Human Action Recognition Based on Fusion Features Extraction of Adaptive Background Subtraction and Optical Flow Model

Figure 4

18 subareas of each subframe in the centers of subframe.