Enhancing fall risk assessment: instrumenting vision with deep learning during walks.
Jason Moore, Robert Catena, Lisa Fournier and 6 others
PMID 38909239WHAT IT FOUND
A new AI tool automatically tracks where pregnant women look while walking over obstacles in a lab.
It accurately identifies gaze on hazards, helping clinicians see if distraction contributes to fall risk without manual video review.
Key findings
01The AI model accurately detected objects and gaze location, with a mean average precision score of 0.93 for object detection.
02The system successfully segmented the walking track to determine if a participant was looking at their own path or a distraction, achieving a mean Intersection over Union score of 0.82.
03The tool was developed to automate the analysis of over 100 hours of video data, which would be impractical to review manually.
STILL TO COME
How it was doneWhat they foundWhat it means for PTsWhat it means for OTs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study was conducted in a highly controlled lab environment with consistent lighting and specific obstacles, so the model may not work in real-world settings with variable conditions. The sample size was small (20 healthy pregnant women), and the model was not tested on clinical populations with actual fall risk or visual impairments. The validation dataset was small, which may have inflated accuracy scores for some object classes.
Declared interests
Funded by the NIHR North East and North Cumbria Applied Research Collaboration, the American Society of Biomechanics, and Washington State University.
The easy way to misread this
Do not assume this tool is ready for clinical use. It was trained on a small, homogeneous dataset from healthy pregnant women in a controlled lab, and has not been validated for diagnosing fall risk in older adults or patients with neurological conditions.