An improved body action recognition method based on manifold learning

The body action recognition is one of the key technologies of the computer vision. As the fact that the features of body action usually reside on low dimensional manifolds embedded in a high dimensional ambient space, a new method of body action recognition based on manifold learning is proposed in this paper. In the proposed method, a Linear Local Embedding of Difference (DLLE) algorithm is applied to get the low dimensional manifolds of the images and achieve human action recognition. The result shows that the DLLE method has more advantage in time-consuming and recognition accuracy rate than the other dimensionality reduction methods. Furthermore, the experimental results demonstrated the feasibility and effectiveness of the proposed algorithm in body action recognition.