Real-Time EOG Signal Baseline Drift Estimation Using Passive VOG Data.

Area of Science:

• Biomedical Engineering
• Neuroscience
• Human-Computer Interaction

Background:

• Electrooculography (EOG) is crucial for eye gaze tracking in human-computer interfaces.
• Baseline drift in EOG signals causes inaccurate gaze estimation, hindering application usability.
• Existing drift mitigation techniques are often limited to offline processing.

Purpose of the Study:

• To develop a novel real-time drift mitigation technique for EOG signals.
• To enable accurate eye gaze tracking without requiring users to fixate on targets.
• To preserve the original signal morphology during baseline correction.

Main Methods:

• A real-time drift mitigation technique using passive videooculography (VOG) was developed.
• The VOG signal models and removes the EOG baseline drift.
• The method operates without user-initiated target fixations.

Main Results:

• The proposed technique effectively de-drifts EOG signals in real-time.
• Performance was evaluated against standard methods using data from ten subjects.
• Demonstrated improved accuracy in horizontal and vertical gaze angle estimation.

Conclusions:

• The novel real-time EOG de-drifting technique offers improved performance.
• This approach enhances the reliability of EOG-based eye gaze tracking for HCI.
• The method preserves signal integrity while correcting for baseline wander.