD models was developed using Rhi- sent; for (ii) the volunteer could not select multiple lines. noceros 3D’s Grasshopper plugin and trialled by poten- tial clinical users of the process. Feedback from the trials was used to prototype a more user-friendly interface to the CAD environment and to simplify the conversion process further to make it more accessible to those with little to no CAD experience. The process used herein was based on Li and Tanaka’s work on the digital design of wrist orthoses in clinical settings [12]. The manuscript is divided into three main tasks: the documentation and assessment of the manual method for CCS design; the development of an automated workflow; and the evalua - tion of the proposed automated workflow. Methods Capture of existing manual workflow The first step in the development of an improved scan- to-print workflow was the capture and documentation of the existing process implemented by an expert clini- cal CAD user (volunteer C0). The researchers recorded the manual CAD method implemented by volunteer C0 Fig. 1 The three key actions in the workflow of converting an STL of in the prescription and manufacture of CCS systems. a custom cushion scan into a CAD surface in Rhinoceros software The original method employed Rhinoceros 3D (Version Nace et al. 3D Printing in Medicine (2023) 9:3 Page 4 of 12 Upon completing the trial, volunteer C1 described the own set of skills to learn beyond the traditional CAD process as too long and suggested that the process should environment. However, Grasshopper has methods of cre- be as automated as possible to simplify the process. They ating separate Graphic User Interfaces (GUIs) with which stated the process should take no longer than 15  min to users of the routine can interact instead of the visual pro- complete. gramming environment of Grasshopper. To coincide with Volunteer C2, who had no previous experience with the user-need of “ease of use” in this project, the authors any CAD tools, took approximately 38  min to complete decided to use Grasshopper’s tools to create a simple user the guide. Volunteer C2 struggled to understand the dif- interface for the automated workflow. ferent viewpoints in the Rhinoceros software and had To create the user interface within Grasshopper, the issues creating the gridlines. The volunteer described the modelling workflow first had to be translated into Grass - process as “time-consuming” and stated that the ideal hopper’s environment. The steps taken in the manual process should take less than 30 min. They suggested that workflow were translated into Grasshopper’s visual pro - a dialogue box on screen identifying the next steps would gramming language. Figure 2 shows the entire Grasshop- be helpful. per program, broken down into action blocks 1 through The key user needs identified from feedback during 7. Blocks 1 through 6 are directly related to making a the trial are listed in Table  1. These (clinical) user needs CAD-editable model from the STL scan and block 7 cre- were used to tailor the proposed automated process ates the user interface that allows the user to set parame- for the anticipated future users. Analysis of these user ters of the model in blocks 1 through 6 without changing needs led to the determination of the following design the Grasshopper file. The user interface is shown in Fig.  3. requirements: The function of blocks 1 through 6 of the scan-to-CAD Grasshopper code, and how the user interface employs 1. Automate as many steps as possible, including the them, are further described in the following sections. A orientation and creation of the surface step and the step-by-step guide for using the Grasshopper scan-to- gridlines step. CAD process with the custom user interface tool in Fig. 3 2. Customise a toolbar or user interface to remove can be found in Additional file 2. tools in Rhino’s interface that are not required and decrease overall time requirement. Block 1: Orienting the scan using Grasshopper The orientation of the STL scan in a 3D workspace dic - tates how it is translated into a 3D CAD model. The Grasshopper program requires the user to orient the Semi‑automated workflow development model such that the surface that conforms to the user This section details the development of a semi-auto - can be seen in the XY-plane and that the front of the seat mated workflow that utilizes Grasshopper in Rhinoc - cushion or the bottom of the back support runs parallel eros 3D with the goal of satisfying the defined user needs to the Y-axis, as shown in Fig. 4. gathered from the trial with clinical users. Grasshopper Block 1 enables the user to orient the STL cushion scan is a plug-in for Rhinoceros that