Channel-Agnostic Semantic Compression for Bandwidth-Limited Visual Communication

2026-07-31

AI Summary

This work addresses the challenges of semantic visual communication under bandwidth constraints and dynamic channel conditions, where existing approaches suffer from poor generalization due to tight coupling between source and channel coding or semantic inconsistency caused by generative reconstruction. To overcome these limitations, the authors propose the RQ-NAC framework, which achieves, for the first time, channel-agnostic semantic compression. The method employs residual quantization to produce scalable discrete semantic representations and integrates n-gram context modeling with arithmetic coding for highly efficient lossless entropy coding, enabling fine-grained rate-distortion trade-offs. Experimental results demonstrate that RQ-NAC achieves over 600× compression relative to raw visual data while preserving high perceptual quality, semantic consistency, flexible adaptability, and precise bitrate control.

Abstract

Bandwidth-limited visual communication systems require efficient transmission of high-dimensional data under dynamic wireless conditions. Existing approaches either rely on joint source-channel coding, which tightly couples representation learning with channel models and lacks flexibility across varying environments, or adopt generative reconstruction techniques that may introduce semantically inconsistent outputs. In this paper, we propose RQ-NAC, a channel-agnostic semantic compression framework for visual communication. The proposed method leverages residual quantization to produce scalable discrete semantic representations, enabling fine-grained and predictable control over the rate-distortion tradeoff. To further enhance compression efficiency, we integrate an n-gram-driven arithmetic coding module that exploits contextual dependencies among latent indices for lossless entropy coding. Extensive experiments demonstrate that RQ-NAC achieves over 600$\times$ compression relative to uncompressed visual data while preserving high perceptual quality. The results show that our approach enables efficient, flexible, and reliable semantic transmission under bandwidth-constrained conditions.

Problem

• semantic compression
• bandwidth-limited communication
• channel-agnostic
• visual data transmission
• rate-distortion tradeoff

Innovation

• semantic compression
• residual quantization
• channel-agnostic
• n-gram arithmetic coding
• visual communication