Research On Fast Motion Estimation Algorithm In H.264

With advancements of video coding technology, computer technology and network infrastructure, multimedia system has become increasingly widely used. In recent years, with the presentation of video coding standard MPEG-4, H.264, etc., there have been a growing number of digital video products. Compared with previous video coding standard, H.264 has a higher coding efficiency and better network affinity. However, H.264's better coding gain is at the cost of coding complexity. Therefore, how to reduce the encoding complexity while maintaining a certain encoding quality becomes a topic worthy of study.Motion Estimation (ME) is an important part of video compression system, since it can achieve significant compression by exploiting the temporal redundancy existing in a video sequence. On the other hand, it is also the most computationally intensive part of the entire encoding process. A good fast motion search algorithm should efficiently speed up the encoding time and keep the quality of encoded video stable at the same time. Experimental results show that some classical motion estimation algorithms, such as Three-step search(TSS), Four-step search(FSS), Diamond Search(DS), Hexagon-Based Search(HEXBS), Hybrid Unsymmetrical-cross Multi-Hexagon-grid Search(UMHexagonS), reduce the computational requirement significantly by checking only some points inside the search window, while keeping a good error performance when compared with Full search (FS) algorithm. However, most of these algorithms get trapped in local minimum, yielding a significant loss in MV estimation performance.This paper proposed a Fast Hybrid Adaptive Motion Estimation (FHAME) algorithm, which consists of the Simple Dynamic Search Range (SDSR) algorithm, Successive Elimination Algorithm (SEA) and Early Termination Algorithm (ETA). Firstly, SDSR is used to reduce the number of search points (SP). Secondly, the SEA is employed to eliminate the number of SP. Finally, the ETA is utilized to early terminate the motion estimation of current block. The proposed algorithm is implemented in the reference software JM9.4. Experimental results indicate that the proposed algorithm can reduce the number of search points by about 93.1% and encoding time by about 42%, while maintaining almost the same bit rate and PSNR.