Pop-Out: A New Cognitive Model of Visual Attention That Uses Light Level Analysis to Better Mimic the Free-Viewing Task of Static Images

Figure 1

The global architecture of pop-out. The system can switch between bottom-up and perceptual fusion processes by changing the value of . The initial image is taken from the MIT dataset [31]. The first step consists in extracting the image intensity and components that will be enhanced during contrast preprocessing. Before contrast enhancement, is also used for light level analysis in order to detect in which lighting conditions the initial image was captured (the main goal of this analysis is to know if the image is taken in photopic, mesopic, or scotopic conditions). After contrast enhancement of , we obtain that will be used for Log-Gabor filtering in order to get the texture of the scene. image enhanced is also used to extract and components. After Log-Gabor filtering of and components, we get and that are, respectively, the blue and the red color information of the image texture. According to the result of the light level analysis (photopic, mesopic, or scotopic condition detected), different weights are given to   and . To obtain the , and are used as a mask that is combined with the luminance texture. The represents the visual attention predicted in case of free-viewing task. when we just use the luminance texture as visual attention map (case of ). The is just based on the edge sensitivity or the luminance texture. The takes into account the color components filtered ( and ).