| With the fast developing of computer and communication technology, the traditional multimedia business have been unsatisfied to the demands of people, and multimedia apply-serves on network are expected, such as Wireless Communication, Streaming Media, Long-distance Medical Treatment, Video-Conference. These multimedia apply-serves have much higher demands on the transmission bandwidth and the transmission speed, especially on video. So, it has been the hotspot of the present researches to research and develop the new video compression technology with higher quality and efficiency.Among the factors which impact the quality and efficiency of video compression, Motion Estimation is one of the most influencing factors. The more exact the motion vector is, the higher the efficiency of the video compression is, and the better the quality of the encoded and decoded video images are. And Motion Estimation contributes to the heaviest computational load in the video coding system. Thus Motion Estimation algorithms have been the important research spot in video coding technology.In this paper, we study video coding (compression) and analyze the existing video coding methods. Then we put more effort on Motion Estimation technology in inter-frames compression method, and we conduct research on the search modes in Motion Estimation, the search strategies used by Motion Estimation algorithms, the design and implementation technology of Motion Estimation algorithm. The main achievements are listed as following:1). We present Parallelogram Search mode. Search mode decides the check point's number in each searching step, so the efficiency of Motion Estimation algorithms is partly decided by the search modes used by Motion Estimation algorithms. In order to overcome the shortcomings of the existing algorithms, we present PS mode, which includes three PS models and an extended PS model.When the present search-direction is in some one search-direction region, the search model corresponding to the search-direction region is selected. When the search-direction is beyond the search-direction region, and enters another search-direction region, the previous search model is replaced by another search model. Compared with other search modes, PS mode can avoid the blindness on searching, and reduce the possibility of getting in the local optimization.2). We put forward the search strategy that search mode and search direction are context-adaptive, and we develop a novel fast motion estimation algorithm (CAPS) based on the search strategy. Based on search strategy, the search direction of next search step is decided by the existing search results, and then one search model is selected. So, CAPS algorithm based on the search strategy can exactly and fastly find the optimization point. CAPS algorithm uses PS mode, and searches in larger step size, so CAPS algorithm overcomes the shortcoming of getting in the local optimization in a large degree. Experiments show that CPAS algorithm improves the search efficiency by 50% in comparison with UMHexagonS algorithm at the same rate-loss, and by 92% in compared with FS algorithm. More acute the motion of video sequence, the higher the improvement speed of CAPS algorithm can obtain.3). We designed Motion Estimation method based on Genetic Algorithm. Most of the existing motion estimation algorithms have the shortcoming of getting in the local optimization, which leads to the quality decline of video images. GA has the advantage of global searching, so we designed MEBGA method to improve the quality of video images. Experiments show that MEBGA method can achieve the similar PSNR of FS algorithm, and largely reduce the chances of being trapped to local optima, and this improves the quality of video coding. In addition, the statistic time cost on ME indicates that MEBGA method largely reduced the ME time, and thus improve the search speed. |
PDF Full Text Request