Complex neural networks as future tools in imagery analysis

Abstract

Brain networks are uniquely capable of generating and integrating information collected from multiple sources in real time. The application of structural and information theoretical measures to such networks has begun to unravel the crucial ingredients that ensure their rapid and robust performance. We suggest the use of information theoretical measures in applications that mimic some of these biological processing principles. We discuss candidate measures, their implementation in neural networks, their applicability to various sets of artificial and natural stimuli, and their future use in the automated analysis of aerial images.