Learning a Single Convolutional Layer Model for Low Light Image Enhancement

Low-light image enhancement (LLIE) aims to improve the illuminance of images due to insufficient light exposure. Recently, various lightweight learning-based LLIE methods have been proposed to handle the challenges of unfavorable prevailing low contrast, low brightness, etc. In this paper, we have streamlined the architecture of the network to the utmost degree. By utilizing the effective structural re-parameterization technique, a single convolutional layer model (SCLM) is proposed that provides global low-light enhancement as the coarsely enhanced results. In addition, we introduce a local adaptation module that learns a set of shared parameters to accomplish local illumination correction to address the issue of varied exposure levels in different image regions. Experimental results demonstrate that the proposed method performs favorably against the state-of-the-art LLIE methods in both objective metrics and subjective visual effects. Additionally, our method has fewer parameters and lower inference complexity compared to other learning-based schemes. Code will be made publicly available at the URL https://gitee.com/zhanghahaxixi/SCLM

I. Introduction

Due to environment or equipment limitations, images taken under low lighting conditions always result in poor pictures with severe noise, low contrast, and many other problems. Improving the perceptual quality of such low-light images has been a long-standing issue. Traditional solutions include histogram curve adjustment methods [1], [2], [3] and Retinex-based methods [4], [5], [6]. Although hand-crafted constraints or priors are helpful in improving the quality of the low-light image, the enhanced output always suffers from over- or under-enhancement in local regions. In recent years, with the surge of deep learning, various data-driven methods have been proposed to tackle this problem [7], [8].