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 4

(a) Initial image. (b) Luminance texture: . (c) “Intuitive space” (color texture): . We are in photopic conditions; more weights are given to the red wavelength magnitudes (). means the binary complement of . (d) Binary complement of the weighted sum of and components: . (e) after closing. (f) Regions of interest according to spectral light sensitivity . We note that when we complement the weighted sum of and components (), we obtain a special colorimetric space reflecting the relative spectral sensitivity of each color (see (d)). In this space white area contains low luminous efficiency and describes the wavelength intensity of less bright colors (since we all know that white is the combination of all colors). In our “intuitive space” (see (c)), dark red colors combined with the dark blue give black area (which corresponds to the useless white area in (d)) which are less bright than orange, for example (in fact, in photopic vision or when we do not consider component, orange can be considered as an enlightened red wavelength; for a painter, it is also obvious that black color can be seen as the merging of red and blue). In (e), we apply a closing (mathematical morphology) for grouping all whitish regions of the scene (the useless one or the less bright one). Whitish areas in are the regions that will attract our attention (f).

(a)

(b)

(c)

(d)

(e)

(f)