Patrick M. Pilarski

This paper introduces a new set of evaluation metrics for myoelectric prosthesis controllers and demonstrates their use in comparing deep learning and traditional controllers.

This paper develops and tests RCNN-based controllers for myoelectric prostheses to mitigate the limb position effect, showing improved performance over traditional methods in multiple metrics.

This study develops a method to generate accurate 3D gaze vectors by combining eye tracking and motion capture, finding that spherical coordinates and binocular data yield the best results.

This study validates the GaMA protocol's reproducibility across different sites and technologies for assessing visuomotor upper limb function in non-disabled participants.