High-Fidelity Generative Image Compression

TL;DR

This paper presents a novel generative image compression method combining GANs and learned compression techniques, achieving high-quality, perceptually similar reconstructions across various bitrates and high resolutions, outperforming previous methods.

Contribution

It introduces a new approach that integrates GANs with learned compression, enabling visually pleasing, high-resolution image reconstructions at multiple bitrates, bridging theory and practical application.

Findings

Produces perceptually similar images to input

Operates effectively across a broad range of bitrates

Preferred over previous methods in user studies

Abstract

We extensively study how to combine Generative Adversarial Networks and learned compression to obtain a state-of-the-art generative lossy compression system. In particular, we investigate normalization layers, generator and discriminator architectures, training strategies, as well as perceptual losses. In contrast to previous work, i) we obtain visually pleasing reconstructions that are perceptually similar to the input, ii) we operate in a broad range of bitrates, and iii) our approach can be applied to high-resolution images. We bridge the gap between rate-distortion-perception theory and practice by evaluating our approach both quantitatively with various perceptual metrics, and with a user study. The study shows that our method is preferred to previous approaches even if they use more than 2x the bitrate.