Fast video object segmentation using Markov random field

A fast video object segmentation algorithm is proposed in this paper. The algorithm utilizes the motion vectors from blocks with variable block sizes to identify background motion model and moving objects. Markov random field is used to model the foreground field to enhance spatial and temporal continuity of objects. To speed up the segmentation time, time-consuming spatial segmentation techniques are avoided. Instead, spatial information in the form of Walsh Hadamard transform coefficients is utilized to improve segmentation accuracy. Experimental results show that the proposed algorithm can effectively extract moving objects from different kind of video sequences. The computation time of the segmentation process is merely about 75 ms per CIF frame using a normal PC, allowing the algorithm to be applied in real-time applications such as video surveillance and conferencing.