Video Objects Behavior Recognition Using Fast MHI Approach

The concept of motion history image (MHI) is widely adopted by many researchers to solve problems of video objects behavior recognition. In the case of predefined behaviors with various durations, the computation time of behavior recognition will increase intensely. In this paper, a fast MHI approach is proposed to reduce the computation time of the MHI approach by storing multiple sets of features for a predefined behavior and using the partial distance computation method. In this paper, 9 local orientations proposed by Cheng et al and squared Euclidean distance are used in our behavior matching process to demonstrate the performance of the proposed approach. Experimental results show that the proposed method can effectively reduce the computation time of the MHI method.