International Conference on Computer and Knowledge Engineering

Home / 15th International Conference on Computer and Knowledge Engineering

VVC-AAR: Adaptive Attention-Aware Resolution and Residual Coding for Perceptually Optimized Ultra-Low Bitrate VVC Compression

Authors :

Yaghoub Saberi¹ Somayeh Arab Najafabadi² Mohammadreza Hemmati³

1- Department of Computer Engineering Na.C. , Islamic Azad University, Najafabad, Iran 2- Department of Computer Engineering Na.C. , Islamic Azad University, Najafabad, Iran 3- Department of Computer Engineering Na.C. , Islamic Azad University, Najafabad, Iran

Keywords :

VVC،Video Compression،Attention Map،Neural Residual Coding،Perceptual Quality،Adaptive Resolution،Low Bitrate Streaming

Abstract :

this paper presents VVC-AAR, a saliency-guided resolution adaptation framework designed to enhance the coding efficiency of the Versatile Video Coding (VVC) standard. The proposed method employs an importance map derived from visual salience cues to classify video regions into multiple perceptual importance levels. High-importance regions are encoded at full resolution with low quantization parameters. Low-importance areas are down-sampled and compressed more aggressively. They are later refined through neural-based enhancement to restore visual quality. This adaptive strategy enables the encoder to achieve more effective rate allocation according to perceptual relevance. Experimental results on standard test sequences and high-resolution datasets demonstrate that VVC-AAR achieves significant bitrate savings up to 35% over the VVC anchor (VTM 12.0) at comparable or improved perceptual quality, as measured by VMAF. These findings confirm that perception-driven compression with resolution adaptation can provide substantial coding gains without compromising subjective visual experience.