{"id":295,"date":"2014-01-22T13:30:18","date_gmt":"2014-01-22T11:30:18","guid":{"rendered":"http:\/\/www.mpateraki.org\/?p=295"},"modified":"2015-01-23T11:25:16","modified_gmt":"2015-01-23T09:25:16","slug":"robust-model-based-3d-torso-pose-estimation-in-rgb-d-sequences","status":"publish","type":"post","link":"https:\/\/mpateraki.org\/?p=295","title":{"rendered":"Robust Model-Based 3D Torso Pose Estimation in RGB-D Sequences"},"content":{"rendered":"<p>Sigalas M., <strong>Pateraki M.<\/strong>, Oikonomidis I., Trahanias P., 2013. Robust Model-Based 3D Torso Pose Estimation in RGB-D Sequences. In Proc. of the 2nd IEEE Workshop on Dynamic Shape Capture and Analysis, held within Intl. Conference on Computer Vision (ICCV), 7 December, Sydney, Australia.\u00a0 [<a href=\"http:\/\/dx.doi.org\/10.1109\/ICCVW.2013.49\" target=\"_blank\">doi<\/a>] [<a href=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2014\/11\/4DMOD2013_ID_8_final.pdf\" target=\"_blank\">pdf<\/a>] [<a href=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2014\/11\/sigalas_iccv13.txt\" target=\"_blank\">bib<\/a>]<\/p>\n<p>Abstract:<\/p>\n<p>Free-form Human Robot Interaction (HRI) in naturalistic\u00a0environments remains a challenging computer vision\u00a0task. In this context, the extraction of human-body pose information\u00a0is of utmost importance. Although the emergence\u00a0of real-time depth cameras greatly facilitated this task, issues\u00a0which limit the performance of existing methods in relevant\u00a0HRI applications still exist. Applicability of current\u00a0state-of-the art approaches is constrained by their inherent\u00a0requirement of an initialization phase prior to deriving body\u00a0pose information, which in complex, realistic scenarios, is\u00a0often hard, if not impossible.\u00a0In this work we present a data-driven model-based\u00a0method for 3D torso pose estimation from RGB-D image\u00a0sequences, eliminating the requirement of an initialization\u00a0phase. The detected face of the user steers the initiation\u00a0of shoulder areas hypotheses, based on illumination, scale\u00a0and pose invariant features on the RGB silhouette. Depth\u00a0point cloud information is subsequently utilized to approximate\u00a0the shoulder joints and model the human torso based\u00a0on a set of 3D geometric primitives and the estimation of the\u00a03D torso pose is derived via a global optimization scheme.\u00a0Experimental results in various environments, as well as using\u00a0ground truth data and comparing to OpenNI User generator\u00a0middleware results, validate the effectiveness of the\u00a0proposed method.<\/p>\n<table align=\"center\">\n<tbody align=\"center\">\n<tr align=\"center\">\n<td align=\"center\" width=\"100%\" colspan=\"3\"><a href=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2015\/01\/method_over.jpg\"><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-297 size-large\" src=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2015\/01\/method_over-1024x516.jpg\" alt=\"\" width=\"540\" height=\"272\" srcset=\"https:\/\/mpateraki.org\/wp-content\/uploads\/2015\/01\/method_over-1024x516.jpg 1024w, https:\/\/mpateraki.org\/wp-content\/uploads\/2015\/01\/method_over-300x151.jpg 300w\" sizes=\"auto, (max-width: 540px) 100vw, 540px\" \/><\/a><\/td>\n<\/tr>\n<tr align=\"center\">\n<td align=\"center\" width=\"33%\"><a href=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2015\/01\/res6.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"300\" height=\"225\" class=\"aligncenter wp-image-298 size-medium\" src=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2015\/01\/res6-300x225.jpg\" alt=\"\" srcset=\"https:\/\/mpateraki.org\/wp-content\/uploads\/2015\/01\/res6-300x225.jpg 300w, https:\/\/mpateraki.org\/wp-content\/uploads\/2015\/01\/res6.jpg 640w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/td>\n<td align=\"center\" width=\"33%\"><a href=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2015\/01\/res3.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"300\" height=\"225\" class=\"aligncenter wp-image-299 size-medium\" src=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2015\/01\/res3-300x225.jpg\" alt=\"\" srcset=\"https:\/\/mpateraki.org\/wp-content\/uploads\/2015\/01\/res3-300x225.jpg 300w, https:\/\/mpateraki.org\/wp-content\/uploads\/2015\/01\/res3.jpg 640w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/td>\n<td align=\"center\" width=\"33%\"><a href=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2015\/01\/res7.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"300\" height=\"225\" class=\"aligncenter wp-image-300 size-medium\" src=\"http:\/\/www.mpateraki.org\/wp-content\/uploads\/2015\/01\/res7-300x225.jpg\" alt=\"\" srcset=\"https:\/\/mpateraki.org\/wp-content\/uploads\/2015\/01\/res7-300x225.jpg 300w, https:\/\/mpateraki.org\/wp-content\/uploads\/2015\/01\/res7.jpg 640w\" sizes=\"auto, (max-width: 300px) 100vw, 300px\" \/><\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n","protected":false},"excerpt":{"rendered":"<p>Sigalas M., Pateraki M., Oikonomidis I., Trahanias P., 2013. Robust Model-Based 3D Torso Pose Estimation in RGB-D Sequences. In Proc. of the 2nd IEEE Workshop on Dynamic Shape Capture and Analysis, held within Intl. Conference on Computer Vision (ICCV), 7 December, Sydney, Australia.\u00a0 [doi] [pdf] [bib] Abstract: Free-form Human Robot Interaction (HRI) in naturalistic\u00a0environments remains [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":298,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[],"class_list":["post-295","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-publications"],"_links":{"self":[{"href":"https:\/\/mpateraki.org\/index.php?rest_route=\/wp\/v2\/posts\/295","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mpateraki.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/mpateraki.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/mpateraki.org\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/mpateraki.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=295"}],"version-history":[{"count":5,"href":"https:\/\/mpateraki.org\/index.php?rest_route=\/wp\/v2\/posts\/295\/revisions"}],"predecessor-version":[{"id":305,"href":"https:\/\/mpateraki.org\/index.php?rest_route=\/wp\/v2\/posts\/295\/revisions\/305"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/mpateraki.org\/index.php?rest_route=\/wp\/v2\/media\/298"}],"wp:attachment":[{"href":"https:\/\/mpateraki.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=295"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/mpateraki.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=295"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/mpateraki.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=295"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}