Motion-aware VVC Inter prediction Enhancement

In the next generation video coding standard VVC (Versatile Video Coding), inter-frame prediction is a central tool for removing temporal redundancy information. In this paper, we propose a motion-aware deep reference frame generation algorithm. Specifically, the reconstructed frames are fed into a trained motion-aware frame interpolation network to generate deep reference frame that match the visual perception characteristics of the human eye and have higher temporal correlation with the currently encoded frame. The deep reference frame is inserted into the reference frame list to provide a more reliable reference for subsequent motion estimation and motion compensation. The experimental results show that the method can save 1.56%, 1.91%, and 1.90% in BD-rate for Y, U, and V components in random access configuration compared with the VVC reference software VTM-15.0, respectively.