Denoising Diffusion Post-Processing for Low-Light Image Enhancement

Abstract

Low-light image enhancement (LLIE) techniques attempt to increase the visibility of images captured in low-light scenarios. However, as a result of enhancement, a variety of image degradations such as noise and colour bias are revealed. Furthermore, each particular LLIE approach may introduce a different form of flaw within its enhanced results. To combat these image degradations, post-processing denoisers have widely been used, which often yield oversmoothed results lacking detail. In this work, a diffusion model post-processing approach is proposed, and the Low-light Post-processing Diffusion Model (LPDM) is introduced in order to model the conditional distribution between under-exposed and normally-exposed images. The LPDM is applied in a manner which avoids the computationally expensive generative reverse process of typical diffusion models, and is able to post-process images in one pass through LPDM. Extensive experiments demonstrate that the proposed approach outperforms competing post-processing denoisers by increasing the