Dataset Groups Activity Stream CrowdHuman dataset The CrowdHuman dataset is a benchmark dataset for human detection, consisting of 15,000 images, 4,370 images for validation, and 5,000 images for testing. BibTex: @dataset{Jiangmiao_Pang_and_Kai_Chen_and_Jianping_Shi_and_Huajun_Feng_and_Wanli_Ouyang_and_Dahua_Lin_2024, abstract = {The CrowdHuman dataset is a benchmark dataset for human detection, consisting of 15,000 images, 4,370 images for validation, and 5,000 images for testing.}, author = {Jiangmiao Pang and Kai Chen and Jianping Shi and Huajun Feng and Wanli Ouyang and Dahua Lin}, doi = {10.57702/wimdntj5}, institution = {No Organization}, keyword = {'crowd detection', 'human detection', 'image classification', 'object detection'}, month = {dec}, publisher = {TIB}, title = {CrowdHuman dataset}, url = {https://service.tib.eu/ldmservice/dataset/crowdhuman-dataset}, year = {2024} }