Gradient based fast mode decision algorithm for intra prediction in HEVC

The High Efficiency Video Coding (HEVC) standard achieves much better efficiency than previous video coding standards. One contributor to this improvement is the large set of intra prediction modes that it supports. While resulting in the better Rate-Distortion (R-D) performance compared to the H.264/AVC equivalent predictor, it suffers from heavier computation load. In this paper, a fast intra mode decision algorithm is proposed to speed up the original intra coding in HEVC. By analysing the gradient information of the current prediction unit (PU), only a small part of modes are selected as the candidate modes for the rate-distortion optimization (RDO) process. Two other performance tuning methods (SBS and SDBS) are also introduced in this paper. Experimental results show that, compared to the newest HEVC test model software HM 10.0, the proposed method provides a reduction of up to 42% in computational complexity with a small increase in the bit-rate and a negligible reduction in PSNR values.