Highly efficient human action recognition using compact 2DPCA-based descriptors in the spatial and transform domains

Human action recognition is considered as a challenging problem in the field of computer vision. Most of the reported algorithms are computationally expensive. In this paper, a novel system for human action recognition based on Two-Dimensional Principal Component Analysis (2DPCA) is presented. This method works directly on the optical flow and / or silhouette extracted from the input video in both the spatial domain and the transform domain. The algorithm reduces the computational complexity and storage requirements, while achieving high recognition accuracy, compared with the most recent reports in the field. Experimental results performed on the Weizmann action and the INIRIA IXMAS datasets confirm the excellent properties of the proposed algorithm.