Toward a unified view of radiological imaging systems. Part II: Noisy images

"The imaging process is fundamentally a sampling process." This philosophy of Otto Schade, utilizing the concepts of sample number and sampling aperture, is applied to a systems analysis of radiographic imaging including some aspects of vision. It leads to a simple modification of the Rose statistical model; this results in excellent fits to the Blackwell data on the detectability of disks as a function of contrast and size. It gives a straightforward prescription for calculating a signal-to-noise ratio, which is applicable to the detection of low-contrast detail in screen-film imaging, including the effects of magnification. The model lies between the optimistic extreme of the Rose model and the pessimistic extreme of the Morgan model. For high-contrast detail, the rules for the evaluation of noiseless images are recovered.