Cellular neural/nonlinear network (CNN) models are introduced to analyze an
d explain human early vision color perception and some of the visual illusi
ons. A CNN model of Horn's algorithm is shown and a new lightness-determina
tion method is introduced. CNN models for the brightness, the Mach-band, th
e Zollner, the cafe-wall, and the Poggendorff illusions are also introduced
.