{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/deep-retinex-decomposition-for-low-light","title":"Deep Retinex Decomposition for Low-Light Enhancement","arxiv_id":"1808.04560","date":"2018-08-14","proceeding":null,"authors":["Chen Wei","Wenjing Wang","Wenhan Yang","Jiaying Liu"],"abstract":"Retinex model is an effective tool for low-light image enhancement. It\nassumes that observed images can be decomposed into the reflectance and\nillumination. Most existing Retinex-based methods have carefully designed\nhand-crafted constraints and parameters for this highly ill-posed\ndecomposition, which may be limited by model capacity when applied in various\nscenes. In this paper, we collect a LOw-Light dataset (LOL) containing\nlow/normal-light image pairs and propose a deep Retinex-Net learned on this\ndataset, including a Decom-Net for decomposition and an Enhance-Net for\nillumination adjustment. In the training process for Decom-Net, there is no\nground truth of decomposed reflectance and illumination. The network is learned\nwith only key constraints including the consistent reflectance shared by paired\nlow/normal-light images, and the smoothness of illumination. Based on the\ndecomposition, subsequent lightness enhancement is conducted on illumination by\nan enhancement network called Enhance-Net, and for joint denoising there is a\ndenoising operation on reflectance. The Retinex-Net is end-to-end trainable, so\nthat the learned decomposition is by nature good for lightness adjustment.\nExtensive experiments demonstrate that our method not only achieves visually\npleasing quality for low-light enhancement but also provides a good\nrepresentation of image decomposition.","url_abs":"http://arxiv.org/abs/1808.04560v1","url_pdf":"http://arxiv.org/pdf/1808.04560v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"deep-retinex-decomposition-for-low-light","repo_url":"https://github.com/bchao1/awesome-image-enhancement","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"deep-retinex-decomposition-for-low-light","repo_url":"https://github.com/weichen582/RetinexNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"deep-retinex-decomposition-for-low-light","repo_url":"https://github.com/zhuangyunliang/dpfnet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"image-enhancement","task_name":"Image Enhancement"},{"task_slug":"low-light-image-enhancement","task_name":"Low-Light Image Enhancement"}],"methods":[],"datasets_introduced":[{"slug":"lol","name":"LOL","full_name":"LOw-Light dataset"}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.04560","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}