A CNN-Based Prediction-Aware Quality Enhancement Framework for VVC

This paper presents a framework for Convolutional Neural Network (CNN)-based quality enhancement task, by taking advantage of coding information in the compressed video signal. The motivation is that normative decisions made encoder can significantly impact type and strength artifacts decoded images. In this paper, main focus has been put on defining prediction signal intra inter frames. used training phase as well input to help process learning are specific each type. Furthermore, retain low memory requirement proposed method, one model all Quantization Parameters (QPs) with Parameter (QP)-map, which also shared between luma chroma components. addition Post Processing (PP) approach, In-Loop Filtering (ILF) codec integration considered, where characteristics Group Pictures (GoP) taken into account boost performance. CNN-based Quality Enhancement (QE) implemented top Versatile Video Coding (VVC) Test Model (VTM-10). Experiments show prediction-aware aspect method improves efficiency gain default QE 1.52%, terms BD-BR, at same network complexity compared filter.