HDCompression: Hybrid-Diffusion Image Compression for Ultra-Low Bitrates

TL;DR

HDCompression introduces a hybrid framework combining diffusion models, VQ-modeling, and LIC to achieve high-fidelity, perceptually pleasing image compression at ultra-low bitrates, outperforming existing methods.

Contribution

This work proposes a novel hybrid image compression method that leverages diffusion models for enhanced fidelity and combines it with VQ-modeling and LIC, improving performance at ultra-low bitrates.

Findings

Outperforms previous LIC, VQ-modeling, and hybrid methods in metrics and visualization.

Enhances fidelity and perceptual quality at ultra-low bitrates.

Uses lightweight diffusion models with simple sampling schedules.

Abstract

Image compression under ultra-low bitrates remains challenging for both conventional learned image compression (LIC) and generative vector-quantized (VQ) modeling. Conventional LIC suffers from severe artifacts due to heavy quantization, while generative VQ modeling gives poor fidelity due to the mismatch between learned generative priors and specific inputs. In this work, we propose Hybrid-Diffusion Image Compression (HDCompression), a dual-stream framework that utilizes both generative VQ-modeling and diffusion models, as well as conventional LIC, to achieve both high fidelity and high perceptual quality. Different from previous hybrid methods that directly use pre-trained LIC models to generate low-quality fidelity-preserving information from heavily quantized latent, we use diffusion models to extract high-quality complementary fidelity information from the ground-truth input, which can enhance the system performance in several aspects: improving index map prediction, enhancing the fidelity-preserving output of the LIC stream, and refining conditioned image reconstruction with VQ-latent correction. In addition, our diffusion model is based on a dense representative vector (DRV), which is lightweight with very simple sampling schedulers. Extensive experiments demonstrate that our HDCompression outperforms the previous conventional LIC, generative VQ-modeling, and hybrid frameworks in both quantitative metrics and qualitative visualization, providing balanced robust compression performance at ultra-low bitrates.