A novel visual attention model using multi-scale cues

A novel visual attention model using multi-scale cues is presented in this paper. Visual data is decomposed into multi-scale sub-images which contain multi-scale details on the basis of Gaussian Pyramid, and contrast features are extracted from these multi-scale images for saliency map generation. Compared with other visual attention models, the proposed model can efficiently generate saliency map with better visual effects of integrated contour and inner region. Experimental results on various types of images achieved better performance, which demonstrates the effectiveness and efficiency of the proposed method.