Research On Block Matching Motion Estimation Algorithm For Video Coding

The video coding plays a vital role in the wide application of multimedia technology. Being an important part of the video coding, the motion estimation technology determines the quality and efficiency of the video coding to a large extent. As the standard of video coding has been reformed continuously, a number of motion estimation algorithms have emerged constantly, and the motion estimation algorithm based on block matching has become a hot spot in the field of the video. In resent years, organizations and scholars in different countries continuous devote themselves to the research, and the performance of the motion estimation algorithm is significantly improved.However, the result of the algorithm can’t achieve satisfactory degree. How to fulfill a real-time coding rapidly and high quality is the key to the block matching motion estimation algorithm. In this paper, we have done a deep research for the method of motion estimation based on block matching and analyzed the advantages and disadvantages of the matching criterion and search strategy. So we have proposed measures both form aspect of the search accuracy and speed for improving the shortcoming.In order to improve the accuracy of the block match in view of subjective visual quality and take characteristics of local pixel into account synthetically when the matching criterion works, we proposed a new strategy of the block matching which increased a new constraint that is local similarity degree in two macroblock based on SAD. To speed up the process of the block matching, the multilevel successive elimination algorithm is adopted. In the end, we done the emulation experiment for analysis that the new matching strategy based on the characteristics of the local pixel improve the quality of reconstructed image effectively when the local image mutations or is covered with shadow.Based on the advantages and disadvantages of the existing block matching algorithm, the prediction of initial searching point and early termination strategy is adopted at the beginning of searching according to the model of motion vector central distribution and motion vector correlation, which speed up the static-block detection. This paper proposes a directional adaptive sampling search (DASS) algorithm.According to the information of matching error in search patterns; the algorithm adopts two sorts of triangle search patterns adaptively, which reduce the matching calculation and improve the searching efficiency.The strategy of local sampling search is employed for the search of big-motion vector because that the local sampling search will be capable of checking comprehensively the region with the highest possibility of existing the minimum error point and makes up for defects of traps by local minima to a certain extent. Experimental results show that the proposed algorithm not only ensure the accuracy but also improve the speed of search.