Fast Motion Estimation Algorithms Based On An Octagon-Grid For Video Coding

With the development of the visual information technology, video has been playing more and more important role in informational communication. Therefore, video coding has been the hot topic for the researcher in this field. The JVT (Joint Video Team), formed by the experts from VCEG(Video Coding Experts Group) and MPEG(Moving Picture Experts Group), worked out a series of video coding standard including H.261, H.263, H.264 and the revising edition. Motion estimation is the key technology of video compression, which dominates the performance of compression and the time for calculating. In the new standard H.264, motion estimation used the three new image processing technology, i.e. multi-reference frame, multi-mode block and fractional-pixel precision. New technologies become new field to be studied. Based on the H.264 standard, the paper studies every part of the motion estimation process, and makes some improvement in the steps. The detailed contents are as follows.Firstly, it introduces the frame of video coding and the motion estimation technology in brief. In the whole frame of video coding, motion estimation is very important. What algorithm for the motion estimation decides the efficiency of video compression, and influences the development of the visual communication indirectly.Secondly, in the paper the technology of fractional-pixel precision block matching for movement estimate is studied. Its three parts, the initialization point for searching and pixel searching and early termination, are studied and their algorithms are improved. A total new search mode, which likes an octagon including 12 search points, is applied in the proposed algorithm. As a result, the number of pixels for searching descends and the efficiency of coding is enhanced.Thirdly, based on the octagon mode and the proposed algorithm, it consider the advantage of varies of mode such as the octagon, the symmetrical cross, the diamond and the triangle. At last, it finds out an effective search mode with appropriate early termination. Shown by the result, the new algorithm for the motion estimation reduces the calculating complex.Finally, Experiments are made to prove the performance of these new algorithms. Then we can find that they are better than others by analyzing the result of the experiment and the anticipated results are realized. This provides some possible solutions for the visual communication and accelerates the information business.