Simple approach to reduce the number of selected intra luminance prediction modes for RDO computation in AVS2

The second generation Audio-video Coding Standard (AVS2) is the latest video standard currently under the development of Audio-video Coding Standard (AVS) Workgroup of China for achieving higher coding efficiency than previous (AVS1). In intra luminance prediction of AVS2, the AVS2 video encoder has to compute the rate-distortion optimization (RDO) of 9 selected candidate prediction modes in order to determine the best luminance prediction mode, which is a time-consuming work. In this paper, a simple approach is proposed for reducing the number of candidate prediction modes to equals or less than 6, where the video encoding experiments based on Reference Design (RD) 14.0 program with the AH-Intra configuration have demonstrated that the proposed method can averagely decrease the video encoding time by 29.37%, at a cost of Peak Signal to Noise Ratio (PSNR) degradation of 0.05dB.