Accurate and Robust Eye Tracking with Ultrasound: A Computational Study
Published in 2023 IEEE International Ultrasonics Symposium (IUS), 2023
Abstract: This paper introduces an ultrasound simulation platform for generating realistic eye-tracking data, accounting for transducer design, sensor noise, occlusions, and headset slippage. Using a face/eye model with adjustable gaze and eyelid opening, synthesized data were input into a machine learning algorithm to estimate gaze and slippage. The system achieved a gaze RMSE of 0.085° without slippage and 0.756° with slippage. This end-to-end pipeline supports the optimization of wearable ultrasound devices and explores ultrasound as a complementary technology to camera-based eye tracking for AR/VR applications.
Recommended citation: N. Lu, et al. (2023). "Accurate and Robust Eye Tracking with Ultrasound: A Computational Study." 2023 IEEE International Ultrasonics Symposium (IUS) Proceedings.
Download Paper