Gesture Recognition for Chinese Traffic Police

In this paper, we present a five-part body model to recognize gestures made by Chinese traffic police in complex scenes for driver assistance systems and intelligent vehicles. Unlike most previous methods which require a training stage or a 3D measuring device to construct the body part appearance model, we propose to use the max-covering scheme to learn a five-part body model in an automatic way. Experimental results show that good recognition results can be obtained using the proposed method.