RSEND: Retinex-based Squeeze and Excitation Network with Dark Region Detection for Efficient Low Light Image Enhancement

TL;DR

RSEND is a novel, efficient one-stage Retinex-based deep learning framework that enhances low-light images by effectively separating illumination and reflectance, utilizing Squeeze and Excitation networks for detail capture, and outperforming existing models in quality and efficiency.

Contribution

The paper introduces RSEND, a concise, one-stage Retinex-based network with Squeeze and Excitation modules for low-light image enhancement, reducing complexity and improving performance over prior multi-stage CNN methods.

Findings

RSEND achieves up to 4.2 dB PSNR improvement on various datasets.

RSEND outperforms CNN-based models and surpasses transformer-based models on LOL-v2-real.

The method is more efficient and simpler than previous multi-stage approaches.

Abstract

Images captured under low-light scenarios often suffer from low quality. Previous CNN-based deep learning methods often involve using Retinex theory. Nevertheless, most of them cannot perform well in more complicated datasets like LOL-v2 while consuming too much computational resources. Besides, some of these methods require sophisticated training at different stages, making the procedure even more time-consuming and tedious. In this paper, we propose a more accurate, concise, and one-stage Retinex theory based framework, RSEND. RSEND first divides the low-light image into the illumination map and reflectance map, then captures the important details in the illumination map and performs light enhancement. After this step, it refines the enhanced gray-scale image and does element-wise matrix multiplication with the reflectance map. By denoising the output it has from the previous step, it obtains the final result. In all the steps, RSEND utilizes Squeeze and Excitation network to better capture the details. Comprehensive quantitative and qualitative experiments show that our Efficient Retinex model significantly outperforms other CNN-based models, achieving a PSNR improvement ranging from 0.44 dB to 4.2 dB in different datasets and even outperforms transformer-based models in the LOL-v2-real dataset.