{"id":18573569,"url":"https://github.com/prs-eth/deflow","last_synced_at":"2025-04-10T07:32:12.060Z","repository":{"id":181978741,"uuid":"623962393","full_name":"prs-eth/DeFlow","owner":"prs-eth","description":"[CVPRW 2023, Best Paper Award] DeFlow: Self-supervised 3D Motion Estimation of Debris Flow","archived":false,"fork":false,"pushed_at":"2023-07-25T07:48:57.000Z","size":16587,"stargazers_count":31,"open_issues_count":0,"forks_count":2,"subscribers_count":3,"default_branch":"master","last_synced_at":"2025-03-24T18:11:25.261Z","etag":null,"topics":["opticalflow","sceneflow","selfsupervision","sensorfusion"],"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/prs-eth.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,"publiccode":null,"codemeta":null}},"created_at":"2023-04-05T13:05:36.000Z","updated_at":"2025-03-21T05:21:25.000Z","dependencies_parsed_at":"2024-11-06T23:11:02.471Z","dependency_job_id":"876b1209-3adc-4c12-841e-9d97e37a6f01","html_url":"https://github.com/prs-eth/DeFlow","commit_stats":null,"previous_names":["prs-eth/deflow"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FDeFlow","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FDeFlow/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FDeFlow/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FDeFlow/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/prs-eth","download_url":"https://codeload.github.com/prs-eth/DeFlow/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":248176448,"owners_count":21060066,"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":["opticalflow","sceneflow","selfsupervision","sensorfusion"],"created_at":"2024-11-06T23:10:56.926Z","updated_at":"2025-04-10T07:32:11.077Z","avatar_url":"https://github.com/prs-eth.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cp align=\"center\"\u003e\n\u003ch2 align=\"center\"\u003e  DeFlow: Self-supervised 3D Motion Estimation of Debris Flow :mountain:\u003c/h2\u003e\n\n\u003cp align=\"center\"\u003e\n    \u003ca href=\"http://zhuliyuan.net/\"\u003e\u003cstrong\u003eLiyuan Zhu\u003c/strong\u003e\u003c/a\u003e, \n    \u003ca href=\"https://github.com/yurujaja\"\u003e\u003cstrong\u003eYuru Jia\u003c/strong\u003e\u003c/a\u003e, \n    \u003ca href=\"https://shengyuh.github.io/\"\u003e\u003cstrong\u003eShengyu Huang\u003c/strong\u003e\u003c/a\u003e,\n    \u003ca href=\"https://gseg.igp.ethz.ch/people/scientific-assistance/nicholas-meyer.html\"\u003e\u003cstrong\u003eNicholas Meyer\u003c/strong\u003e\u003c/a\u003e,\n    \u003ca href=\"https://gseg.igp.ethz.ch/people/group-head/prof-dr--andreas-wieser.html\"\u003e\u003cstrong\u003eAndreas Wieser\u003c/strong\u003e\u003c/a\u003e,\n    \u003ca href=\"https://igp.ethz.ch/personen/person-detail.html?persid=143986\"\u003e\u003cstrong\u003eKonrad Schindler\u003c/strong\u003e\u003c/a\u003e,\n    \u003ca href=\"https://erdw.ethz.ch/en/people/profile.jordan-aaron.html\"\u003e\u003cstrong\u003eJordan Aaron\u003c/strong\u003e\u003c/a\u003e\n  \u003c/p\u003e\n\n\u003cp align=\"center\"\u003e\u003cstrong\u003eETH Zurich\u003c/strong\u003e\u003c/a\u003e\n  \u003ch3 align=\"center\"\u003e\u003ca href=\"https://openaccess.thecvf.com/content/CVPR2023W/PCV/papers/Zhu_DeFlow_Self-Supervised_3D_Motion_Estimation_of_Debris_Flow_CVPRW_2023_paper.pdf\"\u003ePaper\u003c/a\u003e \n  | \u003ca href=\"https://zhuliyuan.net/deflow\"\u003eWebsite\u003c/a\u003e | \u003ca href=\"https://www.research-collection.ethz.ch/handle/20.500.11850/599948\"\u003eDataset\u003c/a\u003e \u003c/h3\u003e \n  \u003cdiv align=\"center\"\u003e\u003c/div\u003e\n\n\n\n\u003cimage src=\"misc/overview.png\"/\u003e\n\u003c/p\u003e\n\nThis repository is the official implementation of paper:\n\u003cb\u003eDeFlow: Self-supervised 3D Motion Estimation of Debris Flow\u003c/b\u003e, CVPRW 2023.