DI-Retinex: Digital-Imaging Retinex Theory for Low-Light Image Enhancement

TL;DR

This paper introduces DI-Retinex, a novel low-light image enhancement method based on an extended Retinex theory that accounts for digital imaging factors, using an unsupervised approach with improved visual quality and downstream task performance.

Contribution

It proposes a new Digital-Imaging Retinex theory with an offset term and non-linear mapping, along with unsupervised loss functions, advancing low-light image enhancement techniques.

Findings

Outperforms existing unsupervised methods in visual quality, size, and speed.

Enhances downstream face detection performance in low-light conditions.

Demonstrates effectiveness through extensive experiments.

Abstract

Many existing methods for low-light image enhancement (LLIE) based on Retinex theory ignore important factors that affect the validity of this theory in digital imaging, such as noise, quantization error, non-linearity, and dynamic range overflow. In this paper, we propose a new expression called Digital-Imaging Retinex theory (DI-Retinex) through theoretical and experimental analysis of Retinex theory in digital imaging. Our new expression includes an offset term in the enhancement model, which allows for pixel-wise brightness contrast adjustment with a non-linear mapping function. In addition, to solve the lowlight enhancement problem in an unsupervised manner, we propose an image-adaptive masked reverse degradation loss in Gamma space. We also design a variance suppression loss for regulating the additional offset term. Extensive experiments show that our proposed method outperforms all existing unsupervised methods in terms of visual quality, model size, and speed. Our algorithm can also assist downstream face detectors in low-light, as it shows the most performance gain after the low-light enhancement compared to other methods.