New models of visual saliency: Contourlet transform based model and hybrid model

Abstract

Modeling of visual attention is a very active research domain which has attracted the attention of researchers over the past years. Several models of saliency detection are now available that have shown successful applications in various fields. In this paper, we initially present three models already existing in the literature, we are talking about Itti's model, Imamoglu's model and Le Moans's model. Then, we present our proposal models in order to improve the results obtained from these three models: the first one is based on the Contourlet Transform and the second is an Hybrid model which combines Imamoglu's model and Le Moans's model.