Action recognition using ensemble weighted multi-instance learning

Chen, Guang; Giuliani, Manuel; Clarke, Daniel; Gaschler, Andre; Knoll, Alois

Abstract

© 2014 IEEE. This paper deals with recognizing human actions in depth video data. Current state-of-the-art action recognition methods use hand-designed features, which are difficult to produce and time-consuming to extend to new modalities. In this paper, we propose a novel, 3.5D representation of a depth video for action recognition. A 3.5D graph of the depth video consists of a set of nodes that are the joints of the human body. Each joint is represented by a set of spatio-temporal features, which are computed by an unsupervised learning approach. However, if occlusions occur, the 3D positions of the joints are noisy which increases the intra-class variations in action classes. To address this problem, we propose the Ensemble Weighted Multi-Instance Learning approach (EnwMi) for the action recognition task. It considers the class imbalance and intra-class variations. We formulate the action recognition task with depth videos as a weighted multi-instance problem. We further integrate an ensemble learning method into the weighted multi-instance learning framework. Our approach is evaluated on Microsoft Research Action3D dataset, and the results show that it outperforms state-of-the-art methods.

Citation

Chen, G., Giuliani, M., Clarke, D., Gaschler, A., & Knoll, A. (2014). Action recognition using ensemble weighted multi-instance learning. IEEE International Conference on Robotics and Automation, 4520-4525. https://doi.org/10.1109/ICRA.2014.6907519

• Journal Article Type: Conference Paper
• Conference Name: Proceedings of the IEEE International Conference on Robotics and Automation 2014 (ICRA 2014)
• Acceptance Date: May 31, 2014
• Publication Date: Jan 1, 2014
• Journal: Proceedings - IEEE International Conference on Robotics and Automation
• Print ISSN: 1050-4729
• Publisher: Institute of Electrical and Electronics Engineers
• Peer Reviewed: Peer Reviewed
• Pages: 4520-4525
• DOI: https://doi.org/10.1109/ICRA.2014.6907519
• Keywords: joints, three-dimensional displays, training, kernel, histograms, feature extraction
• Public URL: https://uwe-repository.worktribe.com/output/811624
• Publisher URL: http://dx.doi.org/10.1109/ICRA.2014.6907519
• Additional Information: Title of Conference or Conference Proceedings : 2014 IEEE International Conference on Robotics and Automation (ICRA)