Eye Tracking Analysis Using Differently Shaped Areas of Interest to Represent Multi- Element Moving Objects

Abstract

Eye tracking analysis can be used to evaluate human performance. However, it is a challenging task to map eye tracking data to multi-element moving objects if the objects take different complex shapes and dynamically move within the display. We propose algorithms that approximate multi-element moving objects into simpler areas of interest (AOIs). Based on minimum AOIs required to represent objects, three AOI types (rectangle, circle, and triangle) were created that maintains the same sizes regardless of object shapes. The algorithms were applied to a task of observing a simulated air traffic control display. Two types of eye fixation data were used for evaluation: Simulated data using uniform distribution and human participant data. The number of eye fixations differed significantly among AOI types when simulated data were used (p < .001) but were insignificant when human data were used (p = .94). We discuss how eye tracking data may be affected by AOI types, and we offer suggestions to improve simulated eye fixations and AOI designs for multi-element moving objects. This pilot study provides ground work for better AOI design and possibly for automating the development of accurate AOIs for multi-element moving objects.