CIMI4D
收藏资源简介:
CIMI4D是由厦门大学福建智慧城市感知与计算重点实验室创建的大型多模态攀岩运动数据集,专注于复杂的人体-场景交互。数据集包含约180,000帧,涵盖RGB视频、激光雷达点云、IMU测量等多种模态数据,以及高精度重建的场景网格。通过精细的人体姿态和全局轨迹标注,CIMI4D旨在深入理解攀岩运动中的复杂动作和场景交互。此外,数据集还提供了岩石握点的详细标注,以支持对人体与攀岩环境交互的深入研究。CIMI4D的应用领域包括人体姿态估计、预测和生成,特别是在需要考虑场景约束的任务中,为计算机视觉算法提供了新的挑战和研究机会。
CIMI4D is a large-scale multimodal climbing motion dataset developed by the Fujian Key Laboratory of Smart City Sensing and Computing, Xiamen University, focusing on complex human-scene interactions. The dataset contains approximately 180,000 frames, covering multimodal data such as RGB videos, LiDAR point clouds, IMU measurements, as well as high-precision reconstructed scene meshes. With fine-grained human pose and global trajectory annotations, CIMI4D aims to facilitate in-depth understanding of complex movements and scene interactions in climbing motions. Additionally, the dataset provides detailed annotations of rock handholds to support in-depth research on human interactions with climbing environments. Application domains of CIMI4D include human pose estimation, prediction and generation, particularly in tasks that require consideration of scene constraints, offering new challenges and research opportunities for computer vision algorithms.

- 1CIMI4D: A Large Multimodal Climbing Motion Dataset under Human-scene Interactions福建智慧城市感知与计算重点实验室,厦门大学 · 2023年



