Motion tracking of iris features to detect small eye movements

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Neuroscience

Background:

• Current video-based eye trackers struggle to detect subtle eye movements like monocular microsaccades.
• Existing algorithms, reliant on pupil and corneal reflection, produce noisy data and low signal-to-noise ratios.
• This limits understanding of the prevalence and existence of monocular microsaccades.

Purpose of the Study:

• To introduce a novel video-based eye tracking methodology.
• To reliably detect small eye movements, specifically monocular microsaccades, with high confidence.
• To improve the accuracy and robustness of miniature eye movement recording.

Main Methods:

• Developed a new video-based eye tracking technique.
• Tracks the motion of stable, higher-order iris features (e.g., texture) instead of pupil or corneal reflection.
• Estimates eye movement velocity rather than precise position.

Main Results:

• Successfully detected small eye movements over 0.2 degrees (12 arcminutes) with high confidence.
• The new method yields a more robust and detailed record of miniature eye movements.
• Overcame signal artifacts and low signal-to-noise ratios inherent in current methods.

Conclusions:

• The novel eye tracking method reliably detects monocular microsaccades.
• Utilizing iris features for velocity estimation offers a significant improvement over position-based tracking.
• This advancement provides a clearer understanding of miniature eye movements and their role.