Vanishing Point Attracts Eye Movements in Scene Free-viewing

TL;DR

This paper demonstrates that vanishing points in scenes attract human eye movements and introduces a combined model of saliency and vanishing point information that improves gaze prediction accuracy.

Contribution

The study reveals the influence of vanishing points on eye movements and proposes a novel combined model outperforming existing saliency models.

Findings

Vanishing points attract human gaze in scene viewing.

The combined model outperforms state-of-the-art saliency models.

Vanishing point information enhances scene understanding.

Abstract

Eye movements are crucial in understanding complex scenes. By predicting where humans look in natural scenes, we can understand how they percieve scenes and priotriaze information for further high-level processing. Here, we study the effect of a particular type of scene structural information known as vanishing point and show that human gaze is attracted to vanishing point regions. We then build a combined model of traditional saliency and vanishing point channel that outperforms state of the art saliency models.