Previously in this forum we have reported the application of multiparameter optimization techniques to the design of a minimum dose mammography system. The approach used a reference system to define the physical imaging performance required and the dose to which the dose for the optimized system should be compared. During the course of implementing the resulting design in hardware suitable for laboratory testing, the state of the art in mammographic imaging changed, so that the original reference system, which did not have a grid, was no longer appropriate. A reference system with a grid was selected in response to this change, and at the same time the optimization procedure was modified, to make it more general and to facilitate study of the optimized design under a variety of conditions. We report the changes in the procedure, and the results obtained using the revised procedure and the up- to-date reference system. Our results, which are supported by laboratory measurements, indicate that the optimized design can image small objects as well as the reference system using only about 30% of the dose required by the reference system. Hardware meeting the specification produced by the optimization procedure and suitable for clinical use is currently under evaluation in the Diagnostic Radiology Department at the Clinical Center, NH.

The development of specialized dyes that essentially prevent light from crossing the film base in double-coated gadolinium oxysulfide (GOS) phosphor-based radiographic systems has made it possible to design screen-film combinations with significantly improved MTF characteristics. Specifically, by using GOS-based screens with reduced light diffusion properties in combination with near-zero-crossover radiographic films, significantly improved MTF can be obtained at competitive speed and effective x-ray attenuation levels. The basic performance characteristics of such screen-film systems are described in some detail, including x-ray attenuation properties, sensitivity to scattered x-radiation, sensitometric data, contrast transfer functions, noise equivalent quanta, and detective quantum efficiency. It is also shown that high-MTF GOS screens are available that meet or exceed the performance characteristics of comparable UV-emitting yttrium tantalate phosphor-based materials.

A 600-speed system based on Ultra-Vision^TM technology has been designed. Since the undoped yttrium tantalate phosphore used in other Ultra-Vision^TM screens is incapable of achieving the required speed with existing films it was necessary to admix a small fraction of lanthanum oxybromide which is a more efficient x-ray to light converter and which emits both visible (approximately 470 nm) and UV (approximately 370 nm) light. The image quality performance of the resulting system approaches that of conventional 400-speed film/screen systems. These image quality characteristics are described and compared with those of conventional 600-speed and 400-speed systems.

We obtained the tube voltage dependence of the Wiener spectra of the radiographic mottles of the front and back emulsions made using a low-speed screen and an anticrossover film. The Wiener spectral values of the front emulsion are almost constant, while those of the back emulsion increase with the tube voltage. The spectral values of the front emulsion are greater than those of the back emulsion at most of densities. In order to investigate the reason of these phenomena, we separated the radiographic mottle into the three factors. As a result, we have found that the reason of the dependence of the radiographic mottle is that the contribution of the quantum mottle is less than that of the structure mottle. The Wiener spectral values of the structure mottle of the front screen are greater than those of the back screen for most of the tube voltages. The Wiener spectral values of the front screen decrease and those of the back screen increase with the tube voltage, and the two curves cross at high voltage. These reasons were explained from the spatial fluctuation of the thickness of the screen due to the nonuniform screen structure and the attenuation curve of the x-ray intensity. The tube voltage dependence of the quantum mottle was explained from that of the number of photons absorbed in the screen.

In the present contribution we have investigated the nature and role of the probability distribution function governing the fluctuations in image density about a mean level, and especially the relationship of this distribution to the Wiener spectrum associated with image noise. Experimental measurements are shown confirming that, at a given mean density level, the probability distribution function is independent of exposure statistics regardless of whether the film is exposed to uncorrelated light photons or to correlated light via an intensifying screen. We therefore conclude that differences in the associated Wiener spectrum are entirely attributable to changes in the extent of the autocorrelation interval and not due to changes in the shape or extent of the probability distribution function.

The resolution of a fiberoptical boule was determined experimentally. A boule is a matrix of scintillating glass fibers arranged parallel to each other and compressed into a thick plate. A 51 mm X 51 mm by 3 mm thick terbium-activated boule was optically coupled to a 1 k X 1 k X 16 bit CCD camera. The image of an edge was used to compute the line spread function, and the modulation transfer function was calculated. After correcting for the MTF of the CCD and optics, the MTF of the fiberoptical plate demonstrated 20% modulation at a spatial resolution of 15 line pairs per millimeter. The x-ray absorption of the 3 mm thick plate was 98% at 70 kV (3.3. mm half value layer). It is shown using simple trigonometry that for a very high resolution detector, x-ray beam divergence (angled photons striking the receptor) at the periphery of the field of view may cause substantial resolution losses. The role that fiberoptical plate technology may play in diagnostic medical imaging is discussed.

