{"id":13471886,"url":"https://github.com/LiewFeng/RayDN","last_synced_at":"2025-03-26T15:30:42.885Z","repository":{"id":227822690,"uuid":"752177999","full_name":"LiewFeng/RayDN","owner":"LiewFeng","description":"[ECCV 2024] Ray Denoising (RayDN): Depth-aware Hard Negative Sampling for Multi-view 3D Object Detection","archived":false,"fork":false,"pushed_at":"2024-09-30T03:33:23.000Z","size":22005,"stargazers_count":72,"open_issues_count":1,"forks_count":1,"subscribers_count":2,"default_branch":"main","last_synced_at":"2024-10-30T04:11:31.822Z","etag":null,"topics":["3d-object-detection","autonomous-driving","denoising","eccv","eccv2024","nuscenes"],"latest_commit_sha":null,"homepage":"https://arxiv.org/abs/2402.03634","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/LiewFeng.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":"2024-02-03T08:59:17.000Z","updated_at":"2024-10-29T13:12:30.000Z","dependencies_parsed_at":"2024-03-25T16:17:27.825Z","dependency_job_id":"a706494c-d9bc-4ce7-8636-abf5a8fdeb6b","html_url":"https://github.com/LiewFeng/RayDN","commit_stats":null,"previous_names":["liewfeng/raydn"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LiewFeng%2FRayDN","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LiewFeng%2FRayDN/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LiewFeng%2FRayDN/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/LiewFeng%2FRayDN/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/LiewFeng","download_url":"https://codeload.github.com/LiewFeng/RayDN/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245681183,"owners_count":20655148,"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":["3d-object-detection","autonomous-driving","denoising","eccv","eccv2024","nuscenes"],"created_at":"2024-07-31T16:00:50.053Z","updated_at":"2025-03-26T15:30:42.474Z","avatar_url":"https://github.com/LiewFeng.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n\u003ch1\u003eRayDN\u003c/h1\u003e\n\u003ch1\u003eRay Denoising: Depth-aware Hard Negative Sampling  for Multi-view 3D Object Detection\u003c/h1\u003e\n\u003c/div\u003e\n\n[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/beam-beta-distribution-ray-denoising-for/3d-object-detection-on-nuscenes-camera-only)](https://paperswithcode.com/sota/3d-object-detection-on-nuscenes-camera-only?p=beam-beta-distribution-ray-denoising-for)\n\u003ca href=\"https://www.ecva.net/papers/eccv_2024/papers_ECCV/html/6549_ECCV_2024_paper.php\"\u003e\u003cimg src=\"https://img.shields.io/badge/ECCV2024-Paper-\u003ccolor\u003e\"\u003e\u003c/a\u003e\n[![arXiv](https://img.shields.io/badge/arXiv-Paper-\u003cCOLOR\u003e.svg)](https://arxiv.org/abs/2402.03634)\n\n[\u003cvideo src=\"figs/RayDN.mp4\"\u003e\u003c/video\u003e](https://github.com/LiewFeng/RayDN/assets/42316773/de2c229b-0f6a-4456-b72c-508ea6161198)\n\n\n\n## Introduction\n\nThis repository is an official implementation of our ***ECCV 2024*** paper [Ray Denoising: Depth-aware Hard Negative Sampling for Multi-view 3D Object Detection](https://arxiv.org/abs/2402.03634). This repository contains Pytorch training code, evaluation code and pre-trained models.\n\n## Framework\n\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"figs/framework.png\" width=\"800\"/\u003e\n\u003c/div\u003e\u003cbr/\u003e\n\n## Getting Started\n\nOur code is built based on [StreamPETR](https://github.com/exiawsh/StreamPETR). Please follow [StreamPETR](https://github.com/exiawsh/StreamPETR) to [setup enviroment](https://github.com/exiawsh/StreamPETR/blob/main/docs/setup.md) and [prepare data](https://github.com/exiawsh/StreamPETR/blob/main/docs/data_preparation.md) step by step.\n\n## Training and Inference\nYou can train the model following:\n\n```angular2html\ntools/dist_train.sh projects/configs/RayDN/raydn_eva02_800_bs2_seq_24e.py 8 \n```\n\nYou can evaluate the detection model following:\n```angular2html\ntools/dist_test.sh projects/configs/RayDN/raydn_eva02_800_bs2_seq_24e.py work_dirs/raydn_eva02_800_bs2_seq_24e/latest.pth 8 --eval bbox\n```\n\n\n## Results on NuScenes Val Set.\n| Model | Setting |Pretrain| Lr Schd | NDS| mAP| Config | Download |\n| :---: | :---: | :---: | :---: | :---:|:---:| :---: | :---: |\n| RayDN | R50 - 428q | NuImg | 60ep | 56.1 | 47.1 | [config](projects/configs/RayDN/raydn_r50_flash_704_bs2_seq_428q_nui_60e.py) | [ckpt](https://mailsucasaccn-my.sharepoint.com/:u:/g/personal/liufeng20_mails_ucas_ac_cn/EYtElqwLxxRMqewZ0qZIz2wBmfLoPrOe3YIVdlLVZSKGcQ?e=wdbkHi) |\n| RayDN | EVA02-L - 900q | EVA02 | 24ep | 62.4 | 54.1 | [config](projects/configs/RayDN/raydn_eva02_800_bs2_seq_24e.py) |[ckpt](https://mailsucasaccn-my.sharepoint.com/:u:/g/personal/liufeng20_mails_ucas_ac_cn/ERYKTAGGSKRFmrDoF6VnUf8BKw96Cw-rNyvbFFrouQWJBw?e=Dkcil3) |\n\n\n\n\n\n\n## Acknowledgements\n\nWe thank these great works and open-source codebases:\n[MMDetection3d](https://github.com/open-mmlab/mmdetection3d), [StreamPETR](https://github.com/exiawsh/StreamPETR), [DETR3D](https://github.com/WangYueFt/detr3d), [PETR](https://github.com/megvii-research/PETR).\n\n\n## Citation\n\nIf you find RayDN is useful in your research or applications, please consider giving us a star 🌟 and citing it by the following BibTeX entry.\n```bibtex\n@article{liu2024ray,\n  title={Ray Denoising: Depth-aware Hard Negative Sampling for Multi-view 3D Object Detection},\n  author={Liu, Feng and Huang, Tengteng and Zhang, Qianjing and Yao, Haotian and Zhang, Chi and Wan, Fang and Ye, Qixiang and Zhou, Yanzhao},\n  journal={arXiv preprint arXiv:2402.03634},\n  year={2024}\n}\n```\n\n\n\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FLiewFeng%2FRayDN","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FLiewFeng%2FRayDN","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FLiewFeng%2FRayDN/lists"}