3DPW
收藏资源简介:
3DPW是野外3D姿势的缩写,是第一个具有精确3D姿势的户外数据集,可用于解决姿势估计问题。尽管存在其他户外数据集,但它们较小。3DPW是第一个包含由手机相机捕获的视频剪辑数据集的产品。 数据集包括: 60个视频序列。2D姿势注释。用本文提出的方法获得的3D姿势。序列中每一帧的摄像机位置。3D身体扫描和3D人体模型 (可重新存储和可重新计算)。每个序列都包含其对应的模型。18种不同服装的3D模型。
3DPW, short for 3D Pose in the Wild, is the first outdoor dataset with accurate 3D poses for solving pose estimation tasks. While other outdoor datasets exist, they are relatively small in scale. 3DPW is the first dataset that includes video clips captured by smartphone cameras. The dataset includes: - 60 video sequences - 2D pose annotations - 3D poses obtained via the method proposed in this paper - Camera positions for each frame in the sequences - 3D body scans and 3D human models (storable and re-computable) - A corresponding human model for each sequence - 3D models of 18 different clothing styles

- 3DPW数据集首次发表,由德国图宾根大学和马克斯·普朗克智能系统研究所联合发布,旨在提供高质量的三维人体姿态估计数据。
- 3DPW数据集首次应用于国际计算机视觉与模式识别会议(CVPR),展示了其在三维人体姿态估计领域的应用潜力。
- 3DPW数据集被广泛应用于多个研究项目中,包括但不限于三维人体姿态估计、动作识别和人体运动分析,进一步验证了其数据质量和多样性。
- 3DPW数据集的扩展版本发布,增加了更多的场景和动作类型,提升了数据集的覆盖范围和应用广度。
- 13DPW: Recovering 3D Human Pose from Monocular ImagesMax Planck Institute for Informatics · 2018年
- 2Monocular 3D Human Pose Estimation In The Wild Using Improved CNN SupervisionMax Planck Institute for Informatics · 2017年
- 3Monocular 3D Pose and Shape Estimation of Multiple People in Natural ScenesUniversity of California, Berkeley · 2020年
- 43D Human Pose Estimation in Video with Temporal Convolutions and Semi-Supervised TrainingUniversity of California, Berkeley · 2019年
- 5Learning to Reconstruct 3D Human Pose and Shape via Model-fitting in the LoopUniversity of Southern California · 2019年



