Object-Gaze Distance: Quantifying Near- Peripheral Gaze Behavior in Real-World Applications

Area of Science:

• Human-Computer Interaction
• Computer Vision
• Cognitive Science

Background:

• Traditional eye tracking (ET) primarily analyzes foveal vision, neglecting peripheral gaze behavior.
• Existing ET methods struggle to quantify visual attention in complex, real-world scenarios.

Purpose of the Study:

• To enhance the Object-Gaze Distance (OGD) method for quantifying near-peripheral gaze behavior.
• To improve the analysis of visual attention in complex environments by integrating peripheral vision data.

Main Methods:

• Developed an algorithmic enhancement to the OGD method for eye tracking analysis.
• Utilized machine learning for Area of Interest (AOI) detection.
• Computed 2D Euclidean pixel distance to create continuous gaze-based time-series data.

Main Results:

• Incorporating near-peripheral vision increased interpretable fixation data significantly: 23.8% to 78.3% for AOI screw, and 4.5% to 67.2% for AOI screwdriver.
• The enhanced method achieved a substantial increase in quantifiable gaze data during a surgical procedure.
• Multi-OGD time series analysis revealed novel gaze patterns.

Conclusions:

• The proposed algorithmic enhancement provides a more accurate depiction of human gaze behavior by including near-peripheral vision.
• This advancement offers richer insights into visual attention dynamics in multi-object environments.
• The method has potential applications in fields requiring detailed analysis of visual behavior, such as surgical training and human-computer interaction.