enables the automation in the Rhinoceros modelling space regardless of the scan’s of many modelling actions through coded routines or the original orientation, one of the three key actions listed in plug-in’s own visual programming language. This visual Fig. 1. The programming block enables the user to rotate programming environment is not ideal for inexperienced the scan about the X-, Y-, and Z-axes separately in such CAD users to interact with when modelling, as it has its Table 1 A tabulated list of user needs and their description for the custom cushion scan-to-CAD object process, discovered through the trial by potential end users of the original Rhino scan-to-CAD process User Need Description The process is efficient Time to complete process does not exceed 15 min There are no unnecessary steps in the process The process is easy to use The process is easy to learn for all users regardless of their CAD skill level The user interface is easy to understand The user interface does not cause confusion The user interface accommodates users with a range of skill levels The controls are easy to use and locate Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 5 of 12 Fig. 2 Picture of the Grasshopper program that automates the conversion of a STL cushion or back support scan into a CAD-editable surface. There are seven action blocks in the Grasshopper program numbered in the picture, with descriptions summarising what each block does a way that any rotation is accounted for the other axes if part of the polyline is wider than the cushion surface, without the user needing to take action to ensure this. it will not project onto the surface and the Grasshop- Each axis of rotation is defined in Grasshopper as the per program will fail to move on to the next steps in the unit vector of its axis. The input of each rotation is con - program. The toggle should be switched to ON once nected to the user interface with sliding bars, enabling the position of the plane and polylines have been set as full rotation about each axis by the end user. needed. Block 2: Define and project edge curve Block 3: Split edge curve Once the correct orientation is set by the user, the scan Once the outline has been defined and projected onto the needs to be “baked” into the Rhinoceros modelling STL model, this outline must be divided parallel to the workspace, so that it can interact with both Rhino and X- and Y-axes to prepare the outline for final steps of the Grasshopper tools. Baking any model into Rhino also Grasshopper program. Block 3 divides the outline from allows for saving and exporting of the model. The user Block 2 into three segments using lines created by the must bake the oriented STL scan by toggling the switch user in the Rhino workspace – one line in the X-direction in the “Launch surface analysis” tab to ON. As soon as and one line in the Y-direction. This is necessary to define this occurs, block 2 of the Grasshopper program begins the surface in the Rhino modelling workspace, instead of its work. in tessellations used in the STL mesh format. Block 2 of the Grasshopper program is set to capture Block 3 is connected to the user interface tabs 5 and 6, the naked edges of the STL scan that has been baked into shown respectively in use in Figs. 6 and 7. Tab 5 asks the Rhino, turn the edges into a set of joined curves, and con- user to split the outline once by drawing a vertical line vert the curves into a polyline. This polyline is projected in the Top viewport using Rhino’s Line tool. The vertical up to a plane above the model surface and copied. The line must intersect the outline placed above the cushion; copy of the polyline can be stretched or shrunk in the once this happens, red x’s appear along the outline show- plane. The polyline copy is used to define a new outline of ing nodes along the outline curve. Tab 6 of the user inter- the CAD cushion surface, as shown in Fig. 5. face functions similarly to tab 5 but for a horizontal line The user interface has 3 interactive tools in this step: in the Top viewport that splits the outline without red x’s two numerical sliders and a toggle switch. The first in two. This horizontal line must intersect the outline, as slider is used to move the plane up or down so that it shown in Fig. 7. is above the cushion surface, while the second slider The newly split outline is now made up of 3 curves: one stretches or shrinks the polyline copy that will define in the general X-direction and two in a perpendicular the new outer edges of the future CAD surface. The user direction. These curves will be the first used in the rest of must set the second polyline such that it resides inside the Grasshopper program to