Fixation sequences made during visual examination of briefly presented 2D images.

TLDR: It is shown that eye-movement traces made by different observers in response to the same image have few common temporal sequences involving the same fixation locations, even for sequences of only two fixations, and that the temporal sequence in which fixations are made is not a significant factor in the analysis of the eye- Movement data.

Abstract: Eye movements made by eighteen observers in response to brief (3 s) presentations of eleven different images, each in three forms (unfiltered, high-pass filtered and low-pass filtered), have been analysed in order to identify both repeated sequences of fixations and image locations which attract re-fixations. It is shown that eye-movement traces made by different observers in response to the same image have few common temporal sequences involving the same fixation locations, even for sequences of only two fixations. There is a greater incidence of such sequences in eye-movement traces made by the same observer in response to two presentations of the same image, but average numbers are still low. Conserved sequences involving more than two identical locations occur at a much lower frequency, and the incidence of repeated sequences is not increased if consideration is restricted to regions of the image which attract large numbers of fixations. It is concluded that the temporal sequence in which fixations are made is not a significant factor in the analysis of the eye-movement data considered in this report. Calculations based on a least squares index of similarity are consistent with this conclusion. The analysis shows a relatively high incidence of re-fixation on certain locations in the images and there is evidence that such re-fixations are a significant factor in the high similarity between fixation locations established by different observers when viewing the same image.