A Review of Attention Models in Image Protrusion and Object Detection

Modelling in visual attention especially the stimulus-driven one, i.e. saliency-based attention,has been a very active research field during the recent 25 years. There are many attentionmodels which, apart from being in other aspects, have been offered in successful functions ofcomputer vision, moving robots, and cognitive systems. The present article surveys theprimary concepts of visual attention, implemented in cognitive, Bayesian network, decisiontheories, and information theory in a computational perspective. It will demonstrate acategorization that provides a critical comparison of the approaches as well as their abilitiesand results. Specifically, the article formulates the criteria, derived from computationalbehaviors and studies in order to compare the quality of visual attention models.

کلیدواژه ها:

visual attention ، bottom-up attention ، top-down attention ، saliency ، eye movement ، regions of interest ، visual search