\n\nExisting work on scene flow estimation focuses on autonomous driving and mobile robotics, while automated solutions are lacking for motion in nature, such as that exhibited by debris flows. We propose \\deflow, a model for 3D motion estimation of debris flows, together with a newly captured dataset. We adopt a novel multi-level sensor fusion architecture and self-supervision to incorporate the inductive biases of the scene. We further adopt a multi-frame temporal processing module to enable flow speed estimation over time. Our model achieves state-of-the-art optical flow and depth estimation on our dataset, and fully automates the motion estimation for debris flows.\n\n## Installation :national_park:\nFirst clone our repository:\n```bash\ngit clone https://github.com/Zhu-Liyuan/DeFlow\ncd DeFlow\n```\n\nYou will need to install conda to build the environment.\n```bash\nconda create -n DeFlow python=3.9\nconda activate DeFlow\npip install -r requirements.txt\n```\n\n## Dataset and pretrained model\nWe provide preprocessed debris flow dataset. The preprocessed dataset and checkpoint can be downloaded by running:\n```shell\nwget --no-check-certificate --show-progress https://share.phys.ethz.ch/~gsg/DeFlow/DeFlow_Dataset.zip\nunzip DeFlow_Dataset.zip -d data\nwget --no-check-certificate --show-progress https://share.phys.ethz.ch/~gsg/DeFlow/checkpoint.zip\nunzip checkpoint.zip\n```\nYou can also build your own dataset following the structure below\n```Shell\n├── Data\n    ├── Cam1\n        ├── 000001.jpg\n        ├── 000002.jpg\n        .\n        .\n        ├── 00000X.jpg\n    ├── Cam2\n        ├── 000001.jpg\n        ├── 000002.jpg\n        .\n        .\n        ├── 00000X.jpg\n    ├── LiDAR\n        ├── 000001.ply\n        ├── 000002.ply\n        .\n        .\n        ├── 00000X.ply\n├── Transformations\n        ├── cam_intrinxics.txt\n        ├── LiCam_tranformations.txt\n```\nTo train a model, run:\n```bash\npython main.py --config_path configs/deflow_default.yaml\n```\nand you can change the mode to eval in the config file for evaluation.\n\n\n## Contact\nIf you have any questions, please let me know: \n- Liyuan Zhu {liyzhu@student.ethz.ch}\n\n## Citation\nIf you use DeFlow for any academic work, please cite our original paper.\n```bibtex\n@InProceedings{zhu2023DeFlow,\nauthor = {Liyuan Zhu and Yuru Jia and Shengyu Huang and Nicholas Meyer and Andreas Wieser and Konrad Schindler, Jordan Aaron},\ntitle = {DEFLOW: Self-supervised 3D Motion Estimation of Debris Flow},\nbooktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},\nmonth = {June},\nyear = {2023}\n}\n```\n\nAdditionally, we thank the respective developers of the following open-source projects:\n- [PCAccumulation](https://github.com/prs-eth/PCAccumulation) \n- [CamLiFlow](https://github.com/MCG-NJU/CamLiFlow) \n- [Self-mono-sf](https://github.com/visinf/self-mono-sf)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fprs-eth%2Fdeflow","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fprs-eth%2Fdeflow","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fprs-eth%2Fdeflow/lists"}