Unpaired Image Enhancement with Quality-Attention Generative Adversarial Network

TL;DR

This paper introduces QAGAN, a novel unpaired image enhancement model that uses a quality attention module within a GAN framework to improve low-quality images by learning domain-specific quality features.

Contribution

The paper proposes a new quality attention module integrated into a bidirectional GAN for unpaired image enhancement, enabling the generator to focus on semantic and style-related features from both domains.

Findings

Outperforms state-of-the-art unpaired enhancement methods in objective metrics.

Achieves superior subjective visual quality in enhanced images.

Demonstrates effective domain-relevant feature learning through the quality attention module.

Abstract

In this work, we aim to learn an unpaired image enhancement model, which can enrich low-quality images with the characteristics of high-quality images provided by users. We propose a quality attention generative adversarial network (QAGAN) trained on unpaired data based on the bidirectional Generative Adversarial Network (GAN) embedded with a quality attention module (QAM). The key novelty of the proposed QAGAN lies in the injected QAM for the generator such that it learns domain-relevant quality attention directly from the two domains. More specifically, the proposed QAM allows the generator to effectively select semantic-related characteristics from the spatial-wise and adaptively incorporate style-related attributes from the channel-wise, respectively. Therefore, in our proposed QAGAN, not only discriminators but also the generator can directly access both domains which significantly facilitates the generator to learn the mapping function. Extensive experimental results show that, compared with the state-of-the-art methods based on unpaired learning, our proposed method achieves better performance in both objective and subjective evaluations.