文章基本信息

标题：Motion Tracking, Retrieval and 3D Reconstruction from Video
本地全文：下载
作者：Hashim Yasin ; Bj"orn Kr"uger ; Andreas Weber 等
期刊名称：International Journal of Multimedia and Ubiquitous Engineering
印刷版ISSN：1975-0080
出版年度：2014
卷号：9
期号：2
页码：265-280
出版社：SERSC
摘要：The work at hand presents a novel data-driven framework for 3D full body human motionreconstruction from uncalibrated monocular video data. To this end, we develop a knowledgebase by taking 2D samples of the motion capture library from di.erent viewing directions.This allows later steps to handle 2D query videos without any information on the viewingdirection. We detect and track features from input video sequences by utilizing low-levelimage based feature detection techniques like MSER and SURF. This process is stabilized byback projection of high-level 3D prior information obtained from the motion capture libraryto the image plane. Extraction of suitable feature sets from both, input control signals andmotion capture data, enables us to retrieve the best relevant prior poses from the motioncapture library by employing fast motion retrieval techniques. Finally, 3D motion sequencesare reconstructed by non-linear energy minimization, that takes into account multiple priorterms. Furthermore, we propose a method to estimate camera parameters from input videoitself and sampling of motion capture library
关键词：Feature detection and tracking; Motion retrieval; Camera parameters; 3D;motion reconstruction