A method to measure the MTF of digital x-ray systems

A method has been devised to accurately measure the modulation transfer function (MTF) of digital x-ray systems up to and, for undersampled systems, beyond the pixel Nyquist frequency (fN). A phantom consisting of an array of parallel tungsten or similar wires is imaged, and discrete Fourier transforms of rows of pixel values are computed. Under suitable conditions of phantom orientation, wire diameter, wire spacing, and image magnification, the envelope of the modulus of the mean Fourier transform represents the system MTF. Experimental results extending beyond fN are presented for an undersampled prototype digital chest x-ray system and shown to be in reasonable agreement with predicted values. Employment of the method with other digital imaging modalities [i.e., computerized tomography (CT) scanners and nuclear magnetic resonance (NMR) units] is also discussed as well as error considerations and practical problems in implementing the method.     

Comparison of radiographic texture analysis from computed radiography and bone densitometry systems

Osteoporosis is a disease that results in an increased risk of bone fracture due to a loss of bone mass and deterioration of bone structure. Bone mineral density (BMD) provides a measure of bone mass and is frequently measured by bone densitometry systems to diagnose osteoporosis. In addition, computerized radiographic texture analysis (RTA) is currently being investigated as a measure of bone structure and as an additional diagnostic predictor of osteoporosis. In this study, we assessed the ability of a peripheral bone densitometry (PD) system to yield images useful for RTA. The benefit of such a system is that it measures BMD by dual-energy x-ray absorptiometry and therefore provides high- and low-energy digital radiographic images. The bone densitometry system investigated was the GE/Lunar PIXI, which provides 512 x 512 digital images of the heel or forearm (0.2 mm pixels). We compared texture features of heel images obtained with this PD system to those obtained on a Fuji computed radiography (CR) system (0.1 mm pixels). Fourier and fractal-based texture features of images from 24 subjects who had both CR and BMD exams were calculated, and correlation between the two systems was analyzed. Fourier-based texture features characterize the magnitude, frequency content, and orientation of the trabecular bone pattern. Good correlation was found between the two modalities for the first moment (FMP) with r=0.71 (p value<0.0001) and for minimum FMP with r=0.52 (p value=0.008). Root-mean-square (RMS) did not correlate with r=0.31 (p value>0.05), while the standard deviation of the RMS did correlate with r=0.79 (p value<0.0001). Good correlation was also found between the two modalities for the fractal-based texture features with r=0.79 (p value<0.0001) for the global Minkowski dimension and r=0.63 (p value=0.0007) for the fractal dimension from a box counting method. The PD system therefore may have the potential for yielding heel images suitable for RTA.     

Validation of MTF measurement for digital mammography quality control

The modulation transfer function (MTF) describes the spatial resolution properties of imaging systems. In this work, the accuracy of our implementation of the edge method for calculating the presampled MTF was examined. Synthetic edge images with known MTF were used as gold standards for determining the robustness of the edge method. These images simulated realistic data from clinical digital mammography systems, and contained intrinsic system factors that could affect the MTF accuracy, such as noise, scatter, and flat-field nonuniformities. Our algorithm is not influenced by detector dose variations for MTF accuracy up to 1/2 the sampling frequency. We investigated several methods for noise reduction, including truncating the supersampled line spread function (LSF), windowing the LSF, applying a local exponential fit to the LSF, and applying a monotonic constraint to the supersampled edge spread function. Only the monotonic constraint did not introduce a systematic error; the other methods could result in MTF underestimation. Overall, our edge method consistently computed MTFs which were in good agreement with the true MTF. The edge method was then applied to images from a commercial storage-phosphor based digital mammography system. The calculated MTF was affected by the size (sides of 2.5, 5, or 10 cm) and the composition (lead or tungsten) of the edge device. However, the effects on the MTF were observed only with regard to the low frequency drop (LFD). Scatter nonuniformity was dependent on edge size, and could lead to slight underestimation of LFD. Nevertheless, this negative effect could be minimized by using an edge of 5 cm or larger. An edge composed of lead is susceptible to L-fluorescence, which causes overestimation of the LFD. The results of this work are intended to underline the need for clear guidelines if the MTF is to be given a more crucial role in acceptance tests and routine assessment of digital mammography systems: the MTF algorithm and edge object test tool need to be publicly validated.     

