KinD-LCE Curve Estimation And Retinex Fusion On Low-Light Image

TL;DR

This paper introduces KinD-LCE, a novel low-light image enhancement algorithm that improves brightness and detail restoration by combining curve estimation, Retinex fusion, and noise reduction, enhancing downstream vision tasks.

Contribution

The paper proposes KinD-LCE, a new low-light enhancement method that integrates light curve estimation and Retinex fusion with a TV loss for noise reduction, addressing information loss in traditional methods.

Findings

Achieved PSNR of 19.72 and SSIM of 0.82 on low-light datasets.

Demonstrated improved performance in downstream tasks like object detection.

Effectively reduces noise and preserves details in low-light images.

Abstract

Low-light images often suffer from noise and color distortion. Object detection, semantic segmentation, instance segmentation, and other tasks are challenging when working with low-light images because of image noise and chromatic aberration. We also found that the conventional Retinex theory loses information in adjusting the image for low-light tasks. In response to the aforementioned problem, this paper proposes an algorithm for low illumination enhancement. The proposed method, KinD-LCE, uses a light curve estimation module to enhance the illumination map in the Retinex decomposed image, improving the overall image brightness. An illumination map and reflection map fusion module were also proposed to restore the image details and reduce detail loss. Additionally, a TV(total variation) loss function was applied to eliminate noise. Our method was trained on the GladNet dataset, known for its diverse collection of low-light images, tested against the Low-Light dataset, and evaluated using the ExDark dataset for downstream tasks, demonstrating competitive performance with a PSNR of 19.7216 and SSIM of 0.8213.