Artificial neural networks and image interpretation: A ghost in the machine

Abstract

The artificial neural network (ANN) is a non-algorithmic parallel processing device modeled after human cerebral architecture. ANNs can be readily simulated on a digital computer and have the potential for significant diagnostic application in both image- and non-image-based medical diagnosis. This paper briefly discusses the operation of ANNs, differences between these and more traditional problem solving methods, and some diagnostic imaging problems amenable to their application. © 2004 Elsevier Inc. All rights reserved.