Human body gait tracking based on the maximum step angle

This paper present a novel method to track the gait of human body. The point distribution models is considered to train the 2D image samples including the points of shapes and skeletons, then construct the global 2D spatio-temporal models framework to track the human gait, among which spatial constraints based on the maximum step angle between legs first proposed are used to find the most probable models. Experimental results show that the proposed algorithm can track the human body gait quickly and accurately.