{"id":4684,"date":"2016-08-19T00:15:35","date_gmt":"2016-08-19T05:15:35","guid":{"rendered":"https:\/\/Engineering.Purdue.edu\/cdesign\/wp\/?p=4684"},"modified":"2021-08-10T15:37:38","modified_gmt":"2021-08-10T20:37:38","slug":"deep-learning-3d-shape-surfaces-using-geometry-images","status":"publish","type":"post","link":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/","title":{"rendered":"Deep Learning 3D Shape Surfaces using Geometry Images"},"content":{"rendered":"

Surfaces serve as a natural parametrization to 3D shapes. Learning surfaces using convolutional neural networks (CNNs) is a challenging task. Current paradigms to tackle this challenge are to either adapt the convolutional filters to operate on surfaces, learn spectral descriptors defined by the Laplace-Beltrami operator, or to drop surfaces altogether in lieu of voxelized inputs. Here we adopt an approach of converting the 3D shape into a \u2018geometry image\u2019<\/em> so that standard CNNs can directly be used to learn 3D shapes. We qualitatively and quantitatively validate that creating geometry images using authalic parametrization on a spherical domain is suitable for robust learning of 3D shape surfaces. This spherically parameterized shape is then projected and cut to convert the original 3D shape into a flat and regular geometry image. We propose a way to implicitly learn the topology and structure of 3D shapes using geometry images encoded with suitable features. We show the efficacy of our approach to learn 3D shape surfaces for classification and retrieval tasks on non-rigid and rigid shape datasets.<\/p>\n

