Real-time estimation of horizontal gaze angle by saccade integration using in-ear electrooculography

Area of Science:

• Biomedical Engineering
• Neuroscience
• Signal Processing

Background:

• Conventional eye gaze tracking often relies on bulky equipment or invasive methods.
• Electrooculography (EOG) offers a potential alternative for non-invasive eye movement detection.
• Utilizing ear canal EOG signals presents an opportunity for discreet and portable gaze sensing.

Purpose of the Study:

• To develop and evaluate a real-time algorithm for estimating horizontal eye gaze angle.
• To assess the feasibility of using single-channel electrooculography (EOG) from the ear canal for gaze tracking.
• To compare the performance of the novel algorithm against conventional EOG signal processing techniques.

Main Methods:

• A real-time algorithm was developed to estimate eye gaze angle from single-channel EOG signals.
• The algorithm employed statistical analysis of detected saccades, differing from traditional high-pass filtering.
• Conductive ear moulds were used to capture EOG signals directly from the ear canal.

Main Results:

• The developed algorithm provided noisy estimations of eye position.
• Performance, measured by Pearson product-moment correlation coefficients, was significantly better than conventional methods in certain cases.
• In-ear EOG signals demonstrated potential for portable horizontal eye gaze estimation.

Conclusions:

• In-ear EOG signals captured via conductive ear moulds can form the basis for lightweight, portable eye gaze estimation.
• The proposed algorithm offers a viable alternative to conventional methods, particularly for horizontal gaze tracking.
• This technology could enable applications such as hearing aids that adjust microphone directivity based on user eye gaze.