Differentiating post-stroke patients from healthy individuals via vision-based skeleton-optical fusion.
Xiao Han, Ziyan Wang, Liping Li and 1 others
PMID 41068869WHAT IT FOUND
A vision-based algorithm correctly classified 87.8% of post-stroke and healthy gait videos using a single camera.
It relies on sagittal-plane walking in a controlled clinical setting, so it is not yet a tool for home monitoring or diverse patient populations.
Key findings
01The best-performing model configuration achieved an accuracy of 0.8778 on the clinical dataset.
02Concatenation fusion significantly outperformed multi-head attention, achieving 87.8% accuracy versus 46.1% on the clinical dataset.
03The study was limited to participants who could walk independently, excluding those with severe mobility impairments.
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 only included patients who could walk independently, so the algorithm's utility for more severely impaired patients is unknown. Data collection was restricted to the sagittal plane (side view), missing frontal plane movements that are often critical in stroke gait analysis. The dataset was small (140 participants) and collected from a single center, limiting generalizability. The model did not incorporate gait event detection (like heel-strike), which is necessary for detailed phase analysis. Demographic factors such as age and height were not fully accounted for in the analysis.
Declared interests
The study was funded by the National Key Research and Development Program of China. No other conflicts of interest were declared.
The easy way to misread this
Do not assume this algorithm works for all stroke survivors or in uncontrolled environments. The model was trained only on patients who could walk independently along a straight 4-meter path filmed from the side, so it has not been validated for those with severe gait deviations, assistive device users, or multi-directional movement.
Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →