Re-imagining sensory substitution through gestural control: Point-To-Tell 2.
Ligao Ruan, Giles Hamilton-Fletcher, Mahya Beheshti and 3 others
PMID 41263249WHAT IT FOUND
The prototype accurately identified pointed-at objects up to 5 metres but consistently overestimated distances by 1.21 metres.
It was tested on one sighted person, so it proves technical feasibility but not that blind users can navigate with it.
Key findings
01Pointing-based object selection accuracy remained robust up to 5 metres, achieving 96.90% accuracy across frames.
02The system demonstrated a consistent systematic bias, overestimating the distance of pointed objects by 1.21 metres on average.
03Object recognition accuracy declined significantly at greater distances, dropping to 3.13% for a backpack at 5 metres.
STILL TO COME
How it was doneWhat they found
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
Tested on only one person with normal vision, not blind or low-vision users. The test was a technical validation of the software, not a usability study of how a human interacts with the device. The system overestimates distance, which is a safety-critical error for navigation. The device operates at 3 to 5 frames per second, which is too slow for real-time interaction. It can only recognise 80 specific object classes, missing many daily items.
The easy way to misread this
Do not assume this device is ready for clinical use or that it improves spatial awareness for blind patients. The test used a sighted participant and a laser pointer to verify the target, meaning it measured software accuracy, not human navigation ability. The 1.21-metre distance overestimation is a significant safety flaw that has not been resolved.
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 →