DocumentCode :
2452010
Title :
Star skeleton for human behavior recognition
Author :
Fan, Zhang ; Li, Guo ; Haixian, Lu ; Shu, Gui ; Jinkui, Li
Author_Institution :
Dept. of Electron. Sci. & Technol., Univ. of Sci. & Technol. of China, Hefei, China
fYear :
2012
fDate :
16-18 July 2012
Firstpage :
1046
Lastpage :
1050
Abstract :
Against to the traditional single-view two-dimensional body shape can not describe the diversity and ambiguity of human action, this paper present a 2D action recognition method based on human 3D star skeleton. It use multi-view two-dimensional body shape to describe and identify human action. The offline training process set up the 2D star skeleton codebook and 3D skeleton sequence. The online recognition process put up the vector quantization, get the matched 3D skeleton and then identify the human action. The experimental results show that this method has high recognition rate.
Keywords :
emotion recognition; geometry; image sequences; shape recognition; vectors; 2D action recognition metho; 2D star skeleton codebook; 3D skeleton sequence; human 3D star skeleton; human behavior recognition; multiview two-dimensional body shape; offline training process; online recognition process; single-view two-dimensional body shape; vector quantization; Databases; Hidden Markov models; Humans; Image recognition; Skeleton; Solid modeling; Training;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Audio, Language and Image Processing (ICALIP), 2012 International Conference on
Conference_Location :
Shanghai
Print_ISBN :
978-1-4673-0173-2
Type :
conf
DOI :
10.1109/ICALIP.2012.6376770
Filename :
6376770
Link To Document :
بازگشت