New adaptive partial distortion search using clustered pixel matching error characteristic

The partial distortion search is a particularly attractive fast block-matching algorithm, because it introduces no prediction error as compared with the full-search algorithm. It reduces the number of necessary matching evaluations for every searching point to save computation. In the literature, many researches have tried to improve block-matching algorithms by making use of an assumption that pixels with larger gradient magnitudes have larger matching errors on average. Based on a simple analysis, we have found that on average, pixel matching errors with similar magnitudes tend to appear in clusters for natural video sequences. By using this clustering characteristic, we propose an adaptive partial distortion search algorithm which significantly improves the computational efficiency of the original PDS. This approach is much better than other pixel gradient based adaptive PDS algorithms. In addition, our proposed algorithm is suitable for motion estimation of both opaque and boundary macroblocks of an arbitrary shaped object in MPEG-4 coding.