Attention Guided Low-light Image Enhancement with a Large Scale Low-light Simulation Dataset

TL;DR

This paper introduces an attention-guided neural network approach for low-light image enhancement, utilizing a large synthetic dataset to improve brightness, noise reduction, and color correction, outperforming existing methods.

Contribution

The paper presents a novel end-to-end multi-branch neural network with attention guidance and a large-scale synthetic dataset for improved low-light image enhancement.

Findings

Outperforms state-of-the-art methods quantitatively.

Produces visually superior enhancement results.

Effective noise and color correction in low-light images.

Abstract

Low-light image enhancement is challenging in that it needs to consider not only brightness recovery but also complex issues like color distortion and noise, which usually hide in the dark. Simply adjusting the brightness of a low-light image will inevitably amplify those artifacts. To address this difficult problem, this paper proposes a novel end-to-end attention-guided method based on multi-branch convolutional neural network. To this end, we first construct a synthetic dataset with carefully designed low-light simulation strategies. The dataset is much larger and more diverse than existing ones. With the new dataset for training, our method learns two attention maps to guide the brightness enhancement and denoising tasks respectively. The first attention map distinguishes underexposed regions from well lit regions, and the second attention map distinguishes noises from real textures. With their guidance, the proposed multi-branch decomposition-and-fusion enhancement network works in an input adaptive way. Moreover, a reinforcement-net further enhances color and contrast of the output image. Extensive experiments on multiple datasets demonstrate that our method can produce high fidelity enhancement results for low-light images and outperforms the current state-of-the-art methods by a large margin both quantitatively and visually.