A dynamic saliency attention model based on local complexity

Efficient saliency analysis based on wavelet transform and entropy theory

“…Inspired by the putative neural mechanism, and hypothesizing that some visual inputs are salient in certain background contexts, a center-surround mechanism was proposed to detect saliency across scales [18][19][20]. Moreover, local complexity [21] and local differences [22][23][24] to the surrounding areas are also employed to describe saliency. Among the local features, color information is the most important, and different color channels -including YCbCr [25], RGBYI [26], and Lab [27] -have been used to produce saliency maps.…”