Yi-Jen Chiu

Self-derivation of motion estimation techniques to improve video coding efficiency

This paper presents the techniques to self derive the motion vectors (MVs) at video decoder side to improve coding efficiency of B pictures. With the MVs information self derived at video decoder side, the transmission of these self-derived MVs from video encoder side to video decoder side is skipped and thus better coding efficiency can be achieved. Our proposed techniques derive the block-based MVs at video decoder side by considering the temporal correlation among the available pixels in the previously-decoded reference pictures. Utilizing the MVs derived at video decoder side can be added as one of coding mode candidates from video encoder where the video encoder can utilize this new coding mode during phase of the coding mode selection to better trade off the rate-distortion performance to improve the coding efficiency. Experiments have demonstrated that the BD bitrate improvement on top of ITU-T/VCEG Key Technology Area (KTA) Reference Software platform with an overall about 7% improvement on the hierarchical IbBbBbBbP coding structure under the common test conditions of the joint call for proposal for the new video coding technology from ISO/MPEG and ITU-T committee on January 2010.

Decoder-side Motion Estimation and Wiener filter for HEVC

This paper presents the coding techniques of Decoder-side Motion Vector Derivation (DMVD) and Adaptive Loop Filter (ALF) to apply on the state-of-the-art High Efficiency Video Coding (HEVC) standard. Both DMVD and ALF can provide the improvement of coding efficiency on HEVC test Model (HM) reference software, which are reported in multiple proposals during the development period of HEVC. In DMVD algorithm, Motion Vectors (MVs) are derived at video decoder side by utilizing the spatial and temporal correlation among decoded pictures without the signaling of motion vectors from video encoder side. The transmission of MVs from video encoder is skipped and thus better coding efficiency of inter prediction can be achieved. In ALF algorithm, a group of adaptive Wiener filters are applied to the samples of reconstructed pictures to improve the output video visual quality. The filter coefficients can be trained by video encoder and will be conveyed to video decoder signaled by the slice header to apply for the entire pixels in the slice. The utilization of DMVD and ALF can be controlled by the process of coding mode decision at video encoder side based on Rate-Distortion Optimization (RDO) metric. Experiments have demonstrated that the DMVD achieves 0.6%-1.1% BD rate reduction when compared to HM10.0 on the Random Access (RA) test condition. For ALF, average BD rate savings of 1.5%-2.6% are observed compared to HM10.0 under the HM common test condition. Experiment results also demonstrate that the coding gain from the integrated pipe of DMVD and ALF can be additive.