Abstract
For many applications in graphics, design, and human computer interaction, it is essential to understand where humans look in a scene with a particular task. Models of saliency can be used to predict fixation locations, but a large body of previous saliency models focused on free-viewing task. They are based on bottom-up computation that does not consider task-oriented image semantics and often does not match actual eye movements. To address this problem, we collected eye tracking data of 11 subjects when they performed some particular search task in 1307 images and annotation data of 2,511 segmented objects with fine contours and 8 semantic attributes. Using this database as training and testing examples, we learn a model of saliency based on bottom-up image features and target position feature. Experimental results demonstrate the importance of the target information in the prediction of task-oriented visual attention.
Highlights
For many applications in graphics, design, and human computer interaction, it is essential to understand where humans look in a scene with a particular task
An understanding of task-oriented visual attention is useful for automatic object recognition [1], image understanding, or image search [2, 3]
It can be used to direct visual search and foveated image video compression [4, 5] and robot localization [6, 7]. It can be used in advertising design or implementation of smart cameras [8]
Summary
For many applications in graphics, design, and human computer interaction, it is essential to understand where humans look in a scene with a particular task. It can be used to direct visual search and foveated image video compression [4, 5] and robot localization [6, 7]. It can be used in advertising design or implementation of smart cameras [8]. Judging the results of experiments by intuitive observation is not precise because different people might focus on different regions of the same scene, even with task To solve this issue, eye tracker equipment pieces that can record human eye fixation, saccades, and gazes are routinely used. There are over two dozen databases with eye tracking data for both images and videos in the public domain [9], which mainly focus on “free-viewing” eye movements
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have