Cellular Nonlinear Networks to Analyze the Complexity of Foggy Paintings

In this paper a new nonlinear signal processing technique, able to remove fog from color images exploiting the optical properties of the fog effect, is introduced. It is based on the use of cellular nonlinear networks (CNNs), a paradigm for nonlinear parallel image processing. The technique is applied to the analysis of several paintings by famous artists of different art movements, in which the presence of fog significantly affects the visual experience. The technique introduced distinguishes whether the fog is a natural element of the painted landscape or if it has been artificially added by the artist to express his or her own feelings.