The next generation Audio Video Coding Standard (AVS4) introduces enhanced intra prediction tools to further improve spatial coding efficiency. To improve directional adaptability for luma prediction and enhance cross-component correlation modeling for chroma, this paper proposes an intra prediction mode optimization method based on directional enhancement and cross-component fusion.

This paper aims to reduce the complexity of encoding screen content videos using Versatile Video Coding (VVC). To address the significant increase in encoding time caused by Screen Content Coding (SCC) tools in VVC, we propose a fast intra mode decision algorithm based on Adaptive Multi-Size Convolutional Neural Network (AMS-CNN).