The Loo and Doi paper reported that the signal-to-noise ratio of the matched filter model has a strong correlation with human visual detectability. Therefore, it can be thought that a human visual system is completely a self-matched situation or near this in relation to it's MTF and internal noise. In the upper Same condition of a visual system, optimum filter functions of image processing for more increasing detectability were derived. After studying a sufficient number of signal and noise patterns before detecting signals, if a human visual system is completely self-matched, the optimum filter function is the whitening noise filter. The whitening noise filter including an inverse number of visual MTF works as a sub-optinum filter, in the case of non-self-matched visual system. On the occasion of not knowing the to signal and noise, the matched filter including inverse number of a visual MTF is the optimum filter. As an application of the matched filter, for eliminating the background of an image, a similar subtraction filter function was derived.
The recent developed anti-crossover film provides more improved sharpness, but less granurality. In that sense a monistic evaluation using conventional individual evaluating methods is difficult. Used with two kinds of screens, high-speed and low-speed, we compared the image characteristics of the crossover effect regarding anti-crossover film and conventional film of similar speeds, designed for a total evaluation of the image quality by introducing a number of noise- equivalent quanta (NEQ(u)). On the same system of the screen, but at a high frequency anti-crossover film possesses superior NEQ(u), when crossover film at a low frequency, but at a high frequency anti-crossover film possesses superior NEQ(u), when compared with conventional film. These results were in good agreement with the results that had been obtained from images of needles and beads, or ROC analysis.
Modulation transfer functions (MTF's) of the screen-film systems which were measured at Kumamoto University, were compared with results obtained at the University of Chicago. There were two categories for the measurements. In the first category, the experimental results of MTF's analyzed at Kumamoto University using slit images and sensitometric strips prepared at the University of Chicago were compared with the results obtained at the University of Chicago. In the second category, measurement results of MTF's using slit images and sensitometric strips which were obtained at Kumamoto University with screens with the same brand name as those used at the University of Chicago with different serial numbers were compared with the University of Chicago's data. The results indicate that the MTF data for screen-film systems in both categories can be reproduced at different laboratories, within the experimental error.