build a contour map of the the visible boundaries of the top of the cushion surface; CAD surface from the STL cushion. Nace et al. 3D Printing in Medicine (2023) 9:3 Page 6 of 12 modelling workspace will make contours from the STL scan. Once the contour direction has been defined by the user, the slider in tab 7 of the user interface is used to define the number of contour lines made and the distance between them in centimetres. The number of contours and space between them are used in block 6 of the Grass- hopper scan-to-CAD program. Block 6 uses a Rhino soft- ware tool Surface from a network of curves that creates a surface using a network of intersecting curves. The tool requires curves in the network to intersect each other not more than once, and to have all curves in one direction intersecting all curves in the other direc- tion at least once. If these requirements are not met by the network of curves, a surface will not be created. To make the selection of curves useable in the Grasshopper program without requiring the user creating the surface to start the process over or conduct heavy computational editing, the Grasshopper scan-to-CAD program uses the slider tool in tab 7, shown in Fig. 8, to enable live defini - tion of the network of curves used in the Rhino tool; the contours made in block 5 are the curves used in block 6, along with the curves made in blocks 3 and 4 of the pro- gram. The user of the custom program is able to see when a surface can be created from the network of curves when a red surface overlaps the imported STL mesh, visible in Fig. 8. There is not one correct set of curves that will cre - ate a surface; thus the slider enables choice for the user. Fig. 3 The user interface created using Grasshopper. This user Once a surface can be made, the user can bake the CAD interface allows users to set some modelling parameters to convert surface into the Rhino modelling workspace by using a an STL cushion scan to a CAD-editable model through 8 steps toggle in tab 8 of the user interface (Fig. 3). Block 7: Creating a user interface for the process Block 4: Create contours in the Y‑direction The user interface was the last step in creating the auto - The next step in the Grasshopper program consists of mated process. Once the general steps that needed to be blocks 4 and 5 from Fig.  2, as they need to be concur- taken to create a CAD-editable surface from an STL scan rent to work correctly. Block 4 takes the three lines in the were programmed into the Grasshopper environment, Y-direction –the horizontal line drawn by the user and the Human UI Grasshopper plug-in was used to make an the two lines created from the split by the user-drawn interactive user interface that leads users through the pro- line—and rebuilds them so that they have an equal num- cess. Human UI makes it so that parametric input tools in ber of segments and control points, or nodes. The control Grasshopper, such as sliders and Boolean toggle switches, points are then projected down onto the STL scan from can instead be placed in a separate window or tab from the three horizontal lines, and these projected points are the Grasshopper environment and the parameters put in used to create NURBS curves along the shape of the scan, the user interface’s window are sent to the Grasshopper running in the Y-direction. This portion of the Grasshop - program live to update the model in the workspace. per process does not require input from the user to run. Workflow evaluation Blocks 5 and 6: Create contours and a surface from contours Once both a manual CAD process and the semi-auto- Block 5 creates NURBS curves running in the X-direc- mated process – referred to as the Grasshopper process tion that conform to the contours of the STL scan using from here on – were created, a tutorial was written for Rhino’s Contour tool. This block correlates to tab 7 in the the Grasshopper process [Additional file  2] similar to that user interface in which the user is asked to use Rhino’s written for the original CAD process [Additional file  1]. Point and Line tools to define in which direction the Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 7 of 12 Fig. 4 The correct orientation setting to begin the Grasshopper scan-to-CAD process. The Grasshopper program requires the user to orient the model such that the surface that conforms to the user can be seen in the XY-plane and that the front of the seat cushion or the bottom of the back support runs parallel to the Y-axis Fig. 5 A depiction of how the polyline defining the outer edge of the cushion surface should be set using the Grasshopper scan-to-CAD program and its user interface Nine volunteers were then recruited and