Measuring Gait Variables Using Computer Vision to Assess Mobility and Fall Risk in Older Adults With Dementia

摘要: Fall risk is high for older adults with dementia. Gait impairment contributes to increased fall risk, and gait changes are common in people dementia, although the reliable assessment of challenging this population. This study aimed develop an automated approach performing assessments based on data that collected frequently unobtrusively, analysed using computer vision methods. Recent developments have led availability open source human pose estimation algorithms, which automatically estimate joint locations a person image. In study, pre-existing model was applied 1066 walking videos 31 dementia as they walked naturally corridor specialized unit over two week period. Using tracked information, features were extracted from video recordings bouts their association clinical mobility scores future falls examined. A significant found between number falls, providing concurrent predictive validation approach.