\"img_eccv\"<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

Surfaces serve as a natural parametrization to 3D shapes. Learning surfaces using convolutional neural networks (CNNs) is a challenging task. Current paradigms to tackle this challenge are to either adapt the convolutional filters to operate on surfaces, learn spectral descriptors defined by the Laplace-Beltrami operator, or to drop surfaces altogether in lieu of voxelized inputs. […]<\/p>\n","protected":false},"author":1,"featured_media":4685,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_et_pb_use_builder":"","_et_pb_old_content":"","_et_gb_content_width":"","_links_to":"","_links_to_target":""},"categories":[357,174,268,286,285],"tags":[],"yoast_head":"\nDeep Learning 3D Shape Surfaces using Geometry Images - Convergence Design Lab, Purdue University<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Deep Learning 3D Shape Surfaces using Geometry Images - Convergence Design Lab, Purdue University\" \/>\n<meta property=\"og:description\" content=\"Surfaces serve as a natural parametrization to 3D shapes. Learning surfaces using convolutional neural networks (CNNs) is a challenging task. Current paradigms to tackle this challenge are to either adapt the convolutional filters to operate on surfaces, learn spectral descriptors defined by the Laplace-Beltrami operator, or to drop surfaces altogether in lieu of voxelized inputs. […]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/\" \/>\n<meta property=\"og:site_name\" content=\"Convergence Design Lab, Purdue University\" \/>\n<meta property=\"article:published_time\" content=\"2016-08-19T05:15:35+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2021-08-10T20:37:38+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2016\/08\/img_eccv.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1280\" \/>\n\t<meta property=\"og:image:height\" content=\"368\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Convergence Design Lab Admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"1 minute\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Organization\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#organization\",\"name\":\"Convergence Design Lab - Distinguished Professor Karthik Ramani - Purdue University\",\"url\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/\",\"sameAs\":[],\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2020\/07\/Convergence-Design-Lab-Logo-2020-07-1.png\",\"contentUrl\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2020\/07\/Convergence-Design-Lab-Logo-2020-07-1.png\",\"width\":960,\"height\":100,\"caption\":\"Convergence Design Lab - Distinguished Professor Karthik Ramani - Purdue University\"},\"image\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/logo\/image\/\"}},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#website\",\"url\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/\",\"name\":\"Convergence Design Lab, Purdue University\",\"description\":\"AR+VR+MR, knowledge and skill transfer, collaborative intelligence, robotics, smart product design, wearables, geometric deep learning, AI, computer vision.\",\"publisher\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"en-US\"},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#primaryimage\",\"url\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2016\/08\/img_eccv.png\",\"contentUrl\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2016\/08\/img_eccv.png\",\"width\":1280,\"height\":368},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#webpage\",\"url\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/\",\"name\":\"Deep Learning 3D Shape Surfaces using Geometry Images - Convergence Design Lab, Purdue University\",\"isPartOf\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#primaryimage\"},\"datePublished\":\"2016-08-19T05:15:35+00:00\",\"dateModified\":\"2021-08-10T20:37:38+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Deep Learning 3D Shape Surfaces using Geometry Images\"}]},{\"@type\":\"Article\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#webpage\"},\"author\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/person\/9b72ab0c8d1d87cbfa06ec2d95ed436b\"},\"headline\":\"Deep Learning 3D Shape Surfaces using Geometry Images\",\"datePublished\":\"2016-08-19T05:15:35+00:00\",\"dateModified\":\"2021-08-10T20:37:38+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#webpage\"},\"wordCount\":180,\"publisher\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#organization\"},\"image\":{\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2016\/08\/img_eccv.png\",\"articleSection\":[\"2016\",\"Ayan Sinha\",\"Recent Publications\",\"Shape Understanding\",\"Spatial Analytics\"],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/person\/9b72ab0c8d1d87cbfa06ec2d95ed436b\",\"name\":\"Convergence Design Lab Admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2020\/10\/CDL-Hand-Logo-1-96x96.png\",\"contentUrl\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2020\/10\/CDL-Hand-Logo-1-96x96.png\",\"caption\":\"Convergence Design Lab Admin\"},\"description\":\"Administrator of the Convergence Design Lab website.\",\"url\":\"https:\/\/engineering.purdue.edu\/cdesign\/wp\/author\/admin\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Deep Learning 3D Shape Surfaces using Geometry Images - Convergence Design Lab, Purdue University","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/","og_locale":"en_US","og_type":"article","og_title":"Deep Learning 3D Shape Surfaces using Geometry Images - Convergence Design Lab, Purdue University","og_description":"Surfaces serve as a natural parametrization to 3D shapes. Learning surfaces using convolutional neural networks (CNNs) is a challenging task. Current paradigms to tackle this challenge are to either adapt the convolutional filters to operate on surfaces, learn spectral descriptors defined by the Laplace-Beltrami operator, or to drop surfaces altogether in lieu of voxelized inputs. […]","og_url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/","og_site_name":"Convergence Design Lab, Purdue University","article_published_time":"2016-08-19T05:15:35+00:00","article_modified_time":"2021-08-10T20:37:38+00:00","og_image":[{"width":1280,"height":368,"url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2016\/08\/img_eccv.png","type":"image\/png"}],"twitter_card":"summary_large_image","twitter_misc":{"Written by":"Convergence Design Lab Admin","Est. reading time":"1 minute"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Organization","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#organization","name":"Convergence Design Lab - Distinguished Professor Karthik Ramani - Purdue University","url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/","sameAs":[],"logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/logo\/image\/","url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2020\/07\/Convergence-Design-Lab-Logo-2020-07-1.png","contentUrl":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2020\/07\/Convergence-Design-Lab-Logo-2020-07-1.png","width":960,"height":100,"caption":"Convergence Design Lab - Distinguished Professor Karthik Ramani - Purdue University"},"image":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/logo\/image\/"}},{"@type":"WebSite","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#website","url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/","name":"Convergence Design Lab, Purdue University","description":"AR+VR+MR, knowledge and skill transfer, collaborative intelligence, robotics, smart product design, wearables, geometric deep learning, AI, computer vision.","publisher":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"en-US"},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#primaryimage","url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2016\/08\/img_eccv.png","contentUrl":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2016\/08\/img_eccv.png","width":1280,"height":368},{"@type":"WebPage","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#webpage","url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/","name":"Deep Learning 3D Shape Surfaces using Geometry Images - Convergence Design Lab, Purdue University","isPartOf":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#website"},"primaryImageOfPage":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#primaryimage"},"datePublished":"2016-08-19T05:15:35+00:00","dateModified":"2021-08-10T20:37:38+00:00","breadcrumb":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/"},{"@type":"ListItem","position":2,"name":"Deep Learning 3D Shape Surfaces using Geometry Images"}]},{"@type":"Article","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#article","isPartOf":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#webpage"},"author":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/person\/9b72ab0c8d1d87cbfa06ec2d95ed436b"},"headline":"Deep Learning 3D Shape Surfaces using Geometry Images","datePublished":"2016-08-19T05:15:35+00:00","dateModified":"2021-08-10T20:37:38+00:00","mainEntityOfPage":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#webpage"},"wordCount":180,"publisher":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#organization"},"image":{"@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/deep-learning-3d-shape-surfaces-using-geometry-images\/#primaryimage"},"thumbnailUrl":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2016\/08\/img_eccv.png","articleSection":["2016","Ayan Sinha","Recent Publications","Shape Understanding","Spatial Analytics"],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/person\/9b72ab0c8d1d87cbfa06ec2d95ed436b","name":"Convergence Design Lab Admin","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/#\/schema\/person\/image\/","url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2020\/10\/CDL-Hand-Logo-1-96x96.png","contentUrl":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-content\/uploads\/2020\/10\/CDL-Hand-Logo-1-96x96.png","caption":"Convergence Design Lab Admin"},"description":"Administrator of the Convergence Design Lab website.","url":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/author\/admin\/"}]}},"_links":{"self":[{"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/posts\/4684"}],"collection":[{"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/comments?post=4684"}],"version-history":[{"count":4,"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/posts\/4684\/revisions"}],"predecessor-version":[{"id":5853,"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/posts\/4684\/revisions\/5853"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/media\/4685"}],"wp:attachment":[{"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/media?parent=4684"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/categories?post=4684"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/engineering.purdue.edu\/cdesign\/wp\/wp-json\/wp\/v2\/tags?post=4684"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}