Optimizing Fixation Filters for Eye-Tracking on Small Screens

Area of Science:

• Human-Computer Interaction
• Applied Neuroscience
• Consumer Psychology

Background:

• Eye-tracking is increasingly used to study consumer responses to advertising.
• Traditional eye-tracking calibration and validation occur on large, static screens.
• Limited understanding exists regarding eye-tracking precision on small screens (smartphones) and in dynamic, scrolling environments (social media).

Purpose of the Study:

• To evaluate the precision of eye-tracking fixation detection algorithms on mobile devices.
• To investigate the impact of different software parameters on eye-tracking metric validity and reliability.
• To propose recommendations for improving eye-tracking data accuracy in mobile and social media contexts.

Main Methods:

• Testing the precision of eye-tracking fixation detection algorithms against raw gaze mapping.
• Analyzing performance in natural scrolling conditions on smartphones.
• Examining the influence of various eye-tracking software parameters on key metrics like Percent Seen and Total Fixation Duration.

Main Results:

• Default fixation detection algorithms show suboptimal performance on mobile phones.
• Specific parameter adjustments significantly affect the validity and reliability of eye-tracking metrics.
• Proposed algorithm adjustments improved Percent Seen accuracy by 19% compared to a leading provider's defaults.

Conclusions:

• Standard eye-tracking fixation detection algorithms require optimization for mobile devices and social media feeds.
• Methodological framework can be adapted for validating other applied neuroscience methods.
• Recommendations are provided for enhancing eye-tracking metric accuracy on small screens and in moving environments.