Reverse Geometry X-radiography imaging uses a unique configuration of x-ray source, patient, and x-ray detector to produce high resolution images. With the current system, produced by Digiray, one can obtain a resolution of 16 lp/mm and a contrast sensitivity of 0.2%. Using a lead focusing grid with 100 keV x-rays at 1 mA a dose of less than 20 mR/sec was measured. Typical acquisition times are 1/16 sec for 256 line resolution to 8 sec for 2048 line resolution. An optimized multi-detector system under development should reduce the dose rate to far less than 1 mR/sec and allow the possibility for fast volume CT applications providing laminographic view of the patient. Also `mini-probe' detectors have been developed which may be used to produce in-vitro unobstructured views.

A hybrid cassette has been developed for simultaneous acquisition of storage phosphor and asymmetric screen-film chest images. This is important for the collection of images for Receiver Operating Characteristic studies comparing conventional radiography and computed radiography, without either increased exposure or non-identical imaging conditions. This hybrid radiographic cassette consists of a computed radiography imaging plate (in front) and an intact, high contrast variant of a commercially available asymmetric screen-film system (in the rear) with a speed of approximately 425. The high contrast, speed and efficiency of this screen-film system allow for positioning of the storage phosphor plate in the front of the cassette. As the imaging plate absorption is approximately 35%, the fast screen-film system provides high quality diagnostic images. There is minimal beam hardening, which is ameliorated by the high contrast of the asymmetric front screen. There is minimal differences in the Plexiglas step wedge phantom gray level values for CR and CR-hybrid images and in optical density values for InSight^TM and InSight^TM-hybrid films. The signal to noise ratio of either hybrid image, while fractionally less than their standard counterparts, is negligibly so. Only a slight modification in radiographic technique is required (10%) for use of this hybrid cassette, providing images that are virtually the same as those obtained through the standard CR and InSight^TM ITC imaging methods.

The image quality of a special high-resolution computed radiography (CR) imaging plate is compared with that of conventional screen-film radiography via measurements of physical factors related to the signal and noise of each detector system. Physical parameters reflecting image contrast, spatial resolution and signal-to-noise characteristics are measured as a function of radiation dose for each detector configuration. Standard signal detector configurations for the acquisition of plain film and CR images are compared with those from a modified film cassette containing both a CR plate and conventional screen-film for a dual-image recording technique. The modified dual detector film cassette configuration allows for a single-exposure, simultaneous acquisition of images for the direct clinical comparison of conventional screen- film and CR radiographs. The small physical size of the newborn chest and the types of lung processes that must be evaluated in the critical care neonate present a particular challenge to the spatial resolution limitations of CR. The clinical utility of the special high-resolution CR plate is assessed for such neonatal intensive care unit applications.

FCR system has an automatic adjustment of image density and contrast by analyzing the histogram of image data in the radiation field. Advanced image recognition methods proposed in this paper can improve the automatic adjustment performance, in which neural network technology is used. There are two methods. Both methods are basically used 3-layer neural network with back propagation. The image data are directly input to the input-layer in one method and the histogram data is input in the other method. The former is effective to the imaging menu such as shoulder joint in which the position of interest region occupied on the histogram changes by difference of positioning and the latter is effective to the imaging menu such as chest-pediatrics in which the histogram shape changes by difference of positioning. We experimentally confirm the validity of these methods (about the automatic adjustment performance) as compared with the conventional histogram analysis methods.

Preliminary results are presented from an examination of a selenium based thoracic radiography system (Philips Thoravision). The system has been evaluated with four criteria: (1) systems repeatability, (2) digital linearity, (3) spatial resolution, and (4) scattered photon detection fractions. Phantoms were imaged at 120 kV. Repeatability was measured using a polystyrene phantom and evaluating the mean exposure value detected in a region of interest over a period of time. Digital linearity was examined by plotting the output digital value as a function of input exposure. Resolution was evaluated using a line pair phantom. To measure scatter fractions, an anthropomorphic phantom was exposed with a superimposed array of lead beam stops. Three configurations were examined: (1) the selenium system (including an air gap), (2) the system with an added 12:1 antiscatter grid, and (3) a photostimulable phosphor system (Philips PCR) for reference. For a 4 day interval, output varied less than 1%. Digital output of the system was linear with exposure (regression Rvalue of 0.998) over the range from 0.2 mR to 10 mR. The system resolved 2.5 line pair per mm. Resolution was comparable to phosphor plate systems. Scatter fractions were improved when a grid was included.

