Egocentric video: a new tool for capturing hand use of individuals with spinal cord injury at home.
Jirapat Likitlersuang, Elizabeth R Sumitro, Tianshi Cao and 3 others
PMID 31277682WHAT IT FOUND
A head-mounted video system detected hand-object interactions in nine people with cervical spinal cord injury.
Its interaction counts and durations tracked human observer labels moderately well, but it was tested in a simulated lab, not real homes.
Key findings
01The automated hand-use measures matched human observer labels moderately for proportion of interaction time, average interaction duration, and number of interactions per hour.
02The paper reports average accuracy values of 0.70 ± 0.16 for the left hand and 0.68 ± 0.18 for the right hand.
03The interaction detector was evaluated on nine participants whose data had not been used to train the hand-detection algorithms.
STILL TO COME
How it was doneWhat they foundWhat 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 tasks were pre-specified and performed in a home simulation laboratory, so the video may not represent the frequency or variety of hand use in a real home. The interaction detector was evaluated on nine participants, and performance varied between participants and sometimes between hands. Wooden objects and flooring caused false hand detections, and poor hand detection or segmentation can break the interaction output. The study did not compare automated hand-use metrics with existing clinical outcome measures such as GRASSP or SCIM. The algorithm was not real-time or mobile; it took 2.70 seconds per frame to process. Caregiver hands were not analysed because participants mostly completed tasks alone.
Declared interests
Funded by Natural Sciences and Engineering Research Council of Canada, Rick Hansen Institute, and Ontario Early Researcher Award. The supplied text lists no competing interests.
The easy way to misread this
Do not conclude this tool is ready to measure hand use in clinical practice or predict rehabilitation outcomes. It was a small feasibility evaluation in a simulated lab, and its hand-use measures only moderately matched human observer labels.