{"id":13704764,"url":"https://github.com/HYUNJS/SGT","last_synced_at":"2025-05-05T12:32:13.604Z","repository":{"id":41156062,"uuid":"418201451","full_name":"HYUNJS/SGT","owner":"HYUNJS","description":"[WACV-2023] Detection Recovery in Online Multi-Object Tracking with Sparse Graph Tracker","archived":false,"fork":false,"pushed_at":"2023-02-14T13:12:55.000Z","size":225,"stargazers_count":56,"open_issues_count":5,"forks_count":4,"subscribers_count":6,"default_branch":"main","last_synced_at":"2024-08-03T22:14:00.134Z","etag":null,"topics":["graph-neural-network","joint-detection-and-tracking","multi-object-tracking","real-time"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/HYUNJS.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null}},"created_at":"2021-10-17T17:12:19.000Z","updated_at":"2024-07-11T03:09:36.000Z","dependencies_parsed_at":"2024-01-12T20:17:14.599Z","dependency_job_id":"2644737a-4202-4066-b1b3-6f9810941e7a","html_url":"https://github.com/HYUNJS/SGT","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/HYUNJS%2FSGT","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HYUNJS%2FSGT/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HYUNJS%2FSGT/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HYUNJS%2FSGT/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/HYUNJS","download_url":"https://codeload.github.com/HYUNJS/SGT/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":224448655,"owners_count":17313090,"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":["graph-neural-network","joint-detection-and-tracking","multi-object-tracking","real-time"],"created_at":"2024-08-02T22:00:16.133Z","updated_at":"2024-11-13T12:30:53.561Z","avatar_url":"https://github.com/HYUNJS.png","language":"Python","funding_links":[],"categories":["算法论文"],"sub_categories":["**2022**"],"readme":"# Sparse Graph Tracker (SGT)\n[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/detection-recovery-in-online-multi-object/multi-object-tracking-on-hieve)](https://paperswithcode.com/sota/multi-object-tracking-on-hieve?p=detection-recovery-in-online-multi-object)\n[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/detection-recovery-in-online-multi-object/multi-object-tracking-on-mot16)](https://paperswithcode.com/sota/multi-object-tracking-on-mot16?p=detection-recovery-in-online-multi-object)\n[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/detection-recovery-in-online-multi-object/multi-object-tracking-on-mot17)](https://paperswithcode.com/sota/multi-object-tracking-on-mot17?p=detection-recovery-in-online-multi-object)\n[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/detection-recovery-in-online-multi-object/multi-object-tracking-on-mot20-1)](https://paperswithcode.com/sota/multi-object-tracking-on-mot20-1?p=detection-recovery-in-online-multi-object)\n\nOfficial code for Sparse Graph Tracker (SGT) based on the Detectron2 framework. Please feel free to leave an ISSUE or send me an email (jhyunaa@ust.hk).\n\n## News\n* (2022.10.11) Our paper is accepted WACV 2023! (arxiv paper will be updated soon)\n* (2022.10.06) Code and pretrained weights are released!\n\n## Installation\n* Please refer [INSTALL.md](INSTALL.md) for the details \n## Dataset Setup\n* Please refer [DATASET.md](DATASET.md) for the details\n\n## Model Zoo\n- [download model weights](https://hkustconnect-my.sharepoint.com/:f:/g/personal/jhyunaa_connect_ust_hk/ErLZ6DG6CndHs-Lo12AxZKAB1kb4AJCh8adtnRAlXTNuzA?e=rKeboA)\n* Please modify the path of checkpoints in the config file based on your checkpoint directory\n\n### MOT17\n| Name | Dataset | HOTA | MOTA | IDF1| Download |\n| :---: | :---: | :---: | :---: | :---: | :---: |\n| SGT | MOT17 | 58.2 | 73.2 | 70.2 | [model](https://hkustconnect-my.sharepoint.com/:u:/g/personal/jhyunaa_connect_ust_hk/EQW4mXblacdDtVc3uxaTsXYB2yqqUTQv9cwnBipAnpKblA?e=9pT2RO) |\n| SGT | MOT17 + CrowdHuman | 60.8 | 76.4 | 72.8 | [model](https://hkustconnect-my.sharepoint.com/:u:/g/personal/jhyunaa_connect_ust_hk/EST0ZaRgqvlJoW2TGFpCUToBRUzGRkgXZQva32rypzWdZQ?e=OzCVgI) |\n\n### MOT20\n| Name | Dataset | HOTA | MOTA | IDF1| Download |\n| :---: | :---: | :---: | :---: | :---: | :---: |\n| SGT | MOT20 | 51.6 | 64.5 | 62.7 | [model](https://hkustconnect-my.sharepoint.com/:u:/g/personal/jhyunaa_connect_ust_hk/EVQks100QaRNp81QlSoQbwMBgncyxw-4cmE_eIrR3JPJoA?e=hQIHxF) |\n| SGT | MOT20 + CrowdHuman | 57.0 | 72.8 | 70.6 | [model](https://hkustconnect-my.sharepoint.com/:u:/g/personal/jhyunaa_connect_ust_hk/EZWXTwJFbPNBs2d32RVAa84BYTjImlYMSsz-Fp4lt8aE6A?e=qiOoHw) |\n\n### HiEve\n| Name | Dataset | MOTA | IDF1 | Download |\n| :---: | :---: | :---: | :---: | :---: | \n| SGT | HiEve | 47.2 | 53.7 | [model](https://hkustconnect-my.sharepoint.com/:u:/g/personal/jhyunaa_connect_ust_hk/EVQks100QaRNp81QlSoQbwMBgncyxw-4cmE_eIrR3JPJoA?e=hQIHxF) |\n\n## How to run?\n\n### Train\n```\npython projects/SGT/train_net.py --config-file projects/SGT/configs/MOT17/sgt_dla34.yaml --data-dir /root/datasets --num-gpus 2 OUTPUT_DIR /root/sgt_output/mot17_val/dla34_mot17-CH\n```\n\n### Inference\n```\npython projects/SGT/train_net.py --config-file projects/SGT/configs/MOT17/sgt_dla34.yaml --data-dir /root/datasets --num-gpus 1 --eval-only OUTPUT_DIR /root/sgt_output/mot17_test/dla34_mot17-CH\n```\n\n### Visualization\n```\n## GT\npython projects/Datasets/MOT/vis/vis_gt.py --data-root \u003c$DATA_ROOT\u003e --register-data-name \u003ce.g., mot17_train\u003e \npython projects/Datasets/MOT/vis/vis_gt.py --data-root \u003c$DATA_ROOT\u003e --register-data-name \u003ce.g., mix_crowdhuman_train\u003e --no-video-flag \n\n\n## model output\npython projects/Datasets/MOT/vis/vis_seq_from_txt_result.py --data-root \u003c$DATA_ROOT\u003e --result-dir \u003c$OUTPUT_DIR\u003e --data-name {mot17, mot20, hieve, mot17_sub, mot20_sub} --tgt-split {val,test}\n```\n\n## Motivation\n![image](https://user-images.githubusercontent.com/29353227/194476858-69c24328-f461-48b9-9262-17f90f38e652.png)\n\n## Pipeline\n![image](https://user-images.githubusercontent.com/29353227/194477178-d31da80b-c215-4acf-ab9d-8519b9f54f9f.png)\n\n## MOT Benchmark Results\n![image](https://user-images.githubusercontent.com/29353227/194478496-39309fea-ced0-4d3f-8be0-cce87f4c9c57.png)\n\n## Ablation Experiment Results\n![image](https://user-images.githubusercontent.com/29353227/194478002-ba6bff6d-6665-45de-80ed-51f384b10094.png)\n\n![image](https://user-images.githubusercontent.com/29353227/194478011-66d3c56d-89bc-40e5-a4d0-22558a9d9159.png)\n\n## Visualization\n![image](https://user-images.githubusercontent.com/29353227/194478129-4a1684ee-7326-4ad1-b3d5-989b13e2b7c5.png)\n\n\n\n## License\nCode of SGT is licensed under the CC-BY-NC 4.0 license and free for research and academic purpose.\nSGT is based on the framework [Detectron2](https://github.com/facebookresearch/detectron2) which is released under the Apache 2.0 license and the detector [CenterNet](https://github.com/xingyizhou/CenterNet) which is released under the MIT license.\nThis codebase also provides Detectron2 version of [FairMOT](https://github.com/ifzhang/FairMOT) which is released under the MIT license.\n\n## Citation\n```BibTeX\n@inproceedings{hyun2023detection,\n  title={Detection recovery in online multi-object tracking with sparse graph tracker},\n  author={Hyun, Jeongseok and Kang, Myunggu and Wee, Dongyoon and Yeung, Dit-Yan},\n  booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},\n  pages={4850--4859},\n  year={2023}\n}\n```\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FHYUNJS%2FSGT","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FHYUNJS%2FSGT","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FHYUNJS%2FSGT/lists"}