Video Motion Segmentation Based on Double Sliding Window

This paper puts forward a motion segmentation method based on double sliding window. In order to learn about the motions of a person in a certain period of time and do further research on the video segmentation, we should do some research on the motions people have every day. So the new method to segment video motions is presented. The new algorithm is to segment the different motions of the human body in the videos. Considering that the frequency of two different kinds of motion is also different, so in order to segment the motions, we mainly segment them according to the unique frequencies of different movements. Finally, this method has smaller computation and is easy to be implemented. Experimental results demonstrate the performance and feasibility of the new method.