An Improved Method of Low Light Image Enhancement Based on Retinex
Qi Wu, Maoling Qin, Jingqi Song, Li Liu · 2021
Low light image enhancement is a highly challenging task, which has received major attention over the decades. Inspired by Retinex Net [1], this paper improves low light image enhancement by mimicking the mechanism of Retinex. The model consists of a Decom-RNet (Decomposition Residual Net) for image decomposition and an Enhance-RNet (Enhancement Residual Net) for illumination enhancement and adjustment. Decom-RNet decomposes the image into illumination and reflection, and Enhance-RNet enhances the illumination. In particular, we introduce an attention module into residual learning, which significantly improves the accuracy of residual learning. Experimental results demonstrate that our method outperforms the state-of-the-art low light image enhancement methods.