{"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/dehazenet-an-end-to-end-system-for-single","title":"DehazeNet: An End-to-End System for Single Image Haze Removal","arxiv_id":"1601.07661","date":"2016-01-28","proceeding":null,"authors":["Bolun Cai","Xiangmin Xu","Kui Jia","Chunmei Qing","DaCheng Tao"],"abstract":"Single image haze removal is a challenging ill-posed problem. Existing\nmethods use various constraints/priors to get plausible dehazing solutions. The\nkey to achieve haze removal is to estimate a medium transmission map for an\ninput hazy image. In this paper, we propose a trainable end-to-end system\ncalled DehazeNet, for medium transmission estimation. DehazeNet takes a hazy\nimage as input, and outputs its medium transmission map that is subsequently\nused to recover a haze-free image via atmospheric scattering model. DehazeNet\nadopts Convolutional Neural Networks (CNN) based deep architecture, whose\nlayers are specially designed to embody the established assumptions/priors in\nimage dehazing. Specifically, layers of Maxout units are used for feature\nextraction, which can generate almost all haze-relevant features. We also\npropose a novel nonlinear activation function in DehazeNet, called Bilateral\nRectified Linear Unit (BReLU), which is able to improve the quality of\nrecovered haze-free image. We establish connections between components of the\nproposed DehazeNet and those used in existing methods. Experiments on benchmark\nimages show that DehazeNet achieves superior performance over existing methods,\nyet keeps efficient and easy to use.","url_abs":"http://arxiv.org/abs/1601.07661v2","url_pdf":"http://arxiv.org/pdf/1601.07661v2.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":"dehazenet-an-end-to-end-system-for-single","repo_url":"https://github.com/muditjoshi98/Image-Dehazing-Using-Residual-Based-Deep-CNN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"dehazenet-an-end-to-end-system-for-single","repo_url":"https://github.com/saintazunya/DehazeNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"dehazenet-an-end-to-end-system-for-single","repo_url":"https://github.com/caibolun/DehazeNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"image-dehazing","task_name":"Image Dehazing"},{"task_slug":"single-image-haze-removal","task_name":"Single Image Haze Removal"}],"methods":[{"method_slug":"maxout","method_name":"Maxout"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-dehazing-on-rs-haze","task":"Image Dehazing","dataset":"RS-Haze","model":"DehazeNet","rank_in_archive_order":7,"of":7,"metrics":{"PSNR":"23.16","SSIM":"0.816"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1601.07661","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}