High-frequency drift coverage for mocap calibration

Develop a calibration approach that covers high-frequency drift in less stable hand-motion-capture hardware, tied to appropriate sensor characterization.

Background

YOCO assumes that mocap bias is approximately stationary within a recording session and uses a calibration HyperNet conditioned on a small set of paired poses. The paper notes that this assumption may fail for less stable hardware exhibiting high-frequency drift, which cannot be addressed adequately through a one-time stationary calibration.

The synthetic augmentation pipeline is designed to model diverse but temporally consistent drift patterns. Extending coverage to high-frequency drift therefore remains unresolved and depends on characterizing the corresponding sensor behavior.

References

The augmentation pipeline lets us scale training across diverse drift patterns at low cost, though covering high-frequency drift remains an open direction tied to sensor characterization.

— YOCO: You Only Calibrate Once! Fast Mocap Calibration for Dexterous Teleoperation  (2610.11657 - Zhang et al., 8 Oct 2026) in Conclusion, Limitations paragraph