Figure 2.

(A) Visual saliency model. Features are extracted from the input image. The center-surround mechanism and normalization are used to generate the individual feature saliency maps. Finally, the saliency map is generated by a linear combination of different individual saliency maps (adapted from Itti et al., 1998); (B) Auditory saliency model. The structure of the model is similar to the visual saliency model by converting sound inputs into a frequency “intensity image” (adapted from Kayser et al., 2005).