Differentiable Vector Quantization for Rate-Distortion Optimization of Generative Image Compression

TL;DR

RDVQ introduces a differentiable, end-to-end optimized vector quantization framework for ultra-low bitrate image compression, achieving high perceptual quality with fewer parameters.

Contribution

It presents a novel differentiable relaxation of the codebook distribution enabling joint rate-distortion optimization in VQ-based image compression.

Findings

Achieves up to 75.71% bitrate reduction on DISTS metric.

Attains competitive or superior perceptual quality at extremely low bitrates.

Uses a lightweight architecture with significantly fewer parameters.

Abstract

The rapid growth of visual data under stringent storage and bandwidth constraints makes extremely low-bitrate image compression increasingly important. While Vector Quantization (VQ) offers strong structural fidelity, existing methods lack a principled mechanism for joint rate-distortion (RD) optimization due to the disconnect between representation learning and entropy modeling. We propose RDVQ, a unified framework that enables end-to-end RD optimization for VQ-based compression via a differentiable relaxation of the codebook distribution, allowing the entropy loss to directly shape the latent prior. We further develop an autoregressive entropy model that supports accurate entropy modeling and test-time rate control. Extensive experiments demonstrate that RDVQ achieves strong performance at extremely low bitrates with a lightweight architecture, attaining competitive or superior perceptual quality with significantly fewer parameters. Compared with RDEIC, RDVQ reduces bitrate by up to 75.71% on DISTS and 37.63% on LPIPS on DIV2K-val. Beyond empirical gains, RDVQ introduces an entropy-constrained formulation of VQ, highlighting the potential for a more unified view of image tokenization and compression. The code will be available at https://github.com/CVL-UESTC/RDVQ.