[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-649003-105":59,"doc-detail-649003-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","data-augmentation-of-engineering-drawings-for-data-driven-component-segmentation","Data Augmentation of Engineering Drawings for Data-driven Component Segmentation","","Data-driven component segmentation of raster engineering drawings is targeted through an automatic pipeline that improves inspection efficiency. The approach vectorizes B&W raster drawings and uses a synthesis method to construct a large labelled dataset from existing DXF examples. It trains a model to predict the type of each vectorized component and deploys on raster inputs. A key requirement is generating arbitrarily large synthetic drawings while enforcing validity of engineering technical rules, avoiding invalid data under domain constraints.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/data-augmentation-of-engineering-drawings-for-data-driven-component-segmentation/649003/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/data-augmentation-of-engineering-drawings-for-data-driven-component-segmentation/649003.png","ImageObject",300,407,{"name":92,"@type":93},"Finn","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-25","2026-10-23",true,{"@type":102,"interactionType":103,"userInteractionCount":8},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What is the ultimate goal of the proposed system?","Question",{"text":112,"@type":113},"The system aims at data-driven component segmentation of raster engineering drawings, enabling automated analysis of all components rather than partial interpretation for a specific scenario.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How does the pipeline prepare training data?",{"text":117,"@type":113},"It uses vectorized drawings (DXF) as a source, synthesizes additional drawings, converts them to raster images, and then trains a model to perform vectorized segmentation and component type prediction.",{"name":119,"@type":110,"acceptedAnswer":120},"Why is data synthesis required to follow technical validity rules?",{"text":121,"@type":113},"Because many generic image augmentations would produce invalid engineering-drawing outputs, the generated synthetic drawings must be checked against technical rules to ensure they remain valid for domain learning.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},649003,1792936774,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":8,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145},34359740700684,"https://ap-avatar.wpscdn.com/avatar/1f400023980c374ae676?_k=1777273430885731487","Data Augmentation of Engineering Drawings for Data-driven Component Segmentation  \nWentai Zhang, Quan Chen, Can Koz, Louise Xie, Amit Regmi, Soji Yamakawa, Tomotake Furuhata, Kenji Shimada, Levent Burak Kara  \nCarnegie Mellon University 1  \nOverview  \nUltimate goal: Data-driven component segmentation of raster drawings  \nAn automatic system  \nB&W raster drawing Vectocontozres,ddsegmentaimensionionandamontextsg  \n2   \nMotivation: Industrial Part Quotation Systems  \nDrawing Format When an order is placed  \n16%  \n84%  \nA survey on the project and issues in Japan’s manufacturing industry, 2017  \nAverage time for a part quote: 7 days  \n 3   \nMotivation: Industrial Part Quotation Systems  \nNotation  \nPart drawings Metadata  \nThis step is time-consuming and dull. We aim at building an automatic system to aid the inspector.  \n 4   \nLiterature Review: Engineering Drawing Analysis  \nPrior works mainly focus on pattern recognition, shape identification and drawing retrieval.  \nDiagram recognition in floor plans, flow charts and electric circuit diagrams and vibratory mechanical systems [Delalandre [et. al](et. al), 2010, [Kara et. al. 2008](Kara et. al. 2008), [Schafer et. al. 2021](Schafer et. al. 2021)]  \nShape identification for part drawings with sampled points, histograms or shape descriptors. [[Liu et. al. 2009](Liu et. al. 2009), [Huet et. al. 2001](Huet et. al. 2001)]  \nDrawing retrieval or matching using lines, pixel blocks or patches.[[Mednonogov et. al. 2000](Mednonogov et. al. 2000), [Jiao et. al](Jiao et. al). 2009, [Sousa et. al. 2010](Sousa et. al. 2010)]  \nOnly achieve a partial interpretation of the drawings for a specific scenario. We aim at developing a data-driven system to analyze all the components for general analysis.  \n 5   \nOverview  \nUltimate goal: Data-driven component segmentation of raster drawings  \nAn automatic system with:  \n1. Avectorization preprocessing  \n2. A synthesis method to automatically construct a large labelled dataset  \n3. A data-driven model that predicts the type of each vectorized component  \nraster drawing  \nB&W  \nVectorized segmentation  \n6   \nRequirements  \nA data synthesis method that:  \n(1) Utilize the information stored in existing labelled examples  \n(2) Generate an arbitrarily large set of synthetic drawings to train a data-driven model for binary component segmentation (contour shape/dimension set)  \n(3) The generated drawings are subjected to validity of technical rules  \n7  \nRelated work  \nGeneral Data Augmentation Methods in CV  \nSimulated Data Augmentation  \nBasic geometric transformation,  \nfiltering, random erasing and mixing.  \n[[Kang et. al. 2017](Kang et. al. 2017), [Zhong et. al. 2020](Zhong et. al. 2020),  \nChatfi[eld et. al. 2014](eld et. al. 2014), Inoue 2018]  \nCarla[[Alexey et. al. 2017](Alexey et. al. 2017)], Udacity[molyakov [et. al. 2018](et. al. 2018)], Kuka[Lukaˇ[c et. al. 2018](c et. al. 2018)]  \nMost of these manipulations will result in invalid image data in the context of engineering drawings  \nThe simulated data can be generated with flexible experiment conditions in a reasonably short time  \nWe aim to create a parametric drawing generator that can synthesize a pool of new drawings in a simulated manner with a handful of existing drawing examples  \n8  \nAlgorithm Overview  \nThe labeling for such drawings requires humans with technical training. But vector drawings can serve to create a dataset for training a model that deploys on raster images.  \nData-driven Model  \nDXF drawings  \nSynthesize and  \nConvert to raster images  \nTraining  \nRaster images Vectorized segmentation  \nDeployment 9  \nApproach Overview  \nFour major modules to parse existing data, generate new data, vectorize the drawing, and predict the component type  \nAvector drawing (DXF)  \nJSON  \nMetadata  \nSynthesis Image  \nVectorized Features  \nType  \nPrediction  \n(To prepare training raster drawings) 10","cbCaii3fucsnu5M1","https://ap.wps.com/l/cbCaii3fucsnu5M1","pdf",4738943,26,"English","# Overview\n## Ultimate goal and automatic system\n# Motivation\n## Industrial part quotation systems\n## Notation and time-consuming steps\n# Literature Review: Engineering Drawing Analysis\n# Requirements\n## Data synthesis from labeled examples\n## Validity under technical rules\n# Related Work\n## General and simulated data augmentation in CV\n# Algorithm Overview\n## From vector drawings to training and deployment\n# Approach Overview\n## Four major modules: parse, synthesize, vectorize, predict","[{\"question\":\"What is the ultimate goal of the proposed system?\",\"answer\":\"The system aims at data-driven component segmentation of raster engineering drawings, enabling automated analysis of all components rather than partial interpretation for a specific scenario.\"},{\"question\":\"How does the pipeline prepare training data?\",\"answer\":\"It uses vectorized drawings (DXF) as a source, synthesizes additional drawings, converts them to raster images, and then trains a model to perform vectorized segmentation and component type prediction.\"},{\"question\":\"Why is data synthesis required to follow technical validity rules?\",\"answer\":\"Because many generic image augmentations would produce invalid engineering-drawing outputs, the generated synthetic drawings must be checked against technical rules to ensure they remain valid for domain learning.\"}]","Data Augmentation of Engineering Drawings for Data-driven Component Segmentation | PDF",1792796523,66]