Bosman, Anna2023-12-142023-12-142024-042023-09*A2024http://hdl.handle.net/2263/93790Dissertation (MSc(Computer Science))--University of Pretoria, 2023.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 theen© 2023 University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria.UCTDDiffusion modelDenoisingLow-Light Image EnhancementPost-ProcessingNeural NetworksComputer VisionDissertationu1721528610.25403/UPresearchdata.24634884