Efficient Feature Representation for Human Action Recognition in DS&SP

Abstract:

Human action often requires a large volume and computation-consuming representation for an accurate recognition with good diversity as the large complexity and variability of actions and scenarios. In this paper, an efficiency combined action representation approach is proposed to deal with the dilemma between accuracy and diversity. Two action features are extracted for combination from a Kinect sensor: silhouette and 3D message. An improved Histograms of Gradient named Interest-HOG is proposed for silhouette representation while the feature angles between skeleton points are calculated as the 3D representation. Kernel Principle Componet Analysis (KPCA) is also applied bidirectionally in our work to process the Interest-HOG descriptor for getting a concise and normative vector whose volume is same as the 3D one aimed at a successful combining. A depth dataset named DS&SP including 10 kinds of actions performed by 12 persons in 4 scenarios is built as the benchmark for our approach based on which Support Vector Machine (SVM) is employed for training and testing. Experimental results show that our approach has good performance in accuracy, efficiency and robustness of self-occlusion.

• Periodical: Advanced Materials Research (Volumes 889-890)
• Pages: 1057-1064
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.889-890.1057
• Online since: February 2014
• Authors: Rui Feng Li, Liang Liang Wang*, Teng Fei Zhang
• Keywords: Action Recognition, Depth Dataset, Feature Representation

References

[1] Sunyoung Cho, Sooyeong Kwak, Hyeran Byun: submitted to Journal of Pattern Recognition Letters (2013).

DOI: 10.1109/icics.2011.6173540

[2] Albert Clapés, Miguel Reyes, Sergio Escalera: submitted to Journal of Pattern Recognition Letters (2013).

[3] Evan A. Suma, David M. Krum, Belinda Lange, Sebastian Koenig, Albert Rizzo, Mark Bolas: submitted to Journal of Computers & Graphics (2013).

[4] Agarwal, A.; Triggs, B.: submitted to Journal of IEEE Transactions on Pattern Analysis and Machine Intelligence (2006).

[5] Li Guan; Franco, J. -S.; Pollefeys, M.: 3D Occlusion Inference from Silhouette Cues (Trans Tech Publications, USA 2007).

DOI: 10.1109/cvpr.2007.383145

[6] Bobick, A.F.; Davis, J.W.: submitted to Journal of IEEE Transactions on Pattern Analysis and Machine Intelligence (2001).

[7] Smisek, J.; Jancosek, M.; Pajdla, T.: 3D with Kinect (Trans Tech Publications, Spain 2011).

[8] Macknojia, R.; Chavez-Aragon, A.; Payeur, P.; Laganiere, R.: Experimental characterization of two generations of Kinect's depth sensors (Trans Tech Publications, Germany 2012).

DOI: 10.1109/rose.2012.6402634

[9] Satpathy, A.; Jiang, X.; Eng, H.L.: submitted to Journal of IEEE Transactions on Image Processing (2013).

[10] Yanwei Pang, Yuan Yuan, Xuelong Li, Jing Pan: submitted to Journal of Signal Processing (2011).

[11] Cruz, L.; Lucio, D.; Velho, L.: Kinect and RGBD Images: Challenges and Applications (Trans Tech Publications, Brazil 2012).

DOI: 10.1109/sibgrapi-t.2012.13

[12] Chang, S.G.; Bin Yu; Vetterli, M.: submitted to Journal of IEEE Transactions on Image Processing (2000).

[13] Dalal, N.; Triggs, B.: Histograms of oriented gradients for human detection (Trans Tech Publications, USA 2005).

[14] Jian Yang; Frangi, A.F.; Jing-Yu Yang; Zhang, D.; Zhong Jin: submitted to Journal of IEEE Transactions on Pattern Analysis and Machine Intelligence (2005).

DOI: 10.1109/tpami.2005.33

[15] Huimin Qian, Yaobin Mao, Wenbo Xiang, Zhiquan Wang: submitted to Journal of Pattern Recognition Letters (2010).