A novel fast motion estimation algorithm for video coding

This paper presents an efficient search algorithm utilizes the relationship between local motion similarity and the variation degree of motion vector. The main feature of proposed algorithm are: the motion vectors are classified as the unchanged, the slightly changed and the greatly changed motion vector according to corresponding local motion similarity; then different search schemes are adopted for different kind motion vectors. The experimental results show the proposed algorithm has a significant computational speedup compared with other popular fast algorithms, but offers a similar, even better performance.