How to choose the size of facial areas of interest in interactive eye tracking

Abstract

Advances in eye tracking technology have enabled the development of interactive experimental setups to study social attention. Since these setups differ substantially from the eye tracker manufacturer’s test conditions, validation is essential with regard to the quality of gaze data and other factors potentially threatening the validity of this signal. In this study, we evaluated the impact of accuracy and areas of interest (AOIs) size on the classification of simulated gaze (fixation) data. We defined AOIs of different sizes using the Limited-Radius Voronoi-Tessellation (LRVT) method, and simulated gaze data for facial target points with varying accuracy. As hypothesized, we found that accuracy and AOI size had strong effects on gaze classification. In addition, these effects were not independent and differed in falsely classified gaze inside AOIs (Type I errors; false alarms) and falsely classified gaze outside the predefined AOIs (Type II errors; misses). Our results indicate that smaller AOIs generally minimize false classifications as long as accuracy is good enough. For studies with lower accuracy, Type II errors can still be compensated to some extent by using larger AOIs, but at the cost of more probable Type I errors. Proper estimation of accuracy is therefore essential for making informed decisions regarding the size of AOIs in eye tracking research.

No evidence that gaze anxiety predicts gaze avoidance behavior during face-to-face social interaction

“…First, scene videos were analyzed by the facial landmark detection tool Open Face 52 to generate facial landmarks. Then the limited-radius voronoi tessellation method 53 with a radius of 2° was applied to receive Areas of Interest (AOI) for the eyes (adding up the right and left eyes), nose, and mouth 54 . Furthermore, we took an ellipse around the face as an additional AOI.…”

How robust are wearable eye trackers to slow and fast head and body movements?

“…More precisely, if the inaccuracy of the eye tracker is less than half of the smallest distance between the objects in the image, the eye tracker may be suitable for conducting an AOI study. Holmqvist, Nyström, & Mulvey (2012), Orquin & Holmqvist (2018), Vehlen, Standard, & Domes (2022) and Hessels, Kemner, van den Boomen, and Hooge (2016) discuss the size of the area of interest in relation to the accuracy and precision.…”

Social attention in the wild - Fixations to the eyes and autistic traits during a naturalistic interaction in a healthy sample

“…Lastly, the size of ellipses fitted to the face, eyes, and mouth was based on the height and width of the facial bounding box returned by the face detection algorithm. AOI size affects the classification of fixations on AOI's: We used comparatively large AOI's, which compensate for false negative classifications in situations of low accuracy, but at the cost of increasing false positives 67 . Choosing smaller AOI's may have resulted in less fixations classified as falling on the eyes and mouth and more fixations falling on the rest of the face and background, which potentially would have influenced results.…”