Results of hierarchical saliency model.

<p>(a) Input images. (b) Ground truth images. (c) Itti <i>et al</i>.'s visual saliency maps, calculated from (a) with all the low level features. (d) Salient region images extracted from (c) using Otsu's global thresholding algorithm. (e) Itti <i>et al</i>.'s visual saliency maps, calculated from (d). (Copyrights of those figures are listed in Acknowledgments.).