24-Gaze-Point Calibration Method for Improving the Precision of AC-EOG Gaze Estimation

Area of Science:

• Biomedical Engineering
• Ophthalmology
• Human-Computer Interaction

Background:

• Gaze estimation using Electro-Occulo-Graphy (EOG) is crucial for various applications.
• Conventional AC-EOG methods often rely on simplified transformations of EOG signals.
• Existing methods face limitations in precision due to how calibration data is processed.

Purpose of the Study:

• To enhance the precision of Alternating Current Electro-Occulo-Graphy (AC-EOG) gaze estimation.
• To develop a more accurate conversion method for EOG signals to gaze position.
• To compare the performance of the proposed method against conventional AC-EOG techniques.

Main Methods:

• Proposed a novel gaze coordinate conversion method using an imaginary center (25th point) for affine transformation of 24-point calibration data.
• Implemented and compared the proposed method with the conventional Quadrant method for AC-EOG gaze estimation.
• Utilized cross-shaped and plus-shaped electrode attachments for data collection and analysis.

Main Results:

• The proposed method demonstrated reduced average eye angle error compared to the conventional method.
• For plus-shaped electrode attachment, the proposed method achieved an average eye angle error of x = 0.94° ± 0.19° and y = 1.48° ± 0.27°.
• The conventional method with cross-shaped electrode attachment resulted in errors of x = 2.27° ± 0.46° and y = 1.83° ± 0.34°.

Conclusions:

• The proposed method provides a simpler and more precise approach to EOG gaze estimation.
• The new conversion technique improves the accuracy of translating eyeball movement signals into gaze positions.
• This advancement has potential implications for more reliable eye-tracking systems.