FreeMan

FreeMan是由香港中文大学（深圳）创建的大规模多视角3D人体姿态估计数据集，包含8000个序列共1100万帧，涵盖40个不同场景下的40名参与者。数据集通过8台智能手机同步采集，旨在解决现有数据集在真实世界条件下的不足，提供更丰富的场景多样性和人体动作变化。FreeMan的应用领域包括动画制作、虚拟现实、元宇宙和机器人交互等，旨在推动3D人体姿态估计技术在真实环境中的应用和发展。

FreeMan is a large-scale multi-view 3D human pose estimation dataset created by The Chinese University of Hong Kong, Shenzhen. It consists of 8000 sequences totaling 11 million frames, involving 40 participants across 40 distinct scenarios. The dataset is synchronously captured using 8 smartphones, aiming to address the limitations of existing datasets under real-world conditions, and provides richer scene diversity and human motion variations. Its application fields include animation production, virtual reality, the metaverse, robot interaction and other areas, with the goal of promoting the application and development of 3D human pose estimation technology in real-world environments.