Automated Computer Analysis of Radiographic Images

Abstract
Lee B. Lusted in his Memorial Fund Lecture (1) anticipated the device: “an electronic ‘scanner-computer’ to look at chest photofluorograms and to separate the clearly normal chest films from the abnormal chest films. The abnormal chest films would be marked for later study by the radiologist.” This automated computer analysis concept was based on some preliminary studies by Pendergrass and Tolles who showed that automatic scanning of photofluorograms could produce satisfactory density tracings. We are now reporting a method of automated computer analysis of radiographic images utilizing algorithms (a rule of procedure for solving a recurrent mathematical or logical problem) for recognizing parts of the radiographic image. Method and Results Thirty-seven 70 mm. photofluorograms were digitized and put on magnetic tape, using an improved image-scanning system similar to one previously reported (2). The radiographic image of the chest on magnetic tape contains 502 digit samples per line times 320 horizontal s...

This publication has 1 reference indexed in Scilit: