{"id":23136691,"url":"https://github.com/naoto0804/synshadow","last_synced_at":"2025-08-17T10:32:20.354Z","repository":{"id":46161708,"uuid":"322502933","full_name":"naoto0804/SynShadow","owner":"naoto0804","description":"Learning from Synthetic Shadows for Shadow Detection and Removal [Inoue and Yamasaki, IEEE TCSVT 2021].","archived":false,"fork":false,"pushed_at":"2024-03-16T04:51:06.000Z","size":1054,"stargazers_count":78,"open_issues_count":0,"forks_count":13,"subscribers_count":5,"default_branch":"main","last_synced_at":"2024-03-16T05:29:39.106Z","etag":null,"topics":["pytorch","shadow-detection","shadow-removal"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/naoto0804.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null}},"created_at":"2020-12-18T05:56:44.000Z","updated_at":"2024-03-07T06:31:34.000Z","dependencies_parsed_at":"2024-03-22T12:01:08.789Z","dependency_job_id":null,"html_url":"https://github.com/naoto0804/SynShadow","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/naoto0804%2FSynShadow","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/naoto0804%2FSynShadow/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/naoto0804%2FSynShadow/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/naoto0804%2FSynShadow/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/naoto0804","download_url":"https://codeload.github.com/naoto0804/SynShadow/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":230115975,"owners_count":18175666,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["pytorch","shadow-detection","shadow-removal"],"created_at":"2024-12-17T12:25:29.059Z","updated_at":"2024-12-17T12:25:29.640Z","avatar_url":"https://github.com/naoto0804.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Learning from Synthetic Shadows for Shadow Detection and Removal (IEEE TCSVT 2021)\n\n## Overview\nThis repo is for the paper \"[Learning from Synthetic Shadows for Shadow Detection and Removal](https://arxiv.org/abs/2101.01713)\". We present SynShadow, a novel large-scale synthetic shadow/shadow-free/matte image triplets dataset and pipeline to synthesize it. We further show how to use SynShadow for robust and efficient shadow detection and removal.\n\n![](teaser.png)\n\nIn this repo, we provide\n- SynShadow dataset: `./datasets`\n- [SP+M](https://arxiv.org/abs/1908.08628) implementation: `./src`\n- Trained models and results: below\n\nIf you find this code or dataset useful for your research, please cite our paper:\n\n```\n@article{inoue2021learning,\n  title={{Learning from Synthetic Shadows for Shadow Detection and Removal}},\n  author={Inoue, Naoto and Yamasaki, Toshihiko},\n  journal={IEEE Transactions on Circuits and Systems for Video Technology},\n  year={2021},\n  volume={31},\n  number={11},\n  pages={4187-4197},\n  doi={10.1109/TCSVT.2020.3047977}\n}\n```\n\n\n## Trained Models and Results\nWe provide the models for shadow detection and removal for convenience. Downloaded models should be placed under `./checkpoints`.\n\n### Shadow Detection\nALl the results are in 480x640. BER is reported for 480x640 images. Below are results evaluated on ISTD test set. DSDNet++ is a modified variant of [DSDNet](https://openaccess.thecvf.com/content_CVPR_2019/html/Zheng_Distraction-Aware_Shadow_Detection_CVPR_2019_paper.html).\n\n| Model | Train | BER |      |\n|  :-:  |   :-:    | :-: | :-:  |\n|DSDNet++|SynShadow|2.74|[results](https://drive.google.com/file/d/1jhVXHSHj4teacLTK_I-Ep4WNDxId-_tp/view?usp=drive_link) / [weights](https://drive.google.com/file/d/13IhP8d5Tb7ZM3kgu0HmgAvSxAS0KMtXZ/view?usp=drive_link)|\n|DSDNet++|SynShadow-\u003eISTD|1.09|[results](https://drive.google.com/file/d/14Ni8fEoQJNj1xz2e_JD1zUKSWwuuwcxc/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1CQ6KjDlKesdb86tJnVcX79B3br_w5jHt/view?usp=drive_link)|\n|BDRAR|SynShadow|2.74|[results](https://drive.google.com/file/d/1IHTf0901Mzm5a2G1sROfAyJYoddFNVbF/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1y-VQBoclNzPsfBZovMwpymlVb4I3ljVy/view?usp=drive_link)|\n|BDRAR|SynShadow-\u003eISTD|1.10|[results](https://drive.google.com/file/d/1Pr7iAxPsjdZMwwEv7aZ54eiw7GMVoEsS/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1pLZr8uRfnPVkUzI3f7Yx0smcNN2W0l0N/view?usp=drive_link)|\n\n### Shadow Removal\nALl the results are in 480x640. For the pre-trained weights, we only provide SP+M weights, since this repository has full implementation of it. RMSE is reported for 480x640 images.\n\nModel: [SP+M](https://arxiv.org/abs/1908.08628)\n| Train | Test | RMSE |      |\n| :-: | :-: | :-: | :-: |\n|SynShadow|ISTD+|4.9|[results](https://drive.google.com/file/d/1r4uM280UIh1XLilHJoPWWiRQ6vBy0fmZ/view?usp=drive_link) / [weights](https://drive.google.com/file/d/10wLm8nVXdmXeBiSO5HfnGT3WNY6AvO8m/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/173WYHwKCV-Rs-sECEQ2Ia_xffY8d_8lS/view?usp=drive_link)|\n|SynShadow-\u003eISTD+|ISTD+|4.0|[results](https://drive.google.com/file/d/17R1byVIJ10dRduy_ZforioH2GQTQCenZ/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1dp28xws9TJ7gEyBeiSc0zLCQ09ALptTz/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/1qPi4dA-0TJvYReE2oCoC9zDyGMyxz8kN/view?usp=drive_link)|\n|SynShadow|SRD+|5.7|[results](https://drive.google.com/file/d/1-hpQzJ7gGiwNcllVgFqBs9O-yFR1S1YN/view?usp=drive_link) / [weights](https://drive.google.com/file/d/10wLm8nVXdmXeBiSO5HfnGT3WNY6AvO8m/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/1zWWIqZV6P09TTuTIfdkz2NlJZ-5z9Bc-/view?usp=drive_link)|\n|SynShadow-\u003eSRD+|SRD+|5.2|[results](https://drive.google.com/file/d/13e0AESmIbQxjjvZP9mXjdRYZz6ZQO0s3/view?usp=drive_link) / [weights](https://drive.google.com/file/d/1h_f7O8D8eqNClQ-uKD0MxyGiwhOhIJSi/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/1MoQPbudiFmHJ1hdcwYOjkdvH_ngRceGb/view?usp=drive_link)|\n|SynShadow|USR|-|[results](https://drive.google.com/file/d/1BM2m9HHXnFN6JLmcD85LLySe3j6E2cON/view?usp=drive_link) / [weights](https://drive.google.com/file/d/10wLm8nVXdmXeBiSO5HfnGT3WNY6AvO8m/view?usp=drive_link) / [precomputed_mask](https://drive.google.com/file/d/1MgH0WWHdzKCWnrMNAMcPNtI6tmBbYhMr/view?usp=drive_link)|\n\nModel: [DHAN](https://arxiv.org/abs/1911.08718)\n| Train | Test | RMSE |      |\n| :-: | :-: | :-: | :-: |\n|SynShadow-\u003eISTD+|ISTD+|4.6|[results](https://drive.google.com/file/d/1ci50ArxqLq5VygyVhvTJH1uUmhuhmYKp/view?usp=drive_link)|\n|SynShadow-\u003eSRD+|SRD+|6.6|[results](https://drive.google.com/file/d/1oiI_Vj5IFMrlT67XMsAF8cHmrxjpVtUY/view?usp=drive_link)|\n|SynShadow|USR|-|[results](https://drive.google.com/file/d/1M4CAUvfCdvFnQ_sXuH8ymjhRtaIOGq8e/view?usp=drive_link)|\n\nNote: we have accidentally removed some files and cannot provide some results.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnaoto0804%2Fsynshadow","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fnaoto0804%2Fsynshadow","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnaoto0804%2Fsynshadow/lists"}