A Novel Fast Mode Decision Algorithm for AVS2 Intra Coding

AVS2 is the latest generation of video coding standard independently developed by China. Compared with the most popular international standard HEVC/H.265, AVS2 achieves the same level of compression efficiency, which doubled in compression efficiency compared to the previous generation coding standard AVC/H.264 or AVS. However, the increased compression efficiency is accompanied by a significant increase in coding complexity. In order to improve the coding speed, this paper proposes a fast intra mode decision algorithm based on texture information and neighbor spatial-temporal information. Firstly, the Sobel operator to obtain the gradient information of the current PU (Prediction Unit) is utilized to determine available candidate intra predicted modes. Furthermore, the predicted candidate modes are updated by the most probable modes (MPMs) obtained from the neighbor spatial-temporal correlation encoded blocks. Experimental results are implemented on RD17.0, and about 36.03% encoding time is reduced on average with negligible coding performance loss.