Abstract
Recently, Joint Video Experts Team (JVET) has started the standardization of a new video coding named as Versatile Video Coding (VVC) with the aim of completing by 2020. In HEVC and VVC, each candidate list of Merge/AMVP (Advanced Motion Vector Prediction) is constructed from the motion information of spatial/temporal neighboring blocks to enhance the performance of inter predictive coding by utilizing available motion information. In this paper, we propose a method to derive candidate lists of Merge/AMVP with more reliable motion information having high correlation by considering the block shape of the current Coding Unit (CU). Experiment results show the performance of the proposed method in the reference SW codec of VVC named as VVC Test Model (VTM).
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