Progress toward the development of a large area, flat-panel imager for diagnostic x-ray imaging is described. The initial fabrication of a prototype array with a format of 1536 X 1920 pixels and a pixel pitch of 127 micrometers giving an active area of 19.5 X 24.4 cm² is reported. With a total of approximately 2.9 million transistors, this ambitious array is on a par with modern microprocessors in terms of transistor count. This work builds upon our concurrent research into the development of a very large area, lower spatial resolution, flat-panel imager for radiotherapy. An overview of the anticipated imaging properties of these devices is presented and future prospects discussed.

The dream of an all-solid state large area x-ray image sensor with digital readout and full dynamic performance will most probably find a first realization in 2D thin-film amorphous silicon arrays. In this paper we address in particular the evaluation of the limits of the signal/noise ratio in this concept. Using small prototype detectors measurements of MTF and noise power spectra have been made as a function of x-ray dose. The results are given in terms of the detective quantum efficiency as a function of dose and spatial frequency. We further present an analysis of the different noise sources and their dependence on the detector parameters, and we provide estimates on the maximum signals that may be achieved per unit dose. The intrinsic lag of the amorphous silicon photodiodes causes a second problem area with this type of x-ray detectors. Especially in radiography/fluoroscopy mixed applications, memory effects may not be negligible.

Hydrogenated amorphous silicon imaging arrays are being developed for numerous applications in medical imaging. Diagnostic and megavoltage images have previously been reported and a number of the intrinsic properties of the arrays have been investigated. This paper reports on the first attempt to characterize the intrinsic spatial resolution of the imaging pixels on a 450 micrometers pitch, n-i-p imaging array fabricated at Xerox P.A.R.C. The pre- sampled modulation transfer function was measured by scanning a approximately 25 micrometers wide slit of visible wavelength light across a pixel in both the DATA and FET directions. The results show that the response of the pixel in these orthogonal directions is well described by a simple model that accounts for asymmetries in the pixel response due to geometric aspects of the pixel design.

The design and construction of a prototype high resolution digital x-ray imaging system, employing a novel scintillating microfiber detector as the x-ray imaging sensor is described. The fiber detector has an active imaging area of 5 cm X 5 cm and thickness of 1 cm. The optical image formed in the fiber detector is read out by a high resolution image intensifier- CCD camera system. Investigations on the image quality of the prototype x-ray system are reported. MTF measurement shows >= 10 line pairs/mm spatial resolution of the scintillating fiber detector over the diagnostic x-ray energy range. The feasibility of the medical diagnostic application of the scintillating fiber detector is analyzed, and the new sensor is shown to be adaptable to imaging tasks where both high resolution and high detector sensitivity are required.

The authors developed and clinically applied digital radiographic facilities suitable for general examinations based on x-ray image intensifier/video systems. Digital imaging permits instantaneous production of images on a monitor which can then be visually evaluated and processed to optimize image quality prior to selection for permanent hard-copy. We think of this system as a multi-modality digital 100 mm camera to replace the universally accepted 100 mm or 105 mm photofluorographic images. The requirements for a digital 100 mm camera are a digital system which can operate effectively at 100 mm or less exposure levels and 100 mm exposure times so as to consistently provide low noise and stop all motion. It has previously been shown that 1024² digital matrices provide adequate resolution for 9 - 10' fields of view and permit the same quality as is used for the gold standard 100 mm photofluorography. Here we investigate whether large field sizes (16') can benefit by increasing the matrix to 2048².

While methods of testing fluroscopic systems are well established, using test objects such as the Leeds test objects, observer influence may affect the results and the process is relatively time consuming. An alternative approach is examined, whereby use of a line selector and oscilloscope are used with a test object to examine resolution, contrast and grey scale steps. Three radiotherapy simulator fluoroscopic imaging systems are tested and, in each case, observed images are found to be limited by the signal response, as present in the video signal, and not by monitor performance. Edge enhancement effects and signal droop are also found to be effects which merit further investigation. It appears that, at low cost, the use of an oscilloscope and a TV line selector may permit the development of quantitative methods for assessing the performance of imaging systems.

