Quantification of Active Visual Attention using RGB camera.

Area of Science:

• Cognitive Science
• Human-Computer Interaction
• Psychophysiology

Background:

• Active visual attention (AVA) is crucial for human behavior and psychophysiological research.
• Current eye-tracking methods are costly and raise privacy concerns due to video recording.
• There is a need for accessible, real-time AVA measurement solutions.

Purpose of the Study:

• To develop an inexpensive and privacy-preserving method for measuring active visual attention (AVA).
• To utilize blink-rate variability (BRV) as a key indicator for AVA assessment.
• To enable real-time AVA monitoring using standard webcams.

Main Methods:

• Employed laptop cameras/webcams to capture visual data.
• Estimated head pose for camera alignment and screen-gaze detection.
• Calculated subject-specific thresholds using eye aspect ratio (EAR) to detect blinks and construct BRV signals.
• Ensured privacy by saving only EAR values, not face videos.

Main Results:

• Developed a novel active visual attention (AVA) score based on blink-rate variability (BRV).
• Demonstrated the robustness of the AVA score across different participants.
• Validated the score's reliability under varying ambient light conditions and occlusions (e.g., spectacles).

Conclusions:

• Blink-rate variability (BRV) offers a viable, low-cost, and privacy-conscious alternative for AVA measurement.
• The proposed method is suitable for real-time applications and various environmental conditions.
• This technique enhances accessibility for psychophysiological and human-behavior research.