DocumentCode :
3007411
Title :
Human action recognition framework by fusing multiple features
Author :
Qian Xiao ; Jun Cheng
Author_Institution :
Shenzhen Inst. of Adv. Technol., Chinese Univ. of Hong Kong, Shenzhen, China
fYear :
2013
fDate :
26-28 Aug. 2013
Firstpage :
985
Lastpage :
990
Abstract :
In this paper, we propose a framework which fuses multiple features for action recognition in depth sequence. The fusion of multiple features is important for recognizing action since a single feature-based representation is inadequate to capture the variants. Hence, we use two types of features: i) a quantized vocabulary of local spatio-temporal descriptor HOG3D, and ii) a global projection based descriptor that computes the HOG from the Depth Motion Maps. To optimally combine these features, we input those features to different classifiers, where SVM is applied to estimate the probabilities of action labels. Then, we weight those probabilities respectively and sum it to find the maximum score of action labels. The proposed approach is tested on publicly available MSR Action3D dataset which demonstrates that fusion of multiple features help to achieve improved performance significantly, outperforming Li et al.[1] in most of the cases.
Keywords :
estimation theory; gesture recognition; image motion analysis; image sequences; probability; support vector machines; HOG3D; MSR Action3D dataset; SVM; action labels; depth motion maps; depth sequence; global projection based descriptor; human action recognition framework; local spatio-temporal descriptor; probability estimation; quantized vocabulary; single feature-based representation; Accuracy; Cameras; Shape; Support vector machines; Testing; Three-dimensional displays; Training; Action Recognition; Depth Maps; Feature Fusion; RGBD Camera;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Information and Automation (ICIA), 2013 IEEE International Conference on
Conference_Location :
Yinchuan
Type :
conf
DOI :
10.1109/ICInfA.2013.6720438
Filename :
6720438
Link To Document :
بازگشت