{
  "paper_id": "1060b0dc2ab5e7d316c99ce78010c538",
  "paper_title": "SLOPER4D: A Scene-Aware Dataset for Global 4D Human Pose Estimation in Urban Environments",
  "paper_authors": [
    "Yudi Dai",
    "Yitai Lin",
    "Xiping Lin",
    "Chenglu Wen",
    "Lan Xu",
    "Hongwei Yi",
    "Siqi Shen",
    "Yuexin Ma",
    "Cheng Wang"
  ],
  "publication_date": "",
  "publication_venue": "2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)",
  "abstract": "We present SLOPER4D, a novel scene-aware dataset collected in large urban environments to facilitate the research of global human pose estimation (GHPE) with human-scene interaction in the wild. Employing a head-mounted device integrated with a LiDAR and camera, we record 12 human subjects' activities over 10 diverse urban scenes from an egocentric view. Frame-wise annotations for 2D key points, 3D pose parameters, and global translations are provided, together with reconstructed scene point clouds. To obtain accurate 3D ground truth in such large dynamic scenes, we propose a joint optimization method to fit local SMPL meshes to the scene and fine-tune the camera calibration during dynamic motions frame by frame, resulting in plausible and scene-natural 3D human poses. Eventually, SLOPER4D consists of 15 sequences of human motions, each of which has a trajectory length of more than 200 meters (up to 1,300 meters) and covers an area of more than 200 m² (up to 30,000 m²), including more than 100k LiDAR frames, 300k video frames, and 500k IMU-based motion frames. With SLOPER4D, we provide a detailed and thorough analysis of two critical tasks, including camera-based 3D HPE and LiDAR-based 3D HPE in urban environments, and benchmark a new task, GHPE. The in-depth analysis demonstrates SLOPER4D poses significant challenges to existing methods and produces great research opportunities. The dataset and code are released at <http://www.lidarhumanmotion.net/sloper4d/>.",
  "keywords": [
    "global human pose estimation",
    "scene-aware dataset",
    "urban environments",
    "LiDAR",
    "IMU",
    "multi-modal",
    "SMPL"
  ],
  "paper_institutions": [
    "Fujian Key Laboratory of Sensing and Computing for Smart Cities, Xiamen University",
    "Key Laboratory of Multimedia Trusted Perception and Efficient Computing, Ministry of Education of China, School of Informatics, Xiamen University"
  ],
  "output_language": "zh",
  "provider": "doubao",
  "model": "ark-code-latest",
  "prompt_template": "simple",
  "extracted_at": "2026-08-16T19:46:24.903024Z",
  "template_tag": "simple",
  "summary": "### 摘要\n本文提出SLOPER4D，一个大规模城市环境下用于全局4D人体姿态估计的场景感知数据集。针对现有数据集在室外大场景中缺乏全局空间信息和场景约束的问题，SLOPER4D采用头戴式LiDAR和相机，同步采集12名受试者在10个不同城市场景中的活动，提供逐帧的2D关键点、3D姿态参数、全局平移以及重建的场景点云。为获得高精度3D标注，提出联合优化方法，将SMPL网格拟合到场景并逐帧优化相机外参，兼顾局部和全局动作的平滑性、场景接触和点云一致性。数据集包含15条序列，轨迹长度200至1300米，覆盖面积达30000平方米，总计超10万帧LiDAR、30万帧视频和50万帧IMU数据。基于此，论文评估了基于相机和LiDAR的3D人体姿态估计任务，并首次建立全局人体姿态估计（GHPE）基准。实验结果表明，现有方法在SLOPER4D上面临较大挑战，而联合优化方法显著提高了标注质量。该数据集为城市级人体动捕和场景理解提供了新契机，可应用于AR/VR、自动驾驶、智慧城市等场景。核心机制是通过最小化包含平滑项、场景接触项、姿态先验项和网格到点云项的总损失来优化全局人体运动。\n\n### 关键要点\n- SLOPER4D是首个城市级、多模态（LiDAR、相机、IMU）且包含丰富人-场景标注的3D人体姿态数据集，支持全局4D人体姿态估计。\n- 数据采集使用头戴式设备，集成128线LiDAR和DJI-Action2相机，并配合17个无线IMU，LiDAR向下倾斜约45°以保证人体在视场内。\n- 数据集包含15个序列，覆盖10个场景（如图书馆、商业街、海滨跑道、足球场等），单序列轨迹长度超过200米，总面积最大达30000平方米。\n- 标注包括SMPL模型参数、全局平移、2D关键点、2D边界框、3D点云分割以及重建的场景网格，并提供TLS高精度彩色点云。\n- 跨数据集实验表明，在LiDAR模态上，使用混合数据集训练可将MPJPE降低约60%，说明不同传感器数据互补能有效减小域差距。\n- 在相机模态上，用3DPW预训练的VIBE模型在SLOPER4D上的误差（MPJPE 102.5mm）高于在3DPW上（93.5mm），但联合微调可提高在SLOPER4D上的性能。\n- 全局姿态估计基准中，GLAMR在足球和花园场景的RPE较低（RMSE分别为0.08和0.06），但在跑步场景中表现最差（ATE RMSE 29.48m），表明高动态场景更具挑战。\n- 联合优化方法结合平滑损失、场景接触损失、姿态先验和视线感知的网格到点云损失，能有效矫正IMU漂移和摄像机抖动，提高标注精度。\n- 相机外参优化通过最小化2D关键点和边界框投影误差，显著降低了投影偏差，提升了数据质量。\n- 数据集划分：训练集含11个序列约80k LiDAR帧，测试集4个序列约20k帧，并设置三个GHPE评估场景。\n- 现有方法在SLOPER4D上性能下降明显，暗示了大规模室外场景中全局姿态估计的难度，为后续研究提供了新基准和挑战。",
  "available_templates": [
    "deep_read",
    "simple"
  ]
}