tasked with tri- file 1]. The new Grasshopper scan-to-CAD program alling both processes by completing the tutorials for each requires 12 steps to complete, a 37% reduction in the process and recording the time it took for each volunteer number of total steps to convert an STL scan to a CAD to complete each process. Each volunteer self-reported model. The types of steps required in each process are their CAD experience level which were categorized using shown in Fig. 9, where the step type refers to what part this study’s experience level organisation method and are of the software the user interacts with in the step. The listed in Table 2 below. figure shows that the number of steps in which the user must work in the modelling workspace, highlighted in orange in Fig.  9, is reduced by 50% in the Grasshopper Results process compared to the original process. This reduc - A total of 19 steps to convert an STL mesh model of a tion suggests an easier process for users unfamiliar with CCS cushion or back support into a CAD model were CAD tools. documented in the original process [see Additonal Nace et al. 3D Printing in Medicine (2023) 9:3 Page 8 of 12 Fig. 6 Using the Grasshopper scan-to-CAD program at step 5, where a vertical (X-direction) line is drawn by the user in the Rhino modelling workspace at the front end of a seat cushion or the bottom end of a back support. The red x’s seen projected from the hovering outline to the STL cushion show that the line intersects the outline as needed by the Grasshopper program Fig. 7 Using the Grasshopper scan-to-CAD program at step 6, where a horizonal (Y-direction) line is drawn by the user in the Rhino modelling workspace across the center of the STL cushion. The red x’s seen projected from the outline to the STL cushion show that the horizontal line intersects the outline as needed for the Grasshopper program Results comparing the time taken for each volunteer volunteers (N = 9). This change equates to a 31% decrease to complete the two modelling processes are shown in in the time needed to complete the modelling process, Fig. 10 below. The figure shows a decrease in the time to regardless of CAD skill level. Additionally, within each complete trials as the CAD experience level of the user CAD experience group there was a decrease in the time increased, regardless of the modelling process. Although to complete the modelling process when following the this result is expected, it confirms that the modelling pro - Grasshopper method as compared to the original model- cess and tutorials are logical to engineering users. Results ling process. These results suggest that using the Grass - also indicated that the Grasshopper process decreased hopper method in the creation of custom wheelchair the time to convert an STL scan to a CAD surface from cushions would reduce the time needed to complete the an overall average of 29 min to an average of 20 min for all digital portion of the manufacturing process. Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 9 of 12 Fig. 8 Tab 7 in the Grasshopper scan-to-CAD program user interface, showing the slider that adjusts the number of contour lines defining the cushion surface. A surface created from the contours is shown in the red areas overlapping the yellow mesh surface Table 2 Description of the different categories of CAD software experience. Representative numbers for each CAD level were assigned to each category for easy reference in the study CAD Experience Level Experience Level Description Experience Level Volunteers Category Representative Number at this Level New user No experience with any CAD software 0 V1, V2, V3 Beginner user Learning one or more CAD software packages or last used CAD soft- 1 V4, V5, V6 ware more than 2 years before trial Intermediate user Uses a CAD software between twice a year and once a month 2 V7, V8, V9 Discussion these volunteers still indicate that the tutorials for both Results from both clinical users and the volunteer com- processes are sufficient to complete the modelling pro - parison trials of this study suggest that a manual process cess, though the Grasshopper method requires less time is valid though not the optimal method, especially for to complete regardless of CAD skill level. inexperienced and beginner CAD users. Feedback from The quantitative results support the positive feedback medium- to high-skill CAD users suggest that the Grass- on the Grasshopper process from clinical users. The hopper process is more accessible and easier to use than decrease in time to convert the STL to a CAD surface is the manual conversion process using only Rhino software similar to the percent decrease in the number of steps, and