{"id":14818,"date":"2021-06-07T08:56:21","date_gmt":"2021-06-07T15:56:21","guid":{"rendered":"http:\/\/jnack.com\/blog\/?p=14818"},"modified":"2021-06-07T08:56:23","modified_gmt":"2021-06-07T15:56:23","slug":"adobe-humor-estimates-3d-human-movements-from-2d-inputs","status":"publish","type":"post","link":"http:\/\/jnack.com\/blog\/2021\/06\/07\/adobe-humor-estimates-3d-human-movements-from-2d-inputs\/","title":{"rendered":"Adobe &#8220;HuMoR&#8221; estimates 3D human movements from 2D inputs"},"content":{"rendered":"\n<p><a href=\"https:\/\/geometry.stanford.edu\/projects\/humor\/\">This Adobe Research collaboration<\/a> with Stanford &amp; Brown Universities aims to make sense of people moving in space, despite having just 2D video as an input:<\/p>\n\n\n\n<figure class=\"wp-block-embed is-type-video is-provider-youtube wp-block-embed-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<iframe loading=\"lazy\" title=\"HuMoR Teaser Video\" width=\"604\" height=\"340\" src=\"https:\/\/www.youtube.com\/embed\/5VWirxUHG0Y?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture\" allowfullscreen><\/iframe>\n<\/div><\/figure>\n\n\n\n<blockquote class=\"wp-block-quote\"><p>We introduce HuMoR: a 3D&nbsp;<strong>Hu<\/strong>man&nbsp;<strong>Mo<\/strong>tion Model for&nbsp;<strong>R<\/strong>obust Estimation of temporal pose and shape. Though substantial progress has been made in&nbsp;<strong>estimating 3D human motion<\/strong>&nbsp;and shape from dynamic observations, recovering plausible pose sequences in the&nbsp;<strong>presence of noise and occlusions<\/strong>&nbsp;remains a challenge. For this purpose, we propose an&nbsp;<strong>expressive generative model<\/strong>&nbsp;in the form of a conditional variational autoencoder, which learns a distribution of the change in pose at each step of a motion sequence. Furthermore, we introduce a&nbsp;<strong>flexible optimization-based approach that leverages HuMoR as a motion prior<\/strong>&nbsp;to robustly estimate plausible pose and shape from ambiguous observations. Through extensive evaluations, we demonstrate that our model generalizes to diverse motions and body shapes after training on a large motion capture dataset, and enables motion reconstruction from multiple input modalities including 3D keypoints and RGB(-D) videos.<\/p><\/blockquote>\n","protected":false},"excerpt":{"rendered":"<p>This Adobe Research collaboration with Stanford &amp; Brown Universities aims to make sense of people moving in space, despite having just 2D video as an input: We introduce HuMoR: a 3D&nbsp;Human&nbsp;Motion Model for&nbsp;Robust Estimation of temporal pose and shape. Though substantial progress has been made in&nbsp;estimating 3D human motion&nbsp;and shape from dynamic observations, recovering plausible [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[18,66],"tags":[],"_links":{"self":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/14818"}],"collection":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/comments?post=14818"}],"version-history":[{"count":6,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/14818\/revisions"}],"predecessor-version":[{"id":14963,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/posts\/14818\/revisions\/14963"}],"wp:attachment":[{"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/media?parent=14818"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/categories?post=14818"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/jnack.com\/blog\/wp-json\/wp\/v2\/tags?post=14818"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}