The Y Combinator startup measures motion and fingertip force in real time, delivering 50 predictions each second. Its tracking model maintains an average joint-angle error of 4.58 degrees, verified on held-out recordings alongside Apple Vision Pro data. When putting the wearable back on, a user must provide two minutes of motion data so the system can finish recalibrating in under 60 seconds.
Key points
Performance results were validated using held-out comparison recordings captured on the A…: Performance results were validated using held-out comparison recordings captured on the Apple Vision Pro.
Re-donning the wearable requires two minutes of active user movement to prepare for calib…: Re-donning the wearable requires two minutes of active user movement to prepare for calibration.
The automated system completes its full re-calibration sequence in less than one minute.: The automated system completes its full re-calibration sequence in less than one minute.
Sources checked by our analysis pipeline: 1
The hand disappears.The sleeve still knows.
Individual finger movements and fingertip pressure, tracked in real time from electrical signals in the forearm. No cameras, no gloves, just a sleeve.
Watch on LinkedIn- average joint-angle error
- 4.58°
- predictions per second
- 50
- of training data
- Only 78 min
Validated against Apple Vision Pro hand tracking on held-out recordings. After re-donning the sleeve, two minutes of movement re-calibrate the model in under a minute.