A television image processing system for diagnostic radiology using the techniques of contrast enhancement, aperture correction, and highlight inversion is described together with a method for objectively evaluating its usefulness. The evaluation is done by comparing error rates of observers viewing a specially prepared set of chest radiographs through the image processor with error rates using a conventional viewbox. The error rates of different processing techniques are given and conclusions about the causes of these error rates are drawn. These conclusions, in turn, yield guides for the future design of circuits to improve the performance of image processors.