In order to quantitate anatomical and physiological parameters such as vessel dimensions and volumetric blood flow, it is necessary to make corrections for scatter and veiling glare (SVG), which are the major sources of nonlinearities in videodensitometric digital subtraction angiography (DSA). A convolution filtering technique has been investigated to estimate SVG distribution in DSA images without the need to sample the SVG for each patient. This technique utilizes exposure parameters and image gray levels to estimate SVG intensity by predicting the total thickness for every pixel in the image. At this point, corrections were also made for variation of SVG fraction with beam energy and field size. To test its ability to estimate SVG intensity, the correction technique was applied to images of a Lucite step phantom, anthropomorphic chest phantom, head phantom, and animal models at different thicknesses, projections, and beam energies. The root-mean-square (rms) percentage error of these estimates were obtained by comparison with direct SVG measurements made behind a lead strip. The average rms percentage errors in the SVG estimate for the 25 phantom studies and for the 17 animal studies were 6.22% and 7.96%, respectively. These results indicate that the SVG intensity can be estimated for a wide range of thicknesses, projections, and beam energies.

Scatter to primary radiation ratios at detector positions were calculated for spherical water phantoms in the cone-beam x-ray 3D CT scanner using the Monte Carlo simulation method. In the Monte Carlo simulation, 1 X 10⁶ photons of cone-beamed 120-kV polyenergetic x-rays with the filter of 4 mm Al have been traced individually as these photons interact within the phantoms with diameters 15 cm and 20 cm. The decrease in CT number due to scatter was found to be about 100. In the process of image reconstruction, the effect of scatter was isolated by applying our new calculation technique. The scatter reduces CT numbers in the whole region of images and the intensity of scattered rays is nearly uniform, so that rounded- off image is reconstructed.

X-ray microcomputed tomograms with resolution of 50 microns or better have been obtained at several research centers by using either synchrotron or microfocus x-ray sources. Full 3D reconstructions have been obtained in these laboratories on specimens or small animals. In our laboratory we have studied embedded bone specimens by using x-ray cone beam microtomography-methods. For conventional, medical x-ray computed tomography systems, well accepted techniques and standards exist for characterizing the performance of an instrument. No comparable standards exist for computed microtomography and previous publications are vague with respect to the performance achieved. We report in this paper the experimental methods that we have developed to measure noise and resolution in our microtomography laboratory and the specific performance characteristics we have achieved.

An image intensifier-based rotational volume tomographic angiography imaging system has been constructed. The system consists of an x-ray tube and an image intensifier that are separately mounted on a gantry. This system uses an image intensifier coupled to a TV camera as a 2D detector so that a set of 2D projections can be acquired for a direct 3D reconstruction. Although an image intensifier offers good detection quantum efficiency and possibly results in a better low contrast resolution than a fluorescent screen, it introduces two types of distortion: S distortion and pincushion distortion. To obtain accurate reconstructions, these distortions must be corrected prior to 3D reconstruction. Techniques for the correction of these distortions have been developed. These techniques were tested using experimental data acquired with the image intensifier-based volume tomographic angiography imaging system. The results indicate that the distortion correction techniques work well.

In the reconstruction of positron emission tomography images, each slice of the image volume is individually reconstructed from a sinogram, in which the statistics of the data elements are Poisson and the image data is hidden by the mechanism of projection. We propose a method of image reconstruction which incorporates the given data set and also reflects the a prior knowledge that the image consists of smooth noiseless regions that are separated by sharp edges. This method uses both maximum likelihood and maximum a posteriori techniques in a manner that is similar to techniques used by others, but our method incorporates a bounded prior term and adaptive annealing techniques. These advancements prevent excessive smoothing and address the difficulties presented by parameter selection and image convergence.

It is discovered that for circular scanning orbits a function to be reconstructed, f (, can be broken into three terms: f () = fMo(f) + fMl(f) +fN(i)' where f,,(i) represents the null space component of f (i), fçi')corresponds to the reconstruction proposed by Feldkamp etal, while fM1() representsthe remaining part ofthe measurement space component, denoted as f, that is not provided in the Feldkamp reconstruction. Thus, a new cone beam reconstruction algorithm for circular scans is proposed as follows: 1) compute f) using Feldkamp algorithm; 2) compute fMtfr) using the formula developed inthis paper; and 3) estimate f4). This new algorithm has the following merits: 1) By including the correction term fMl(?),itprovides a more accurate reconstruction of f,fl) than Feldkamp algorithm; 2) It computes f(r) and fMl?) in a filtered—backprojection fashion, assuring an accurate and computationally efficient reconstruction of f4(?). For this reason, we expect that it would compare favorably in practice to Grangeat algorithm. This mathematical framework also provides a new perspective to understand the analytical relation between Grangeat algorithm and Feldkamp algorithm.