Evaluation of an algorithm for the assessment of the MTF using an edge method

An algorithm to calculate the presampling modulation transfer function (MTF) of an imaging system from an angled edge image has its own inherent transfer function. Factors such as the angle of the sampling aperture to the edge, registration of edge function profiles using the determined edge angle, differentiation, smoothing, and folding all combine to produce the frequency response of the algorithm. In this work, the profile registration transfer function accounting for an error in the determined edge angle has been derived. This has been incorporated with other, previously reported, algorithm component transfer functions to fully characterize the MTF calculation algorithm. When registering profiles, small errors in the edge angle determination were found to result in large errors in the MTF, as the misalignment errors increase with the number of profiles. For example, registering 50 profiles a 0.07 degree error in a 7 degree edge angle (1% error) produces a 36% error in the MTF at the system cutoff frequency f=f(c) when profiles are oversampled at a frequency f(s)=8f(c)(f(c) is defined as the maximum frequency reproducible without aliasing when sampling at the limiting system Nyquist frequency f(s) = 2f(c)). These results highlight the importance of quantifying the transfer function of the algorithm used to determine an imaging system modulation transfer function. The MTF calculation algorithm and the transfer function analysis have been incorporated into a Windows-based software program to be made available for general use.     

Comparison of three different techniques for dual-energy subtraction imaging in digital radiography: A signal-to-noise analysis

Dual-energy subtraction imaging techniques allow the tissue and bone structures in the patient to be visualized and studied in two separate images, thus removing the obscurity associated with overlapping of the two structures. In addition, they allow the subtraction image signals to be used for quantifying the tissue and bone thicknesses. Thus, capability for dual-energy subtraction imaging is often incorporated with new digital radiography systems. There are three different approaches to dual-energy image subtraction imaging techniques. Among them, the dual-kilovolt (peak) [kV(p)] and sandwich detector techniques have been two widely used approaches. A third approach is the single-kV(p) dual-filter technique, which allows some flexible control of the spectra while avoiding the technical complexity of kV(p) value switching in slit-scan imaging. In this report, the noise properties associated with these three techniques are studied and compared by computing the noise variances in the subtraction image signals as a function of the kV(p) values and filter thicknesses. It was found that the dual-kVp technique results in the least noisy subtraction images, whereas the dual-filter technique results in slightly less noisy subtraction images than the sandwich detector technique. Following optimization of the kV(p) value and filter thicknesses, the dual-filter and sandwich detector techniques result in a noise level of approximately three and four times higher than that resulted from the dual-kV(p) technique, respectively.     

Accuracy of a simple method for deriving the presampled modulation transfer function of a digital radiographic system from an edge image

Several methods for accurately deriving the presampled modulation transfer function (MTF) of a pixelated detector from the image of a slightly slanted edge have been described in the literature. In this paper we report on a simple variant of the edge method that produces sufficiently accurate MTF values for frequencies up to the Nyquist frequency limit of the detector with little effort in edge alignment and computation. The oversampled ESF is constructed in a very simple manner by rearranging the pixel data of N consecutive lines corresponding to a lateral shift of the edge by one pixel. A regular subsampling pitch is assumed for the oversampled ESF, which is given by the original pixel sampling distance divided by the integer number N. This allows the original data to be used for further computational analysis (differentiation and Fourier transform) without data preprocessing. Since the number of lines leading to an edge shift by one pixel generally is a fractional number rather than an integer, a systematic error may be introduced into the presampled MTF. Simulations and theoretical investigations show that this error is proportional to 1/N and increases with spatial frequency. For all frequencies up to the Nyquist limit, the relative error delta MTF/MTF is smaller than 1/(2N). It can thus be kept below a given threshold by suitably selecting N, which furnishes a certain maximum edge angle. The method is especially useful for applications where the presampled MTF is needed only for frequencies up to the Nyquist frequency limit, such as the determination of the detective quantum efficiency (DQE).     

