Morphological multiscale contrast approach for gray and color images consistent with human visual perception

Some algorithms for morphological multiscale contrast enhancement for gray and color images, based on a psychophysical model of contrast perception, are introduced. The color space for applying the transformations is the space of chromatic coordinates $(u,v,Y)$. The brightness $Y$ of the space is transformed separately to increase its contrast; subsequently, the gravity center law is used to enhance hue. One of the algorithms applied to $Y$ is based on a composition of contrast mappings that are built by means of the openings and closings by reconstruction. This composition successfully finds the principal regions according to a morphological contrast criterion. This contrast criterion is consistent with Weber’s law. The second approach is derived from the Retinex approach that uses a family of Gaussian transformations. Instead of working with Gaussian transformations, a family of openings by reconstruction is used. Again, Weber’s law is used for building the operators, although this law is not used as a criterion for selecting the primitives but as a concept for building rational morphological operators. For color images, once the brightness is enhanced, the color is improved by increasing the saturation of the image.