{"id":13644633,"url":"https://github.com/ZijianWang-ZW/PPE_detection","last_synced_at":"2025-04-21T10:33:56.348Z","repository":{"id":41253206,"uuid":"286366731","full_name":"ZijianWang-ZW/PPE_detection","owner":"ZijianWang-ZW","description":"Real-time PPE detection based on YOLO. Open high-quality dataset.","archived":false,"fork":false,"pushed_at":"2022-11-03T08:35:47.000Z","size":7829,"stargazers_count":54,"open_issues_count":2,"forks_count":9,"subscribers_count":2,"default_branch":"master","last_synced_at":"2025-04-12T10:05:41.025Z","etag":null,"topics":["construction-safety","construction-site","deep-learning","ppe-detection","yolo"],"latest_commit_sha":null,"homepage":"https://github.com/ZijianWang1995/PPE_detection","language":null,"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/ZijianWang-ZW.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}},"created_at":"2020-08-10T03:27:59.000Z","updated_at":"2025-04-10T14:21:04.000Z","dependencies_parsed_at":"2023-01-20T22:17:47.592Z","dependency_job_id":null,"html_url":"https://github.com/ZijianWang-ZW/PPE_detection","commit_stats":null,"previous_names":["zijianwang1995/ppe_detection"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZijianWang-ZW%2FPPE_detection","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZijianWang-ZW%2FPPE_detection/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZijianWang-ZW%2FPPE_detection/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ZijianWang-ZW%2FPPE_detection/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ZijianWang-ZW","download_url":"https://codeload.github.com/ZijianWang-ZW/PPE_detection/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":249528384,"owners_count":21286389,"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":["construction-safety","construction-site","deep-learning","ppe-detection","yolo"],"created_at":"2024-08-02T01:02:09.827Z","updated_at":"2025-04-21T10:33:51.336Z","avatar_url":"https://github.com/ZijianWang-ZW.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"# Real-time PPE Detection \u0026 Open Dataset ([Open Access paper link](https://www.mdpi.com/1424-8220/21/10/3478))\n\n![avatar](figures/ppe_sample.gif)\n\n## Introduction\n\nThe repository introduces eight DL models built on You Look Only Once (YOLO) architecture for PPE detection. Meanwhile, a novel high-quality dataset is constructed for detecting the person, the vest, and four helmet colors.\n\n## Framework\n\n![avatar](figures/methodology.PNG)\n\n## CHV Dataset\n\nA novel dataset  is constructed for detecting the helmet, the helmet colors and the person for this project, named **Color Helmet and Vest (CHV)** dataset. \n\nInstead of just accepting exiting images, strict criteria are designed at the beginning, and only  1,330 high-quality images among 10,000 ones from the Internet and open datasets are selected. \n\n\u003cdiv align=center\u003e\u003cimg src=\"figures/size_class.png\" width=\"60%\" /\u003e\u003c/div\u003e\n\n\n\nThe dataset is open for free use, please download at [Google Drive](https://drive.google.com/file/d/1fdGn67W0B7ShpBDbbQpUF0ScPQa4DR0a/view?usp=sharing) or [Baidu Yunpan (password: f003)](https://pan.baidu.com/s/1G9EbLKUgF1tcOPCeWSEeMw ). \n\nIf the dataset helpes you, please cite the repository in your article:\n\n`\n@Article{wang2021ppe,\nAUTHOR = {Wang, Zijian and Wu, Yimin and Yang, Lichao and Thirunavukarasu, Arjun and Evison, Colin and Zhao, Yifan},\nTITLE = {Fast Personal Protective Equipment Detection for Real Construction Sites Using Deep Learning Approaches},\nJOURNAL = {Sensors},\nVOLUME = {21},\nYEAR = {2021},\nNUMBER = {10},\nARTICLE-NUMBER = {3478},\nURL = {https://www.mdpi.com/1424-8220/21/10/3478},\nISSN = {1424-8220},\nDOI = {10.3390/s21103478}\n}`\n\nOr\n\t\n`Wang, Z.; Wu, Y.; Yang, L.; Thirunavukarasu, A.; Evison, C.; Zhao, Y. Fast Personal Protective Equipment Detection for Real Construction Sites Using Deep Learning Approaches. Sensors 2021, 21, 3478. https://doi.org/10.3390/s21103478`\n\n## Results\n\n- **YOLO v5x owns the best mAP, 86.55%.**\n- **YOLO v5s has the faster processing speed, 52 FPS.**\n- For YOLO v3 models, different detection layers are tested, while the more layers cannot improve the performance. \n- For YOLO v4 models,  the increase of training image size cannot contribute to better performance. \n\n\u003ccenter\u003e\u003cimg src=\"figures/map.png\" style=\"zoom:100%;\"/\u003e \n\n\u003cdiv align=center\u003eFigure: Mean average precision in each model.\u003c/div\u003e\n\n\u003cdiv align=center\u003e\u003cimg src=\"figures/time.png\" width=\"60%\"/\u003e\u003c/div\u003e\n\n\u003cdiv align=center\u003eFigure: Average time for processing one image in each model (GPU: Tesla P40 with 24 GB; CPU: 4 cores with 8 GB). \u003c/div\u003e\n\n\n\n## \nIf you are interested in my work, please visit \u003chttps://zijianwang1995.github.io/\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FZijianWang-ZW%2FPPE_detection","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FZijianWang-ZW%2FPPE_detection","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FZijianWang-ZW%2FPPE_detection/lists"}