Suguru Uchida

Extended random-dot model

The random-dot model has been extended to the analysis of photographic granularity that contains grain aggregations. In this model grains can have arbitrary transmittance profiles and make circular-shaped aggregations. Doubly stochastic Poisson processes are used to construct the model, and statistical properties of the model are investigated theoretically. Rigorous expressions for average transmittance and transmittance autocorrelation function are obtained explicitly. For circular and monosized grains, simple expressions of average transmittance and transmittance autocorrelation function are obtained approximately. By means of numerical examples of the results, the influences of grain aggregations on granularity properties are discussed.

Assessment of Radiological Image Quality by Redundancy: Subject Contrast

The relationship between the subject contrast and the total image performance of a radiograph can be assessed by a single number. In this study, the radiological image quality containing object information in two screen-film systems was assessed by the redundancy, C(Y). The quantity C(Y) in the LTII-QS system has a lower value at low subject contrast but a higher value at high subject contrast, compared with the LHII-RX system. The value of C(Y) is affected more by the radiographic mottle of the system at low subject contrast and more by the MTF of the system at high subject contrast.

Evaluation of Radiographic Images by Entropy: Application to Development Process

A method for the quality evaluation of radiographic images in terms of entropy is presented. By this method, the image quality can be synthetically evaluated by a single number. The method presented is used to calculate the amount of information contributed by the image of a uniform lucite step-wedge object. The new method is also applied to the evaluation of development processes. The calculated results show that the information quantities conveyed by tank developed and automatic processor developed images are 1.76 and 1.51 bits per image on the average (the maximum of information quantity can be equally transmitted with 2.32 bits per image). The performance of tank development is found to be superior to that of automatic processor development.

Assessment of Radiographic Granularity by a Single Number. II. Interpretation of Conditional Entropy Hx(y) in the Bivariate Entropy Method

The meaning of conditional entropy Hx(y) in the bivariate entropy method is interpreted theoretically. It is shown that Hx(y) is related to the geometric mean of individual standard deviations of the output distributions when these are Gaussian. This interpretation is used to assess the radiographic granularity by a single number in terms of entropy. It is indicated that Hx(y) in this application is related to the geometric mean of the RMS granularity and represents entropy granularity as defined in this paper. The results of the above analysis and application show that Hx(y) in the bivariate entropy method is a significant measure of the average of the scatter in measured data.

Assessment of Radiographic Granularity by a Single Number

The granularity of radiographic screen-film systems is assessed by a single number in terms of the entropy method. This method is evaluation on the basis of the precision of measurements, and in the present paper the precision is directly related to radiographic granularity which refers to both RMS granularity (the standard deviation of density fluctuations) and film contrast (the gradient of H & D curve). From the results of entropy calculations, the ability of transmission of radiographic images in two screen-film systems is obtained by comparing the values of transmitted information or the relative efficiency of transmission.

Reciprocity-Law Failure in Medical Screen-Film Systems and Its Effects on Patient Exposure and Image Quality

Reciprocity effect or the failure of the reciprocity law is of considerable importance in medical radiography, because most radiographs are made with a pair of intensifying screens in intimate contact with both emulsions of an X-ray film. This failure of the law means that the photographic density depends upon both intensity and exposure time. Recently we have developed an easy and practical method to determine the reciprocity-law failure curve in medical screen-film systems [1–3]. By use of this method, two kinds of reciprocity-law failure curves, expressed as photographic density vs. exposure time for constant exposure and relative exposure vs. exposure time for constant density, can readily be obtained from several time-scale characteristic curves taken experimentally for several focus-film distances(FFDs) [3]. In the present paper, this new method is described with its application to a green-sensitive X-ray film for use in combination with rare-earth screens. The reciprocity effect on the patient exposure will be discussed based on the results of our study.

Noise Smoothing in Image Improvement

We have already reported the correction of a spectrum observed by a monochromator by means of the optical transfer functions (OTF). In these reports, the grain of a film produced a noise which could be contained in the line profile of a corrected spectrum and the line spread function was used to determine OTF. Therefore, in this study the process of smoothing noise by means of the method of least squares was performed, and then the correction of the spectrum was made. We already knew that the signals and the grains make products which represent the intensity of the light which transmits through films. After considering and examining these relationships, we conclude that the correction of a spectral image by means of OTF is difficult to make, unless the process of smoothing is performed on the noise included in signals.

A New Development in the Method of Measurement of Reciprocity-Law Failure and Its Application to Screen/Green-Sensitive X-Ray Film Systems

Since it has been confirmed by experiment that the intensity of X-rays varies approximately as the focus-film distance (FFD) to the minus 2.12th power, the X-ray intensity can be changed by varying the FED. It is shown in this paper that two types of reciprocity failure curve, density vs. exposure time for constant exposure and relative exposure vs. exposure time for constant density, can easily be obtained from several time-scale characteristic curves taken experimentally for several FEDs in the rare-earth screen-film systems used. Only low-intensity reciprocity failure is present for exposure times of more than about 0.1 sec for one film, but both low-intensity and high-intensity reciprocity failures occur in the other one. The effects of reciprocity failure on the H & D curves can be seen in the shape of the curves and the relative speed.

The Relationship between Signals and Graininess on Radiographs

The relationship between signals and graininess on radiographs is discussed here in the following report. It has been reported so far that the signals are reproduced by Fourier inverse transform even though the spectra of signals are cut off. For this reason, it is considered that the intensity of light which is transmitted through films makes the products of the signals and the radiographic mottles. In opposition to these, however, the following could be obtained experimentally. Those products which are made by the signals are the grains of films out of the radiographic mottles, and the brightness of reproduced images is related to the size of grains.

Reliability of the Modulation Transfer Function of Radiographic Screen-Film Systems Measured by the Slit Method

The modulation transfer function (MTF) is one of the most commonly used tools in the analysis of radiographic screen-film systems. It is suggested that the reliability of a systems MTF should be taken into account when the quality of radiographic images is assessed by means of the MTF. The reliabilities of three screen-film system MTFs are obtained by the entropy method (transmitted information) proposed in our previous paper. Numerical results are presented and the reliability rankings of system MTFs are discussed.