A model for image formation and image quality prediction in diagnostic radiology

Abstract

A computer program modelling the formation of radiological images and predicting values of physical quantities that determine diagnostic image quality has been developed. The quantities are the modulation transfer function (MTF), the noise power spectrum (NPS), and the detective quantum efficiency (DQE) associated with signal to noise ratio (SNR). The program is based on mathematical models that describe the effects of x-ray interactions with the imaged object and the image detector as well as phenomena concerning the optical signal propagation within the detector. All data on x-ray effects necessary for computer calculations were derived from published work whereas optical data were determined in our laboratory using experimental techniques. Model predictions were compared with direct quality measurements performed after image formation, on the image itself.