Modified inverse square sensitometry for the determination of the characteristic curve of radiographic screen/film systems

To determine the characteristic curve of the radiographic screen/film systems in a short focal spot-film distance, the inverse square sensitometric method was modified by changing the radiation intensity with two kinds of filters. The characteristic curves obtained in the two exposure series with these two kinds of filters were overlapped to obtain a complete one. The characteristic curve thus obtained was almost the same as the one obtained by the original inverse square sensitometric method. The accuracy of the characteristic curves obtained by the modified method was well-reflected in the clinical radiographs.     

Optimized radiographic spectra for small animal digital subtraction angiography

The increasing use of small animals in basic research has spurred interest in new imaging methodologies. Digital subtraction angiography (DSA) offers a particularly appealing approach to functional imaging in the small animal. This study examines the optimal x-ray, molybdenum (Mo) or tungsten (W) target sources, and technique to produce the highest quality small animal functional subtraction angiograms in terms of contrast and signal-difference-to-noise ratio squared (SdNR2). Two limiting conditions were considered-normalization with respect to dose and normalization against tube loading. Image contrast and SdNR2 were simulated using an established x-ray model. DSA images of live rats were taken at two representative tube potentials for the W and Mo sources. Results show that for small animal DSA, the Mo source provides better contrast. However, with digital detectors, SdNR2 is the more relevant figure of merit. The W source operated at kVps >60 achieved a higher SdNR2. The highest SdNR2 was obtained at voltages above 90 kVp. However, operation at the higher potential results in significantly greater dose and tube load and reduced contrast quantization. A reasonable tradeoff can be achieved at tube potentials at the beginning of the performance plateau, around 70 kVp, where the relative gain in SdNR2 is the greatest.     

Comparison of radio-immunoassay and lymphocytotoxicity inhibition techniques for the determination of beta2 microglobulin.

Rabbit anti-human beta2 microglobulin antisera can lyse human lymphocytes in the presence of rabbit complement. Inhibition of the lymphocytotoxic reaction by highly purified beta2m was applied to the measurement of beta2m concentration in biological fluids. Parallel determinations were also performed using a radioimmunoassay. Lymphocytotoxicity inhibition is simple and more sensitive than radial immunodiffusion but less sensitive than the radioimmunoassay. beta2m was measured by these two techniques in serum and urine from normal individuals, uremic or transplanted patients.     

A method for modifying the image quality parameters of digital radiographic images

A new computer simulation approach is presented that is capable of modeling several varieties of digital radiographic systems by their image quality characteristics. In this approach, the resolution and noise characteristics of ideal supersampled input images are modified according to input modulation transfer functions (MTFs) and noise power spectra (NPS). The modification process is separated into two routines-one for modification of the resolution and another for modification of the noise characteristics of the input image. The resolution modification routine blurs the input image by applying a frequency filter described by the input MTF. The resulting blurred image is then reduced to its final size to account for the sampling process of the digital system. The noise modification routine creates colored noise by filtering the frequency components of a white noise spectrum according to the input noise power. This noise is then applied to the image by a moving region of interest to account for variations in noise due to differences in attenuation. In order to evaluate the efficacy of the modification routines, additional routines were developed to assess the resolution and noise of digital images. The MTFs measured from the output images of the resolution modification routine were within 3% of the input MTF The NPS measured from the output images of the noise modification routine were within 2% of the input NPS. The findings indicate that the developed modification routines provide a good means of simulating the resolution and noise characteristics of digital radiographic systems for optimization or processing purposes.     

