A Fast Algorithm for Coding Depth Decision for VVC Intra-prediction

To reduce the coding complexity of the latest video coding standard—Versatile Video Coding (H.266/VVC), this paper proposes a fast algorithm for CU coding depth decision. We analyzed the CU division process of VVC intra-prediction for 360° panoramic video. On the one hand, according to the sampling characteristics of 360° panoramic video in ERP format, the region is divided, and according to the average weight sum of each CTU, the region where the current CTU is located is determined, and different coding depth intervals are assigned to it. On the other hand, the image features of video are analyzed, and the number of edge pixels is used as the standard to measure the texture complexity. The Sobel operator is used to calculate and count the number of edge pixels of each CTU, and the texture complexity is classified by a real-time updating double-threshold algorithm, so as to further control the interval size of coding depth. The experimental results show that in the all intra modes, compared with the VVC Test Model 11.0, the coding time of the method in this paper is reduced by 15.45% on average, the BDPSNR is reduced by 0.058 dB, and the Bjøntegaard Delta Bitrate (BDBR) is increased by 1.31%.