Investigation of camera-free eye tracking glasses compared to a video-based system

Poster Presentation 43.408: Monday, May 22, 2023, 8:30 am – 12:30 pm, Pavilion
Session: Eye Movements: Individual differences, novel measurement

Claudia Martin Calderon1 (), Abdullah Zafar1, Anne Marie Yeboah2, Kristine Dalton2, Elizabeth Irving2, Ewa Niechwiej-Szwedo1; 1University Of Waterloo, Faculty of Health, 2University Of Waterloo, School of Optometry & Vision Science

Eye movements provide insight into human cognitive and sensorimotor control. However, most research has been constrained to traditional laboratories. Recent advances in eye tracking technology could facilitate research and application in ecological settings such as aviation training or sideline assessment. It is important to confirm the accuracy of the emerging systems. Here we compare a system consisting of camera-free eye tracking glasses with a research grade video-based system. Eleven healthy participants (22.6±2.8 years) were tested using a saccade task (target range 3-25°). Eye movements were recorded simultaneously using a video-based system (EyeLink-2, SR Research) and a portable eye/gaze tracking system (Mindlink, AdHawk Microsystems) at 250 Hz. Calibration for each system was performed separately following guidelines (accuracy error was <1deg). The head was stabilized in a chinrest throughout the experiment. Results showed a mean amplitude difference between the systems of 0.05°±2.05 and mean peak velocity difference of 28.98°/sec ±65.33. Analyses of the results revealed two important factors associated with larger errors between the systems. First, greater error was found in individuals wearing contact lenses (n=4) compared to non-contact lens wearers (n=7) (amplitude: 0.24°±2.13 vs -0.05°±2.00; peak velocity: 45.96°/sec ±66.6 vs 20.32°/sec ±62.94). Second, linear regression analysis revealed Mindlink calibration precision explained 50% of variance in amplitude difference, and 0.3% variance in peak velocity difference between the systems. The Mindlink system offers unprecedented advantage for recording of eye and gaze movements in ecological settings. Our study provides important insight regarding two factors that could contribute to significant errors in saccade amplitude and peak velocity: 1) errors may be higher in individuals wearing contact lenses; 2) in addition to accuracy, Mindlink’s precision calibration variable has a significant impact on the quality of the recording. These results should be considered when adopting the Mindlink for research applications.

Acknowledgements: NSERC