Abstract: The retina is constituted by an array of cells formed on a substrate and each comprising a photosensitive element, means for converting a signal delivered by the photosensitive element into a binary information, storage means connected to the converting means and an elementary processor formed on the substrate and connected to the storage means. The storage means are arranged as a bidimensional array of shift registers to allow transfer of binary information from one cell to the neighboring ones, and the shift registers and elementary processors are connected in parallel to control inputs so as to control the acquisition of an image by simultaneously storing of the binary informations, the realization of simultaneous and parallel processings on the stored binary informations, and the write-in of the results in the storage means. Thus, successive processing steps of the image can be performed on the retina itself, the image being re-written on itself after each step.