Human Action Recognition Based on Non-linear SVM Decision Tree

Even great efforts have been made for decades, the recognition of human actions is still an immature technology that attracted plenty of people. This paper presents a novel classification method for multiple kinds of actions from videos, which is represented by centroid-radii model descriptor and classified by Non-linear SVM Decision Tree (NSVMDT). It solves the problem that trades off activity recognition rate and computational complexity, rather than highlight the former exclusively. We extract binary silhouette images after the background model is created. Then the low-level features are described by centroid-radii model. We utilize NSVMDT to train and classify video sequences, and demonstrate the usability with many sequences. Compared with others, our method is applicable to intelligent surveillance, and its advantage lies in robustness, computational complexity, geometric invariance and classification performance.