L-Shape-Based Iterative Prediction and Residual Median Edge Detection (LIP-RMED) Algorithm for VVC Intra-Frame Lossless Compression

Abstract

Versatile Video Coding (VVC) is the latest video coding standard beyond High Efficiency Video Coding (HEVC), offering higher compression efficiency. The algorithm in this letter called L-shape-based iterative prediction and residual median edge detection (LIP-RMED) divides each block into a series of close neighboring 1-D inverted L-shapes. They are employed as reference samples in the subsequent recursively L-shape prediction. Because of the inherited close neighboring characteristics of L-shapes, LIP can achieve much more accurate prediction results. To further maximize the efficiency of our prediction method, this letter also proposes a residual median edge detection (RMED) method, which allows a second process over the residual that still possesses rich edge information. Experimental results on VVC Test Model (VTM) 12.3 show that our proposed merge method is able to achieve 5.43% average bit-rate savings, with a maximum of 13.67% bit-rate savings on VVC intra-frame lossless compression.