The abstract is read at the publisher; the summary is JClub's.
Freebie, a real-time markerless multi-camera system for non-human primate finger tracking, achieves median end-to-end latency of 15.8 ms at 30 FPS and 16.2 ms at 60 FPS.
- Why it matters: Neural decoders for dexterous finger movements are limited by how accurately ground truth kinematics can be measured, and conventional instrumentation constrains movement dimensionality in NHPs.
- What they did: Two NHPs flexed fingers to move a virtual hand using bend sensors or Freebie-derived keypoints, with Kalman filter (KF) decoding from forearm EMG trained on either Freebie or bend-sensor labels.
- The result: Freebie keypoints closely matched bend sensor positions (median |r|=0.96) with median keypoint errors of 1.7–6.5 pixels, enabling online control comparable to bend sensors and similar EMG KF performance.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.