Media.Player

Real-time Body Segmentation System

总结

Segmenting human Body automatically is very important for human behavior analysis, security system, and computer vision. In this paper, we present an efficient segmentation system for cutting human body out from video sequences in real-time. First, a learning based face detector is developed to rapidly find human faces. To speed up the detection process, a face rejection cascade is constructed to remove most of negative samples while retaining all face samples. Then, we develop a coarse-to-fine segmentation approach to extract the body based on a min-cut optimization. Finally, a new matting algorithm has been proposed to estimate the alpha-matte based on an adaptive trimap generation method. Experimental results demonstrate the effectiveness and robustness of our proposed method that can segment human body in real-time.

附加资料

Patent Number: US8233676B2
Application Number: US200844416A
Inventor: Ngan, King Ngi | Li, Hongliang
Priority Date: 7 Mar 2008
Priority Number: US8233676B2
Application Date: 7 Mar 2008
Publication Date: 31 Jul 2012
IPC Current: G06K000900
US Class: 382118 | 358538 | 382190 | 382195 | 382224 | 382254 | 382282
Assignee Applicant: The Chinese University of Hong Kong
Title: Real-time body segmentation system
Usefulness: Real-time body segmentation system
Summary: Method for recognizing characteristics and identity of a selected human body part i.e. human face (claimed).
Novelty: Selected human body part i.e. human face, characteristics and identity recognizing method, involves capturing portion of human body having known human body part, and reporting sub-dataset identities as output to output device

主要类别

信息和通信技术/电信

细分类别

软件/应用

申请号码

07/ENG/271

其他

Inventor(s): Professor King Ngi NGAN, Department of Electronic Engineering

Patent Status: US Patent Pending

Licensing Status: Available for licensing

国家/地区

香港