Moving object segmentation in H.264/AVC compressed domain using ant colony algorithm

This paper presents a novel approach for moving object segmentation in the H.264/AVC compressed domain, which based on ant colony clustering algorithm. Firstly, the motion vector (MV) field is extracted from the H.264/AVC compressed video, and then merges motion vectors with the same characteristic. Secondly, an improved ant colony clustering algorithm is used to classify the MV field into different motion homogenous regions. Finally, the orientation histogram of the MV field and the final cluster centers are exploited to determine the moving object regions. Experimental results for several video sequences demonstrate that the proposed approach can extract moving object effectively.