A bioinspired model of early visual processing with hue-feature saliency for a cognitive architecture
Cynthia Ávila-Contreras, Daniel Madrigal, Félix Ramos, Juan Luis del Valle‐Padilla · Procedia Computer Science · 2018
We present a computational model that describes the early stages of visual processing with selective attention mechanisms to generate feature-based (hue or color) activations of salient localizations, based on neurophysiology evidence of activations and inhibitions in visual pathway: photoreceptors in retina, simple opponent cells in thalamus with increases by top-down spatial-related signals, double opponent cells in and its lateral inhibitions to determine salient regions. The model is part of the selective attention aspect for a broader cognitive architecture, which focus in the identification of related brain areas, the likely operations of each one, and the type of data generated and shared among the components. Such approach aims to develop a system able to grow as more components and operations from other cognitive functions will be integrated, keeping its biological and computational feasibility.