Objective image quality measures for evaluating advanced MRI reconstruction methods

Abstract

We propose some new objective image quality measures for use with magnetic resonance images. These measures are based on the visual differences predictor (VDP) due to Daly (1992,1993). The VDP is a computer program that implements a complex model of the human visual system, and generates a map indicating the probability of the visual system perceiving a difference between two images on a pixel by pixel basis. A brief overview of the VDP algorithm is given, and the image quality measures we have developed based on the VDP are then described. The preliminary validation methods used to evaluate these image quality measures are then given. The results of these preliminary investigations indicate that these measures are promising as useful indicators of perceived image quality.