Reduced computation using adaptive search window size for H.264 multi-frame motion estimation

Motion estimation of the H.264 takes the advantage of multi-reference frames to select the best matching position. The full process of searching up to 5 past reference frames leads to high computational complexity and casts a heavy burden on the encoder. After a careful analysis, it is found that the motion vector obtained in the 1^st reference frame can be used as a guide to adaptively set an appropriate size of the search window for the rest of the reference frames. Meanwhile, if several small blocks have the same motion vector, the combined large block can utilize such a motion vector as an initial point for further refinement within a small search window size. Results of our experimental work using the JM9.6 show that the proposed algorithm can reduce the computational complexity to 10% of the original full search and meanwhile keep the PSNR drop less than 0.02dB with less than 1% increase in bitrate.