ProGait
收藏资源简介:
ProGait数据集由匹兹堡大学的研究团队创建,旨在支持视频目标分割、2D人体姿态估计和步态分析等视觉任务。数据集包含412个视频片段,记录了4位截肢者在使用不同假肢进行行走试验时的运动情况。视频片段涵盖了不同的场景,包括在平行杆内独立行走和在有辅助的情况下在走廊行走。每个视频片段都有详细的标注,包括边界框、分割掩码和姿态关键点,以及研究人员提供的步态分析文本描述。该数据集旨在提高视觉模型对截肢者步态的检测和分析能力,为假肢设计和优化提供支持。
The ProGait dataset was created by a research team at the University of Pittsburgh, aiming to support visual tasks such as video object segmentation, 2D human pose estimation and gait analysis. It contains 412 video clips recording the movements of 4 amputees during walking trials using different prostheses. The video clips cover diverse scenarios including independent walking within parallel bars and corridor walking with assistance. Each video clip is accompanied by detailed annotations, including bounding boxes, segmentation masks, pose keypoints, as well as gait analysis text descriptions provided by researchers. This dataset is intended to enhance the detection and analysis capabilities of visual models for amputee gaits, providing support for prosthetic design and optimization.




