Neural networks for the simulation of a visual system

In this paper, we present a general system for image interpretation based on neurobiological and physiological concepts. All processing is performed by neural networks. Our system emulates a robot with a single eye. Its architecture solves the problem of communication between low and high level processes in picture analysis. In a classical scheme of image understanding, good segmentation is necessary to obtain clear interpretation. In our neural network approach, we use the concept of mental images. Noisy pictures and partially occluded objects are clearly recognized.