Regressive rate-distortion trade-off with weighted entropy coding for HEVC encoding

Abstract

High-Efficiency Video Coding (HEVC) is the standard employed for the subsequent compressing of the video without degrading the quality of the image. HEVC renders effective performance compared to the existing compression standards as the encoding time is minimal. With the concern for quality and video compression, this paper proposes a modified version for HEVC encoding using the Ordered Tree-based Hex-Octagon based block Search and Rate-Distortion trade-off (OrTHO-Search-based RD) for motion estimation and weighted Context-Adaptive Binary Arithmetic Coding (CABAC). The proposed OrTHO-Search-based RD-dependent HEVC renders a good-quality video after compression with half the compression standard when compared with the other existing compression standards. In the motion estimation block, the OrTHO-Search is employed with a new RD trade-off in such a way that the Conditional Autoregressive Value at Risk concept modifies the existing RD trade-off measure. The bit rates of the proposed method are reduced with effective coding. The experimentation and analysis of the methods are performed using four videos from CIPR SIF Sequences and one video from Xiph.org Video Test Media datasets and the analysis based on Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), reveals that the proposed method acquired the maximal PSNR of 45.1132 dB, maximal SSIM of 0. 9918, and minimal computational time of 0.2135 min.