Lecture to American College of Radiology

Image processing applied to medicine is more constrained than in other fields. Lacking means of verification, as is often the case, it is difficult to judge what constitutes a success or how the ideal image should look. Most successful processing efforts to date have been directed at correcting distortions introduced by shortcomings in the imaging equipment itself or methods for filling in missing data sets. Some important headway has been made in processing which is directed at providing a better match to human visual psychophysics particularly in matrix images. The greatest future opportunities appear to lie in image integration of several different imaging modalities of the same body part and utilization of a priori information as a factor in image formation.