QMAR (Quality of Movement Assessment for Rehabilitation)

使用 6 个 Primesense 相机（3 个不同的正面视图和 3 个不同的侧面视图）记录 QMAR，其中有 38 名健康受试者，8 名女性和 30 名男性。受试者由物理治疗师训练，在模拟三种疾病的同时进行两种不同类型的运动，从而产生五种总体可能性：在模拟帕金森 (W-P)、中风 (W-S) 和跛行 (W-L) 时返回到大致原始位置，与帕金森 (SS-P) 和中风 (SS-S) 一起站立和坐下。 数据集包括 RGB（以及深度和骨架）数据，尽管在此版本中，仅发布了 RGB 数据。我们可以根据要求提供深度/骨架数据，但请注意它们仅适用于两个视图。QMAR was recorded using 6 Primesense cameras, with 3 distinct frontal views and 3 distinct lateral views. A total of 38 healthy subjects participated in the data acquisition, including 8 females and 30 males. Subjects were trained by physical therapists to perform two distinct types of movements while simulating three medical conditions, resulting in five overall scenarios: returning to a roughly original position while simulating Parkinson’s disease (W-P), stroke (W-S), and claudication (W-L), as well as standing up and sitting down while simulating Parkinson’s disease (SS-P) and stroke (SS-S). The dataset encompasses RGB (along with depth and skeleton) data; however, only RGB data is released in this version. Depth and skeleton data can be provided upon request, but please note that they are only available for two camera views.

搜集汇总

背景与挑战

背景概述

QMAR数据集是一个用于康复研究的运动质量评估数据集，包含38名健康受试者在模拟帕金森、中风和跛行等疾病状态下的RGB运动数据，由布里斯托尔大学于2020年发布。