Abstract

To learn the preferential visual attention given by humans to specific image content, we present NUSEF- an eye fixation database compiled from a pool of 758 images and 75 sub jects. Eye fixations are an excellent modality to learn semantics-driven human understanding of im- ages, which is vastly different from feature-driven approaches employed by saliency computation algorithms. The database comprises fixation patterns acquired using an eye-tracker, as sub jects free-viewed images corresponding to many semantic categories such as faces (human and mammal), nudes and actions (look, read and shoot ). The consistent presence of fixation clusters around specific image regions confirms that visual attention is not sub jective, but is directed towards salient ob jects and ob ject-interactions. We then show how the fixation clusters can be exploited for enhanc- ing image understanding, by using our eye fixation database in an active image segmentation application. Apart from proposing a mechanism to automatically determine characteristic fixation seeds for segmentation, we show that the use of fixation seeds generated from multiple fixation clusters on the salient ob ject can lead to a 10% improvement in segmen- tation performance over the state-of-the-art.