CrowdHuman
收藏资源简介:
CrowdHuman数据集由旷视科技创建,专注于解决人群中的人体检测问题。该数据集包含15000张训练图像、4370张验证图像和5000张测试图像,总计47万个人体实例,平均每张图像有22.6个人。数据集通过爬取Google图像搜索结果并进行严格筛选和标注,每个实例都标注了头部、可见区域和全身的边界框。CrowdHuman数据集的应用领域包括自动驾驶、智能监控和机器人技术等,旨在提高人群中人体检测的准确性和鲁棒性。
CrowdHuman dataset was created by Megvii Technology, focusing on solving the problem of human detection in crowded scenarios. This dataset includes 15,000 training images, 4,370 validation images and 5,000 test images, with a total of 470,000 human instances, averaging 22.6 individuals per image. It is constructed by crawling results from Google Image Search and conducting strict filtering and annotation. Each instance is annotated with bounding boxes for the head, visible area and full body. The application fields of CrowdHuman dataset include autonomous driving, intelligent surveillance, robotics and other areas, aiming to improve the accuracy and robustness of human detection in crowded environments.

- 1CrowdHuman: A Benchmark for Detecting Human in a Crowd旷视科技(Face++) · 2018年



