Backpropagation of an image similarity metric for Autoassociative Neural Networks

Autoassociative Neural Networks (AANNs) are most commonly used fur image da ta compression. The goal of an AANN for image data is to have the network o utput be 'similar' to the input. Most of the research in this area use back propagation training with Mean-Squared Error (MSE) as the optimisation crit eria. This paper presents an alternative error function called the Visual D ifference Predictor (VDP) based on concepts from the human-visual system. U sing the VDP as the error function provides a criteria to train an AANN mor e efficiently, and results in faster convergence of the weights, while prod ucing an output image perceived to be very similar by a human observer.