HUMAN4D

HUMAN4D数据集是由国家技术大学雅典分校电气与计算机工程学院等多个机构合作创建的大型多模态4D数据集。该数据集包含超过50,000个样本，涵盖日常、物理和社会活动，通过专业的标记基于多视角RGBD、体积捕捉和音频录制系统同步捕捉。HUMAN4D数据集首次提供使用消费者级深度传感设备进行硬件同步的多视角RGBD帧，以及标记的基于标记的运动捕捉和音频数据线索。数据集旨在支持广泛的计算机视觉和图形学研究，包括人体姿态估计、实时体积视频重建和压缩，以及4D表示的视觉评估。

The HUMAN4D dataset is a large-scale multimodal 4D dataset developed in collaboration with multiple institutions, among which the School of Electrical and Computer Engineering of the National Technical University of Athens is included. It contains over 50,000 samples covering daily, physical and social activities, which were synchronously captured using multi-view RGBD, volumetric capture and audio recording systems with professional annotations. Notably, the HUMAN4D dataset is the first to provide multi-view RGBD frames with hardware synchronization using consumer-grade depth sensing devices, alongside annotated marker-based motion capture and audio data cues. This dataset aims to support a broad spectrum of computer vision and graphics research topics, including human pose estimation, real-time volumetric video reconstruction and compression, as well as visual assessment of 4D representations.