Image quality in digital chromosome analysis systems

This paper reports on an investigation into the differences in image quality of different components used in a digital image processing system for chromosome analysis. As chromosome aberrations are important tools in the cloning of genes, it is important to know if the introduction of computerized analysis systems increases the risk of missing small aberrations. In this investigation the number of visible bands on a number of chromosomes has been used as a measure of quality. The images compared are microscope ocular images, photographs from a microscope builtin camera, digital images from a high and from a standard resolution camera, presented both on screen and print-out on paper. The main conclusions are that: (1) the view in the microscope ocular gives the best resolution, (2) there are risks of losing vital information using the digital image processing system for chromosome analysis, and (3) this risk is significantly reduced when using a high resolution camera.     

Scatter estimation for a digital radiographic system using convolution filtering

The use of a convolution-filtering method to estimate the scatter distribution in images acquired with a digital subtraction angiography (DSA) imaging system has been studied. Investigation of more than 175 convolution kernels applied to images of anthropomorphic head, chest, and pelvic phantoms using 15-, 25-, and 36-cm fields of view (digitized onto a 512 X 512 pixel image matrix) showed that two-dimensional exponential kernels with a full width at half maximum (FWHM) of 50-150 pixels best reproduced the scatter fields within these images with a root-mean-square percentage error from 4% to 8%. A two-dimensional exponential kernal with a FWHM of 75 pixels in each dimension applied to ten different anatomic presentations and fields of view, resulted in an average root-mean-square percentage error of 6.6% for the ten cases studied. The method should be implementable using an array of small lead beam stops placed in the field of only a single mask image and the above described convolution kernel applied to both mask and postopacification images. The mask beam-stop data are used to scale both mask and postopacification convolution-filtered images. This scaled, convolution-filtered image is then subtracted from the original image to produce a largely scatter-corrected image.     

Comparison of several test systems used for determination of rubella immune status.

Hemagglutination inhibition (HAI) is currently the most widely used technique for the determination of rubella immune status. However, two new methods, enzyme-linked immunosorbent assay (ELISA) and indirect immunofluorescence (FIAX), have also been adapted for this purpose. In comparing a commercially available ELISA system (BIO-BEAD, Litton Bionetics) with an HAI system (RUBA-tect, Abbott Laboratories), some ELISA-positive sera were found to be rubella antibody negative by the HAI system. To determine which of these results more accurately reflected the immune status of the patient, 74 RUBA-tect-negative sera were retested by ELISA BIO-BEAD, FIAX (International Diagnostic Technology) and by modified HAI, employing fresh erythrocytes (using Flow Laboratories reagents). Eleven RUBA-tect-negative sera (15%) were positive by ELISA, FIAX, and modified HAI. Two sera were positive only by ELISA and FIAX, two sera were positive by ELISA and HAI, four sera were positive by ELISA only, and one serum was positive by FIAX only. Neutralization assays were subsequently performed on sera positive by only one or two of the procedures to determine the presence of protective rubella antibodies in these sera; all but three of the sera were positive for neutralizing antibody. Commercial ELISA and FIAX systems appear to be more sensitive indicators of rubella immune status than are commercial HAI kits which use stabilized erythrocytes. Neither ELISA nor FIAX require extraction of serum; moreover, the ELISA BIO-BEAD test assay can be performed without an expensive instrument for reading.     

On modelling the kinestatic charge detector for digital radiographic diagnostic and portal imaging

The kinestatic charge detection (KCD) principle has been a digital radiography technique for more than a decade. The advances of the KCD technique have gone from diagnostic imaging to portal imaging. However, little work has been done on understanding the selection of key KCD parameters and relationships between them. In the present study, an engineering model was established that could be used to optimize the placements of key parameters in terms of KCD system mechanical design. In the proposed KCD engineering model, the basic energy conservation law was applied to the process of ion transmission. It allows for the computation of the KCD design parameters such as the optimum grid placement, high voltage board tilt angle and grid wire space, as well as to provide recommendations on high voltage board and electric potentials and their ratio.