Digit recognition in a simplified visual cortex model

Abstract

In this paper a simplified visual cortex model is developed and application of the model to digit detection is explored. Retina and Lateral Geniculate Nucleus (LGN) is modeled as a combined 64×64 input cell array. An applied digit pattern produces Poisson spikes at the output of the combined retina-LGN cell array. Retina-LGN cells are connected to primary visual cortex (V1) cell network. Spike rate pattern of any digit applied to the input is used in digit detection.