Zero-LEINR: Zero-Reference Low-light Image Enhancement with Intrinsic Noise Reduction

Zero-reference deep learning-based methods for low-light image enhancement sufficiently mitigate the difficulty of paired data collection while keeping the great generalization on various lighting conditions. However, color bias and unin-tended intrinsic noise amplification are still issues that remain unsolved. This paper proposes a zero-reference end-to-end two-stage network (Zero-LEINR) for low-light image enhancement with intrinsic noise reduction. In the first stage, we introduce a Color Preservation and Light Enhancement Block (CPLEB) that consists of a dual branch structure with different constraints to correct the brightness and preserve the correct color tone. In the second stage, Enhanced-Noise Reduction Block (ENRB) is applied to remove the intrinsic noises being enhanced during the first stage. Due to the zero-reference two-stage structure, our method is generalized to enhance low-light images with correct color tone on unseen datasets and reduce the intrinsic noise simultaneously.

I. Introduction

Images captured under poor lighting conditions usually suffer from poor visibility, unexpected noise, and less color information. Apart from poor perceptual quality, these images also affect the performance of computer vision tasks such as segmentation, object detection, and recognition. Some conventional approaches have been proposed to resolve the issues. Histogram Equalization-based methods [1], [2] enhance the image by expanding the dynamic range. Retinex-based methods [3]–[5] assume an image can be decomposed into the pixel-wise product of reflectance and illumination, and the reflectance component is consistent under any lighting conditions. Therefore, an enhanced image can be obtained by estimating the illumination map. Although these traditional approaches are not sufficient to enhance the low -light image, they do not rely on a large amount of training data.