How a distractor influences fixations during the exploration of natural scenes.

Hélène Devillez,Nathalie Guyader,Anne Guérin-Dugué

doi:10.16910/jemr.10.2.2

Hélène Devillez, Nathalie Guyader + Show 1 more

Open Access

https://doi.org/10.16910/jemr.10.2.2

Copy DOI

Abstract

The distractor effect is a well-established means of studying different aspects of fixation programming during the exploration of visual scenes. In this study, we present a taskirrelevant distractor to participants during the free exploration of natural scenes. We investigate the control and programming of fixations by analyzing fixation durations and locations, and the link between the two. We also propose a simple mixture model evaluated using the Expectation-Maximization algorithm to test the distractor effect on fixation locations, including fixations which did not land on the distractor. The model allows us to quantify the influence of a visual distractor on fixation location relative to scene saliency for all fixations, at distractor onset and during all subsequent exploration. The distractor effect is not just limited to the current fixation, it continues to influence fixations during subsequent exploration. An abrupt change in the stimulus not only increases the duration of the current fixation, it also influences the location of the fixation which occurs immediately afterwards and to some extent, in function of the length of the change, the duration and location of any subsequent fixations. Overall, results from the eye movement analysis and the statistical model suggest that fixation durations and locations are both controlled by direct and indirect mechanisms.

Highlights

Human vision during visual perception is an active process where observers select information relevant to their exploration goal (Findlay & Gilchrist, 2003)
The distractor effect was measured on both fixation durations and locations based on the fixation order in scene exploration
Analysis of the distractor effect took into account the saliency of the scene, computed by empirical saliency maps

Summary

Introduction

Human vision during visual perception is an active process where observers select information relevant to their exploration goal (Findlay & Gilchrist, 2003). When viewing a visual scene, we typically make three to four eye fixations per second. Decisions are made regarding the fixation, and the saccade is programmed. This programming involves the decision to terminate the current fixation (when) and the choice of location for the fixation (where). Both characteristics have been widely studied, most of the time separately, in order. How a distractor influences fixations during the exploration of natural scenes.

Objectives

Methods

Results

Conclusion