[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-664175-105":3,"detail-sidebar-cat-0-en-105":80,"doc-detail-664175-en":130},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":73,"head_meta":75,"extra_data":77,"updated_unix":79},105,"en","learning-to-parse-wireframes-in-images-of-man-made-environments","Learning to Parse Wireframes in Images of Man-Made Environments","","Learning to Parse Wireframes in Images of Man-Made Environments introduces a learning-based pipeline for extracting a “wireframe” representation from cluttered man-made scenes. The wireframe encodes salient straight lines and their junctions to capture efficient large-scale geometry and object shapes. A new dataset of 5,000+ human-labeled images supports two convolutional networks for junction and line extraction with large spatial support. Experiments show clear gains over prior methods and demonstrate benefits for correspondence, 3D reconstruction, mapping, localization, and navigation.",{"@graph":14,"@context":72},[15,34,55],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/learning-to-parse-wireframes-in-images-of-man-made-environments/664175/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/learning-to-parse-wireframes-in-images-of-man-made-environments/664175.png","ImageObject",300,407,{"name":42,"@type":43},"cimolpedes","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-30","2026-10-29",true,{"@type":52,"interactionType":53,"userInteractionCount":26},"InteractionCounter",{"@type":54},"ViewAction",{"@type":56,"mainEntity":57},"FAQPage",[58,64,68],{"name":59,"@type":60,"acceptedAnswer":61},"What is the wireframe representation proposed in the paper?","Question",{"text":62,"@type":63},"The wireframe contains salient straight lines and their junctions, encoding large-scale geometry and object shapes for man-made environments.","Answer",{"name":65,"@type":60,"acceptedAnswer":66},"How do the authors approach wireframe extraction?",{"text":67,"@type":63},"They build two convolutional neural networks: one for junction extraction and one for line segment extraction, both using large spatial support and trained on a newly collected labeled dataset.",{"name":69,"@type":60,"acceptedAnswer":70},"Why is parsing wireframes useful for downstream vision tasks?",{"text":71,"@type":63},"The paper argues that wireframes provide compact structural information that can benefit feature correspondence, 3D reconstruction, vision-based mapping, localization, and navigation.","https://schema.org",{"og:url":32,"og:type":74,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":76,"canonical":32},"index,follow",{"doc_id":78,"site_id":7},664175,1793242337,{"code":4,"msg":81,"data":82},"success",[83,87,91,95,100,105,110,114,119,122,126],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":84,"show_sort_weight":85,"slug":86},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":88,"show_sort_weight":89,"slug":90},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":92,"show_sort_weight":93,"slug":94},"Exam",70,"exam",{"id":96,"doc_module":4,"doc_module_name":25,"category_name":97,"show_sort_weight":98,"slug":99},5,"Comic",60,"comic",{"id":101,"doc_module":4,"doc_module_name":25,"category_name":102,"show_sort_weight":103,"slug":104},6,"Technology",50,"technology",{"id":106,"doc_module":4,"doc_module_name":25,"category_name":107,"show_sort_weight":108,"slug":109},7,"Healthcare",40,"healthcare",{"id":111,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":112,"slug":113},8,30,"research-report",{"id":115,"doc_module":4,"doc_module_name":25,"category_name":116,"show_sort_weight":117,"slug":118},9,"Religion & Spirituality",20,"religion-spirituality",{"id":117,"doc_module":4,"doc_module_name":25,"category_name":120,"show_sort_weight":117,"slug":121},"World Cup","world-cup",{"id":123,"doc_module":4,"doc_module_name":25,"category_name":124,"show_sort_weight":123,"slug":125},10,"Lifestyle","lifestyle",{"id":127,"doc_module":4,"doc_module_name":25,"category_name":128,"show_sort_weight":96,"slug":129},19,"General","general",{"code":4,"msg":81,"data":131},{"doc_id":78,"user_id":132,"nickname":42,"user_avatar":133,"doc_module":4,"category_id":111,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":26,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":123,"language":139,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":140,"faqs":141,"seo_title":142,"seo_description":12,"update_tm":79,"read_time":143},962094243610,"https://ap-avatar.wpscdn.com/davatar_3d24733baf745e90a7e4bdd5f77d97b2","Learning to Parse Wireframes in Images of Man-Made Environments  \nKun Huang 1 , Yifan Wang 1 , Zihan Zhou2 , Tianjiao Ding 1 , Shenghua Gao 1 , and Yi Ma3  \n1 ShanghaiTech University {huangkun, wangyf, dingtj, [gaoshh](gaoshh}@shanghaitech.edu.cn)[}](gaoshh}@shanghaitech.edu.cn)[@shanghaitech.edu.cn](gaoshh}@shanghaitech.edu.cn)[ ](gaoshh}@shanghaitech.edu.cn)[2](2 The Pennsylvania State University zzhou@ist.psu.edu)[ The Pennsylvania State University](2 The Pennsylvania State University zzhou@ist.psu.edu)[ zzhou@ist.psu.edu](2 The Pennsylvania State University zzhou@ist.psu.edu)  \n3 University of California, [Berkeley](Berkeley yima@eecs.berkeley.edu)[ yima@eecs.berkeley.edu](Berkeley yima@eecs.berkeley.edu)  \nAbstract  \nIn this paper, we propose a learning-based approach to the task of automatically extracting a “wireframe” representation for images of cluttered man-made environments. The wireframe (see Fig. 1) contains all salient straight lines and their junctions of the scene that encode ef􀀂ciently and accurately large-scale geometry and object shapes. To this end, we have built a very large new dataset of over 5,000 images with wireframes thoroughly labelled by humans. We have proposed two convolutional neural networks that are suitable for extracting junctions and lines with large spatial support, respectively. The networks trained on our dataset have achieved signi􀀂cantly better performance than stateof-the-art methods for junction detection and line segment detection, respectively. We have conducted extensive experiments to evaluate quantitatively and qualitatively the wireframes obtained by our method, and have convincingly shown that effectively and ef􀀂ciently parsing wireframes for images of man-made environments is a feasible goal within reach. Such wireframes could bene􀀂t many important visual tasks such as feature correspondence, 3D reconstruction, vision-based mapping, localization, and navigation. The data and source code are available at [https:](https:)// [github.com/huangkuns/wireframe](github.com/huangkuns/wireframe).  \n1. Introduction  \nHow to infer 3D geometric information of a scene from 2D images has been a fundamental problem in computer vision. Conventional approaches to build a 3D model typically rely on detecting, matching, and triangulating local image features (e.g. corners, edges, SIFT features, and patches) . One great advantage of working with local features is that the system can be somewhat oblivious to the scene, as long as it contains suf􀀂cient distinguishable features. Meanwhile, modern applications of computer vision systems often require an autonomous agent (e.g., a car, a robot, or a UAV) to ef􀀂ciently and effectively negotiate with a physical space in cluttered man-made (indoor or outdoor) environments. Such scenarios present signi􀀂cant  \nFigure 1 . First row: Examples of typical indoor or outdoor scenes with geometrically meaningful wireframes labelled by humans; Second row: Wireframes automatically extracted by our method.  \nchallenges to the current local-feature based approaches: Man-made environments typically consist of large textureless surfaces (e.g. white walls or the ground); or they maybe full of repetitive patterns hence local features are ambiguous to match; and the visual localization system is required to work robustly and accurately over extended routes and sometimes across very large baseline between views.  \nNevertheless, the human vision system seems capable of effortlessly localizing or navigating among such environments arguably by exploiting larger-scale (global or semiglobal) structural features or regularities of the scene. For instance, many works [6, 21, 14, 11, 35] have demonstrated that prior knowledge about the scene such as a Manhattan world could signi􀀂cantly bene􀀂t the 3D reconstruction tasks. The Manhattan assumption can often be violated in cluttered man-made environments, but it is rather safe to assume that man-made environments are dominantly piece","cbCaitaBAUwchgTc","https://ap.wps.com/l/cbCaitaBAUwchgTc","pdf",2199156,"English","# Abstract\n# Introduction","[{\"question\":\"What is the wireframe representation proposed in the paper?\",\"answer\":\"The wireframe contains salient straight lines and their junctions, encoding large-scale geometry and object shapes for man-made environments.\"},{\"question\":\"How do the authors approach wireframe extraction?\",\"answer\":\"They build two convolutional neural networks: one for junction extraction and one for line segment extraction, both using large spatial support and trained on a newly collected labeled dataset.\"},{\"question\":\"Why is parsing wireframes useful for downstream vision tasks?\",\"answer\":\"The paper argues that wireframes provide compact structural information that can benefit feature correspondence, 3D reconstruction, vision-based mapping, localization, and navigation.\"}]","Learning to Parse Wireframes in Images of Man-Made Environments | PDF",25]