NWPU-Crowd
收藏资源简介:
NWPU-Crowd是一个大规模的人群计数和定位数据集,由西北工业大学计算机科学与技术学院与光学影像分析与学习中心创建。该数据集包含5109张图像,总计2,133,375个标注头部,适用于监督学习算法。数据集的特点包括高分辨率、包含负样本以及大的外观变化,旨在解决极端拥挤场景下的人群计数和定位问题。数据集的创建过程涉及从互联网和自拍摄影中收集数据,并通过高效的标注工具进行标注。NWPU-Crowd的应用领域包括人群监控、公共安全和空间设计等。
NWPU-Crowd is a large-scale crowd counting and localization dataset developed by the School of Computer Science and Technology and the Center for Optical Image Analysis and Learning of Northwestern Polytechnical University. It contains 5109 images with a total of 2,133,375 annotated heads, and is applicable to supervised learning algorithms. The dataset features high resolution, inclusion of negative samples, and large appearance variations, aiming to address crowd counting and localization tasks in extremely crowded scenarios. The dataset was constructed by collecting data from the Internet and self-captured photography, followed by annotation via efficient annotation tools. Application fields of NWPU-Crowd include crowd surveillance, public safety, spatial design, and other related areas.

- 1NWPU-Crowd: A Large-Scale Benchmark for Crowd Counting and Localization西北工业大学计算机科学与技术学院与光学影像分析与学习中心 · 2020年



