{"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/clearing-the-skies-a-deep-network","title":"Clearing the Skies: A deep network architecture for single-image rain removal","arxiv_id":"1609.02087","date":"2016-09-07","proceeding":null,"authors":["Xueyang Fu","Jia-Bin Huang","Xinghao Ding","Yinghao Liao","John Paisley"],"abstract":"We introduce a deep network architecture called DerainNet for removing rain\nstreaks from an image. Based on the deep convolutional neural network (CNN), we\ndirectly learn the mapping relationship between rainy and clean image detail\nlayers from data. Because we do not possess the ground truth corresponding to\nreal-world rainy images, we synthesize images with rain for training. In\ncontrast to other common strategies that increase depth or breadth of the\nnetwork, we use image processing domain knowledge to modify the objective\nfunction and improve deraining with a modestly-sized CNN. Specifically, we\ntrain our DerainNet on the detail (high-pass) layer rather than in the image\ndomain. Though DerainNet is trained on synthetic data, we find that the learned\nnetwork translates very effectively to real-world images for testing. Moreover,\nwe augment the CNN framework with image enhancement to improve the visual\nresults. Compared with state-of-the-art single image de-raining methods, our\nmethod has improved rain removal and much faster computation time after network\ntraining.","url_abs":"http://arxiv.org/abs/1609.02087v2","url_pdf":"http://arxiv.org/pdf/1609.02087v2.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":"clearing-the-skies-a-deep-network","repo_url":"https://github.com/jinnovation/rainy-image-dataset","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"clearing-the-skies-a-deep-network","repo_url":"https://github.com/MyNameIsPHP/DerainNet-Installation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"image-enhancement","task_name":"Image Enhancement"},{"task_slug":"rain-removal","task_name":"Rain Removal"},{"task_slug":"single-image-deraining","task_name":"Single Image Deraining"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/single-image-deraining-on-rain100h","task":"Single Image Deraining","dataset":"Rain100H","model":"DerainNet","rank_in_archive_order":18,"of":19,"metrics":{"SSIM":"0.592"},"uses_additional_data":false},{"leaderboard":"/sota/single-image-deraining-on-rain100l","task":"Single Image Deraining","dataset":"Rain100L","model":"DerainNet","rank_in_archive_order":17,"of":19,"metrics":{"SSIM":"0.884"},"uses_additional_data":false},{"leaderboard":"/sota/single-image-deraining-on-test100","task":"Single Image Deraining","dataset":"Test100","model":"DerainNet","rank_in_archive_order":12,"of":12,"metrics":{"SSIM":"0.810"},"uses_additional_data":false},{"leaderboard":"/sota/single-image-deraining-on-test1200","task":"Single Image Deraining","dataset":"Test1200","model":"DerainNet","rank_in_archive_order":14,"of":14,"metrics":{"SSIM":"0.835"},"uses_additional_data":false},{"leaderboard":"/sota/single-image-deraining-on-test2800","task":"Single Image Deraining","dataset":"Test2800","model":"DerainNet","rank_in_archive_order":11,"of":12,"metrics":{"PSNR":"24.31","SSIM":"0.861"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1609.02087","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}