Most persons, perhaps, think of computers as devices which write pay checks or calculate satellite orbits. Although the processing of pictorial information promises dramatic and significant contributions to the health sciences and many other fields, it is only in connection with recent deep space probes that most persons have heard the words "computer" and "picture" used together. This discussion will concern the use of computers to detect characteristics and measure specific parameters in pictures. These measurements may then be analyzed, to produce "answers" in the form of conventional probabilistic statistics, such as the likelihood of occurence by chance. Far more complex taxonomic methods 1 are being developed and may be expected to be of importance in the future. However, these usually require extensive use of prediagnosed material in order to "teach" the program the characteristics of desired concepts such as "normal" or "abnormal." Although in the future we envision widespread