An efficient low-complexity block partition scheme for VVC intra coding

The newest versatile video coding (VVC) adopts a novel quadtree with a nested multi-type tree (QTMT) partition structure for intra-frame coding and splits the coding unit (CU) into not only square sub-blocks but also rectangular sub-blocks. The more flexible CU sizes improve the encoding efficiency while cause much heavier computational complexity than High-Efficiency Video Coding (HEVC). In this paper, an efficient low-complexity CU partition method based on the texture characteristic is proposed to achieve a good trade-off between the coding efficiency and the complexity reduction. Specifically, the texture complexities of vertical and horizontal directions are quantitatively measured in terms of the sum of the mean absolute deviation (SMAD) of the sub-blocks. Then, the vertical and horizontal texture complexities are compared to eliminate some unlikely partition modes. Moreover, the threshold of directional SMAD ratios for choosing the texture direction is adjusted according to the quantization parameter (QP) to improve the prediction accuracy. Experimental results show that the proposed method achieves the average complexity reduction efficiency by 54.60% than the VVC test model (VTM). The implementation of the proposed method is publicly available at http://sites.google.com/site/wangmiaohui/ and http://github.com/csust-sonie/fastVVC_RTIP_2107 .