Gait Dataset for Knee Osteoarthritis and Parkinson's Disease Analysis With Severity Levels

Computer-vision seems to be the most focused area in order to perform efficient diagnosis in clinical applications. The direct connection among human brain and the musculoskeletal system directed our effort towards the analysis of such diseases that have the significant impact on a person's gait such as knee osteoarthritis (KOA) and Parkinson's Disease (PD). Previous research on these diseases based on gait involves certain drawbacks including unavailability of a vision-based public and authenticated dataset. Thus, we created and presented a new dataset namely "KOA-PD-NM" while keeping various factors into consideration (e.g. age, gender, disease severity levels, etc.). This dataset involves both normal/healthy (NM) and abnormal (KOA, PD) subjects and don't allow the analysis of only lower body (limbs) but the upper body (arm, posture) movement too. The key aim of this dataset is to analyze the deviations among patient's and normal's gait. The construction of this dataset will surely help the researchers and the society towards this area for prediction of diseases with different stages and the development of better techniques and strategies.

在临床应用中实现高效诊断，计算机视觉（Computer Vision）无疑是受关注度最高的研究领域之一。人脑与肌肉骨骼系统的直接关联，促使我们将研究重心投向对步态影响显著的疾病分析，例如膝骨关节炎（Knee Osteoarthritis, KOA）与帕金森病（Parkinson's Disease, PD）。过往基于步态开展的此类疾病研究存在一定局限，其中包括缺乏可公开获取且经过权威认证的计算机视觉类数据集。为此，我们综合考量年龄、性别、疾病严重程度等多项因素，构建并发布了一款全新数据集，命名为“KOA-PD-NM”。该数据集涵盖正常/健康（NM）受试者与异常（KOA、PD）受试者，不仅可用于下肢（肢体）运动分析，还可同步开展上肢（手臂、姿态）运动研究。本数据集的核心目标是分析患者与健康人群的步态差异。该数据集的构建，将为该领域研究者与社会各界助力不同病程疾病的预测研究，并推动更优技术与策略的开发。

搜集汇总

背景与挑战

背景概述

该数据集是一个用于膝骨关节炎和帕金森病分析的步态数据集，包含健康人群和患者的数据，记录了不同严重程度的上下肢运动信息，旨在帮助研究人员分析疾病对步态的影响。