An Effective Hybrid Search Algorithm for Fast Block Motion Estimation

Motion estimation plays a key role in video coding, but its computational complexity is very high. In order to improve the performance of motion estimation, a hybrid search (HS) algorithm is proposed, which uses coarse-to-fine search strategy based on the ergodicity feature of chaotic search (CS) and the accurate feature of small cross search (SCS). Coarse motion vector is found by CS algorithm. And SCS algorithm is used to refine the search to get the final motion vector. Meanwhile, HS algorithm adopts the static block prejudgment and smooth motion prejudgment to further reduce the computational complexity. The simulation results show that HS algorithm has higher efficiency than the several famous fast motion estimation algorithms such as new three-step search (NTSS), four-step search (4SS), adaptive rood pattern search (ARPS) and diamond search (DS). Furthermore, compared with UMHexagonS adopted by H.264/AVC video coding system, HS algorithm can achieve very similar rate-distortion performance while saving the average ME time ratio 25.52%.