its modelling workspace. The scan-to-CAD process suggesting that the custom user interface in the Grass- cannot be completely automated due to the variance in hopper scan-to-CAD process is less confusing or less the shape of input scans, thus all users new to the process overwhelming than just interacting with the Rhino inter- will find tutorials on how to use the Grasshopper pro - face and its CAD tools. As with any skill, it is expected cess useful. Results from Li and Tanaka’s work on custom that the average time to complete the Grasshopper pro- wrist splint design suggest that users of all skill levels can cess would decrease further as users become more famil- learn a custom scan preparation process with sufficient iar with the method. The time decrease could enable training [12]. Those implementing such custom methods faster production per seating system and increased pro- into their practice will benefit from training their users, duction capacity for any clinical team producing custom and training only the use of the custom process is shorter seating, enabling further outreach to wheelchair users. than that required to be an expert Rhino or other CAD The user interface could be further optimised, software user. Training was not performed with the sec- for example by creating a plugin for Rhino using ond cohort of volunteers in this study; the results from C/C + + with its own unique and streamlined GUI, Nace et al. 3D Printing in Medicine (2023) 9:3 Page 10 of 12 Fig. 9 A comparison of the types of steps required in the original scan-to-CAD process in Rhino and the new process using Grasshopper. The type of each step is classified by what part of the software the user interacts with in the step Fig. 10 Time to complete each process trial, grouped and averaged (N = 3 per group) by CAD experience level of the volunteers, with standard deviation bars shown Nac e et al. 3D Printing in Medicine (2023) 9:3 Page 11 of 12 Abbreviations though that is outside the scope of this study. This study CCS Custom-contoured seating was designed to demonstrate whether it is feasible for STL Standard tessellation language clinical teams without CAD experts or skilled pro- CAD Computer-aided design NURBS Non-uniform rational basis splines grammers to prepare STL scans for 3D printing or fur- GUI Graphic user interface ther CAD manipulation. The semi-automated process outlined here has been trialled with non-expert users Supplementary Information of CAD software and has recorded an average time to The online version contains supplementary material available at https:// doi. independent completion of 20.2  min, thereby demon- org/ 10. 1186/ s41205- 022- 00165-5. strating feasibility. The process outlined here to create a custom user interface and scan-to-CAD conversion Additional file 1. process can be used by other researchers, manufactur- Additional file 2. ers, and clinical teams as a baseline process to prepare Additional file 3. custom orthoses and seating for different manufactur - ing processes. Detailing the Grasshopper process here Acknowledgements allows for customisation and alteration of the method The authors would like to thank Bart Van der Velde for initial guidance on the potential use of the Rhinoceros 3D software in wheelchair seating design and as needed for other applications. As the Rhino software manufacturing. and Grasshopper plug-in are updated and new tools are created, the Grasshopper scan-to-CAD process can be Authors’ contributions All authors organized the protocol for the studies. SN conducted the studies changed to suit the desires and needs of users. and collected and analysed data from the studies. SN was a main contributor in the writing of the manuscript. All authors read, edited, and approved the final manuscript. Conclusion The goal of the modelling process in this study was to Funding convert an STL format scan of a custom moulded seat This research is supported by the Irish Research Council and Enable Ireland under the Irish Research Council’s Employment Based Postgraduate Pro- or back support into a CAD object, editable not only in gramme (Grant No.: irc72e6e373cd8ee4981332f32b5e9773be). Additionally, Rhino but in other CAD software packages if desired. the authors wish to acknowledge I-Form, funded by Science Foundation The manual workflow accomplishes this task but had Ireland (SFI) Grant Number 16/RC/3872. room for improvement for use in a clinical setting with Availability of data and materials users who may not be familiar with CAD software The datasets generated from the study are available upon reasonable request tools. As more clinical teams turn to different forms in an anonymized, quantitative form to protect