Color in context and spatial color computation

The purpose of this dissertation is to contribute in the field of spatial color computation models.We begin introducing an overview about different approaches in the definitionof computational models of color in digital imaging. In particular, we present a recent accurate mathematical definition of the Retinex algorithm, that lead to the definition of a new computational model called Random Spray Retinex (RSR). We then introduce the tone mapping problem, discussing the need for color computation in the implementation of a perceptual correct computational model. At this aim we will present the HDR Retinex algorithm, that addresses tone mappingand color constancy at the same time. In the end, we present some experiments analyzing the influence of HDR Retinex spatial color computation on tristimulus colors obtained using different Color Matching Functions (CMFs) on spectral luminance distribution generated by a photometric raytracer.