DigestAI news desk

Cut through the AI noise.

Research

GaitVista reduces gait measurement error by 27.7% in lab tests

Researchers introduced GaitVista, a reliability-aware measurement layer designed to improve the accuracy of 3D gait assessment using small camera sets and body-worn sensors. The system addresses the issue that sensing failures in these accessible setups can be mistaken for actual changes in a patient's walking function. By assigning specific visual contributions based on camera coverage and…

1 source primary source

Key points

  • GaitVista reduces average full-body gait error by 27.7% on the TotalCapture dataset.
  • The system narrows the error gap to an oracle from 2.76–5.33 cm to 1.11 cm.
  • Tests used healthy participants in controlled settings, not clinical patient data.

In tests on the TotalCapture dataset, GaitVista reduced average full-body and lower-body error by 27.7% and 27.8%, respectively. It also narrowed the gap to an ideal oracle from 2.76–5.33 cm down to 1.11 cm. On the MoVi dataset, it was the only deployable fusion method to outperform both unimodal streams, reducing marker-supported error by 6.4% compared to the strongest learned fusion baseline. The system also improved bilateral knee-flexion waveform accuracy by 18.9%.

The authors note that the current benchmarks involve neurologically healthy participants in controlled settings with specific sensor calibration. Consequently, the paper frames the results as progress toward accessible assessment rather than validated clinical deployment for patients.

Read the original at arXiv cs.AI · by Nethmi Jayasinghe, Mihir Parashar, Amit Ranjan Trivedi primary sourceOpen source ↗
Topics · follow one to build your own front page
GaitVista

The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.

Comments

via GitHub Discussions

More in Research

All →

Related stories