the privacy of participants. Step-by-step guides used by participants in the study are supplemented as of scanning to make custom wheelchair seating, hav- Appendices to the article. The Rhinoceros 3D plug-in tool developed through ing tools to fit the needs of the team and the wheelchair the work in this study is available upon reasonable request. user is key to increasing use of high-end technology in clinical settings and orthosis manufacturing. Publish- Declarations ing details of processes such as the semi-automated Ethics approval and consent to participate process described herein further enables the use of new Need for ethics approval was waived and informed consent for participation technology by non-profits and low-budget teams. in the software trials were obtained from all participants. It also allows other teams the ability to custom- Consent for publication ise a digital workflow for processing scans for other Not applicable. 3D printed medical product design applications. Any process that requires an STL model to be parametri- Competing interests The authors declare that they have no competing interests. cally edited after a scan, such as upper and lower limb prosthetic design or moulds for custom dental devices, could take the building blocks of the Grasshopper tool Received: 1 November 2022 Accepted: 13 December 2022 created and tested in this paper and restructure them to suit the needs of the product. Devices in the medi- cal industry are already benefiting from moving to 3D References printing as a manufacturing method [14, 15]. To make 1. Nace S, Tiernan J, Ni Annaidh A. Manufacturing custom-contoured custom 3D printed products more widely available to wheelchair seating: a state of the art review. Prosthet Orthot Int. patients, creating and supplying software tools that 2019;43(4):382–95. 2. Tasker LH, Shapcott NG, Holland PM. The use and validation of a laser ease the pain of device design without majorly increas- scanner for computer aided design and manufacturing of wheelchair ing costs or design difficulty is key. seating. J Med Eng Technol. 2011;35(6–7):377–85. Nace et al. 3D Printing in Medicine (2023) 9:3 Page 12 of 12 3. Mitsouras D, Liacouras P, Imanzadeh A, Giannopoulos A, Cai T, Kum- amaru K, et al. Medical 3D Printing for the Radiologist. Radiographics. 2015;35:1965–88. 4. Fernandez-Vicente M, Escario Chust A, Conejero A, Fernandez M. Low cost digital fabrication approach for thumb orthoses. Rapid Prototyp J. 2017;23:6. 5. Rybicki F. The impact of regulation, reimbursement, and research on the value of 3D printing and other 3D procedures in medicine. 3D Print Med. 2022;8(1):6. 6. Mitsouras D, Liacouras P, Wake N, Rybicki F. RadioGraphics Update: Medi- cal 3D Printing for the Radiologist. Radiographics. 2020;40(4):E21–3. 7. Smith R. Development of a 3D printed, custom contoured back piece for wheelchair users. Dublin: University College Dublin; 2016. 8. Blaya F, D’amato RO, Pedro PS, Juanes JA, Lopez-Silva JA, Lagándara JG. Study, design and prototyping of arm splint with additive manufacturing process. ACM Int Conf Proc Series. 2017;2:Part F1322-7. 9. Baronio G, Harran S, Signoroni A. A critical analysis of a hand orthosis reverse engineering and 3D printing process. Appl Bionics Biomech. 2016;9(2016):1–7. 10. Palousek D, Rosicky J, Koutny D, Stoklásek P, Navrat T. Pilot study of the wrist orthosis design process. Rapid Prototyp J. 2014;20(1):27–32. 11. Paterson AM, Bibb RJ, Campbell RI. Evaluation of a digitised splinting approach with multiple-material functionality using Additive Manufacturing technologies. 23rd Annual International Solid Freeform Fabrication Sympo- sium, Austin, Texas, United States. 2012. 12. Li J, Tanaka H. Feasibility study applying a parametric model as the design generator for 3D–printed orthosis for fracture immobilization. 3D Print Med. 2018;4(1):1–15. 13. Li J, Tanaka H. Rapid customization system for 3D-printed splint using programmable modeling technique – a practical approach. 3D Print Med. 2018;4(1):5. 14. Kadakia RJ, Wixted CM, Kelly CN, Hanselman AE, Adams SB. From patient to procedure: the process of creating a custom 3D-printed medical device for foot and ankle pathology. Foot Ankle Spec. 2021;14(3):271–80. 15. Burnard JL, Parr WCH, Choy WJ, Walsh WR, Mobbs RJ. 3D-printed spine surgery implants: a systematic review of the efficacy and clinical safety profile of patient-specific and off-the-shelf devices. Eur Spine J. 2020;29(6):1248–60. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

3D Printing in MedicineSpringer Journals

Published: Feb 13, 2023

Keywords: Wheelchair; Cushion; Grasshopper; Scan to solid; Mesh editing

There are no references for this article.