Joint Parsing and Segmentation of Articulated Human Bodies From Videos

DOI

10.2991/asei-15.2015.21How to use a DOI?

Keywords

Human pose; Segmentation; Parsing; Grabcut; Articulated Model

Abstract

Human body parsing and segmentation are two fundamental problems in computer vision. In this paper we build an automatic system for solving the two problems together. We reconstruct the mixture-of-part model by adding various features, because of the high relevance constructed by the new model, we are able to execute precise parsing inference on the whole human body poses. For segmentation we replace the user interaction input with the detection boxes, since the detection boxes fit the human body part well, it is easy for the refined segmentation to distinguish the foreground and background parts. Experiment results show apparent improvements compared with former methods.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhao Liu
AU  - Jingrun Sun
AU  - Chun Chen
PY  - 2015/05
DA  - 2015/05
TI  - Joint Parsing and Segmentation of Articulated Human Bodies From Videos
BT  - Proceedings of the 2015 International conference on Applied Science and Engineering Innovation
PB  - Atlantis Press
SP  - 93
EP  - 97
SN  - 2352-5401
UR  - https://doi.org/10.2991/asei-15.2015.21
DO  - 10.2991/asei-15.2015.21
ID  - Liu2015/05
ER  -