HSC4D

该数据集名为HSC4D，包含了三个大型场景（分别是攀岩馆、多层建筑和户外闭环道路），这些场景中准确记录了动态人体动作和位置信息，专为人体-场景建模而设计。该数据集包含了25万帧IMU数据（100Hz），5万帧时间同步的激光雷达数据（20Hz），SLAM结果以及彩色的地面真实点云。规模方面，三个大型场景的面积从1千到5千平方米不等。该数据集的任务是进行4D场景中动态人体动作捕捉与定位。

This dataset is named HSC4D, which includes three large-scale scenarios: a climbing gym, a multi-story building, and an outdoor closed-loop road. These scenarios accurately record dynamic human motions and positional information, and are specifically designed for human-scene modeling. The dataset contains 250,000 frames of IMU data (100Hz), 50,000 frames of time-synchronized LiDAR data (20Hz), SLAM results, and colored ground-truth point clouds. In terms of scale, the area of these three large scenarios ranges from 1,000 to 5,000 square meters. The task of this dataset is dynamic human motion capture and localization in 4D scenarios.