Morphological sampling based on foveal mechanisms

摘要

Research realized in the field of human vision offers a tremendous potential for computer vision. In our opinion, if we are expected to write algorithms and to build models for image processing, it is necessary to take into account some features of the organization of the human visual system. In that way we have chosen to base our investigation on the retina. Indeed the retina is the first neuronal structure involved in visual perception and it is largely studied by neurobiologists as mentioned by Wandell (1995). One goal of this paper is to show how the knowledge of a particular organ implied in vision can be used to conceive new sampling operators (space-variant and morphological) to realize selective analysis of a scene.