Clustering-based fast intra prediction mode algorithm for HEVC

The High Efficiency Video Coding (HEVC) is the next generation video coding standard. The HEVC provides equivalent perceptual quality with bit-budget saving greater than 50% compared to the H.264/AVC. In this paper, we propose a new fast intra prediction mode decision algorithm for the HEVC. We apply an early termination method based on the statistics of the resulted IPMs from both rough mode decision and most probable mode stages. The resulted IPMs are clustered into a K cluster by means of the K-medoid clustering algorithm, and each IPM cluster center represents all the IPMs within a cluster for the RDO process. The sugeested algorithm has been evaluated on high resolution test video sequences. Compared with the current HM16.0 and state-of-the-art scheme in all intra high efficiency configuration cases, the proposed algorithm outperforms the state-of-the-art scheme in terms of encoding time with similar coding efficiency.