Research On Integer Transform And Motion Estimation Of Video Coding Section Of AVS

In the end of February 2006, AVS (Audio Video coding Standard) was officially approved as a national standard, which is of great significance for China's digital audio and video industry development. Although the pace of AVS industrialization has already begun since the enacting-standard, so far it is in start-up period of the large-scale industrialization, the real industrialization have not come yet, the research on the key techniques of AVS standard has important theoretical value and practical significance for the development of AVS industrialization.In this paper, around AVS video coding standard, on the base of in-depth analysis of AVS video encoding principle and key techniques, the integral transform all-zero blocks detection technique and motion estimation technique have been researched in detail. These two techniques play a decisive role for improving video compression coding efficiency.The main contents in this paper can be summarized as the following three aspects:(1) The development of video standard is introduced, then the key techniques of AVS video encoding standard are analyzed, and it is pointed out that the research of AVS standard is of great significance for the development of China's audio and video industrialization. (2) A new all-zero blocks criteria which is suitable for AVS standard is deduced. The all-zero blocks detection technology in H.264 standard has been discussed in depth, but the use of 8x8 block in AVS Integer Transform and the use of quantizer which adopted the new exponent rank quantitization step, the past all-zero blocks criteria suitable to H.264 can not be directly applied to AVS. According to the features of AVS Integer Transform, a new all-zero blocks criteria suitable for the AVS has been deduced. It has a better image quality, and preventes miss-judgment largely.(3) On the basis of the analysis of existing typical motion estimation algorithms, a predictive adaptive Square-Diamond search algorithm using successive elimination (PA-SEA-SDS) is proposed. Based on the cross-center-biased motion vector distribution characteristic of the image sequences and high space-time correlation of adjacent blocks' motion vectors, Square-Diamond search (SDS) pattern is adopted and it is designed that Square-Diamond search strategy based on successive elimination, with such effective techniques as prediction of initial search point and adaptive search modes. Experiments show that this algorithm has a better search precision, and largely reduces search points and improves the speed of motion estimation.