Abstract
Pupil size is correlated with important cognitive variables and is increasingly being used to study cognition. Pupil data can be recorded inexpensively and non-invasively by many commonly used video-based eye tracking cameras, but researchers often underestimate the methodological challenges associated with controlling for confounds that can result in misinterpretation of their pupil data. One serious confound that often is not properly controlled is pupil foreshortening error (PFE)—the foreshortening of the pupil image as the eye rotates away from the camera. Here we report two studies. The first study formally established a strong ratio scale between the "arbitrary" pupil units reported by the EyeLink 1000 eye tracker and known physical units. The second study systematically mapped PFE as a function of gaze position using an artificial eye model and then applied a geometric model correction. Data were collected across 3 experimental layouts of the eye tracking camera and display using 3 spherical artificial eyes with different fixed pupil diameters. The 9 resulting maps showed large PFE that increased as a monotonic function of the oblique angle between the eye-to-camera axis and the eye-to-target axis and was invariant across the 3 different artificial eye pupil diameters. The measured PFE was corrected using a geometric model that expressed the foreshortening of the pupil area as a function of the cosine of the angle between the eye-to-camera axis and the eye-to-target axis. The model reduced the root mean squared error of pupil measurements by 82.5% when the model parameters were pre-set to the physical layout dimensions, and by 97.5% when they were optimized to fit the empirical error surface. Our data and correction procedure provide an unprecedented reduction in PFE while preserving the freedom to study tasks such as reading or visual search that involve free viewing of the display.
Meeting abstract presented at VSS 2015