Application of head flexion detection for enhancing eye gaze direction classification

Area of Science:

• Computer Vision
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Eye gaze and head movement tracking are crucial for developing alternative interfacing devices.
• Existing methods for eye gaze classification have limitations, particularly with rotated facial images.

Purpose of the Study:

• To propose enhancements for accurate eye gaze direction classification.
• To improve the accuracy and robustness of eye-tracking systems for interfacing devices.

Main Methods:

• Utilized Viola Jones face detector for eye region identification.
• Employed Circular Hough Transform for precise iris localization.
• Applied Support Vector Machine (SVM) for eye gaze classification.
• Integrated head flexion angle calculation using a microcontroller and flex sensors to correct for head rotation.

Main Results:

• The proposed system demonstrates enhanced accuracy in eye gaze direction classification.
• Head movement detection effectively compensates for facial rotations, overcoming limitations of traditional face detectors.
• The integration of head direction significantly improves the overall control method's effectiveness.

Conclusions:

• Head direction is a key determinant in enhancing eye gaze-based control systems.
• The combined classification of eye gaze direction and head pose offers a more robust and accurate interfacing solution.
• This research advances the development of advanced human-computer interaction technologies.