{"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/recurrent-squeeze-and-excitation-context","title":"Recurrent Squeeze-and-Excitation Context Aggregation Net for Single Image Deraining","arxiv_id":"1807.05698","date":"2018-07-16","proceeding":"ECCV 2018 9","authors":["Xia Li","Jianlong Wu","Zhouchen Lin","Hong Liu","Hongbin Zha"],"abstract":"Rain streaks can severely degrade the visibility, which causes many current\ncomputer vision algorithms fail to work. So it is necessary to remove the rain\nfrom images. We propose a novel deep network architecture based on deep\nconvolutional and recurrent neural networks for single image deraining. As\ncontextual information is very important for rain removal, we first adopt the\ndilated convolutional neural network to acquire large receptive field. To\nbetter fit the rain removal task, we also modify the network. In heavy rain,\nrain streaks have various directions and shapes, which can be regarded as the\naccumulation of multiple rain streak layers. We assign different alpha-values\nto various rain streak layers according to the intensity and transparency by\nincorporating the squeeze-and-excitation block. Since rain streak layers\noverlap with each other, it is not easy to remove the rain in one stage. So we\nfurther decompose the rain removal into multiple stages. Recurrent neural\nnetwork is incorporated to preserve the useful information in previous stages\nand benefit the rain removal in later stages. We conduct extensive experiments\non both synthetic and real-world datasets. Our proposed method outperforms the\nstate-of-the-art approaches under all evaluation metrics. Codes and\nsupplementary material are available at our project webpage:\nhttps://xialipku.github.io/RESCAN .","url_abs":"http://arxiv.org/abs/1807.05698v2","url_pdf":"http://arxiv.org/pdf/1807.05698v2.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":[],"tasks":[{"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":"RESCAN","rank_in_archive_order":14,"of":19,"metrics":{"PSNR":"26.36","SSIM":"0.786"},"uses_additional_data":false},{"leaderboard":"/sota/single-image-deraining-on-rain100l","task":"Single Image Deraining","dataset":"Rain100L","model":"RESCAN","rank_in_archive_order":18,"of":19,"metrics":{"SSIM":"0.881"},"uses_additional_data":false},{"leaderboard":"/sota/single-image-deraining-on-test100","task":"Single Image Deraining","dataset":"Test100","model":"RESCAN","rank_in_archive_order":9,"of":12,"metrics":{"SSIM":"0.835"},"uses_additional_data":false},{"leaderboard":"/sota/single-image-deraining-on-test1200","task":"Single Image Deraining","dataset":"Test1200","model":"RESCAN","rank_in_archive_order":13,"of":14,"metrics":{"SSIM":"0.882"},"uses_additional_data":false},{"leaderboard":"/sota/single-image-deraining-on-test2800","task":"Single Image Deraining","dataset":"Test2800","model":"RESCAN","rank_in_archive_order":7,"of":12,"metrics":{"PSNR":"31.29","SSIM":"0.904"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.05698","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}