A method to construct anthropomorphic vascular flow phantoms which incorporate disease states (e.g., stenosis) has been developed. The new technique is an adaptation of an existing technique that fabricates disease free vascular phantoms. Using the modified process, three carotid artery bifurcation phantoms have been built. One phantom was free of stenotic disease, and the other two phantoms have geometries corresponding to (1) a 70% concentric stenosis, and (2) a 70% eccentric stenosis in the internal carotid artery. Collectively, these three phantoms allow for comparative research between and in vitro evaluation of a number of medical imaging instruments, including (1) x-ray based systems (e.g., digital subtraction angiography and computed tomography), (2) Magnetic Resonance Imaging (MRI) scanners, and (3) Doppler Ultrasound (US) instruments. Here, representative images from x-ray and MR imaging instruments are shown. Pulsed Doppler US velocity spectra collected within the lumen of the phantoms are also presented.

We propose a mathematical model for analyzing grid line digitization artifacts on high resolution film. In this paper, aliasing frequencies and the amplitudes of the higher frequency components of grid lines have been derived based on a theoretical model given in an earlier paper. Then the experimental results obtained from digitizing a mammography film are compared with the theoretical results. We propose a method of choosing the optimum digitization parameters in order to minimize grid line aliasing artifacts.

A new technique to digitize X-ray radiographs was investigated. This technique is based on the principle of light photon equalization (LPE). In this method, the light source is modulated so that the dynamic range of the transmitted light into a photo-detector is precisely controlled to obtain improved signal-to-noise ratio. The modulation of the light source is derived from a smoothed image of the original. The final output digital image is reconstructed by multiplying the modulation function and the resultant LPE image. The LPE was implemented in a laser film digitizer as a prototype. The results show that, although the dynamic range of X-ray films is about 10⁴:1, the range of the transmission light into a photo detector was reduced to 10²:1 with this technique. Also the SNR improved by maximum factor of five in the lung area of a digitized chest radiograph.

The development of CCDs for use in dental radiology is now well established and provides a base technology for the development of mammography screening detectors. Techniques for overcoming the critical issues of image quality and detection area are discussed and include a 20 line pair/mm resolution capability.

Medical x-ray imaging systems must be carefully designed to ensure that images can be produced with the highest possible signal-to-noise ratio (SNR) for a given x-ray dose to the patient. In most practical systems, images result from the conversion of primary x-rays into secondary quanta such as optical photons or electrical charge, in multiple cascaded stages. The average number of quanta at each stage (per incident x ray) is often evaluated as the product of all preceding system gain factors and displayed graphically as a "quantum-accounting diagram" (QAD). The stage with the fewest quanta forms the quantum sink, and is generally the noise-determining stage. This "zero spatial-frequency" type analysis is simplistic, however, as it ignores the spatial spreading of secondary quanta that will further degrade image noise. We have recently extended the above approach by introducing a spatial-frequency dependent QAD, in which the number of quanta at each stage is expressed by the product of all preceding system gains and squared modulation transfer functions (MTFs). The results are displayed graphically and used to determine the quantum-sink stage as a function of spatial frequency. The visual impact of the non-zero spatial frequency quantum sink is illustrated in a Monte Carlo simulation of the cascading process. A hypothetical system consisting of a scintillating phosphor optically coupled to a CCD camera is used for illustrative purposes. It is shown that an inefficient optical system results in the addition of a quantum mottle to the images due to a secondary quantum sink in the number of optical quanta. The mottle appears to be uniform in frequencies, but in combination with the effect of the screen MTF masks high-frequency detail more than low-frequency detail. This secondary quantum sink can be minimized both by: i) increasing the efficiency of the optical system; and, ii) improving the highfrequency response of the screen. Increasing the optical efficiency reduces the secondary quantum mottle, thus improving visualization particularly at high frequencies. Improving the high-frequency response makes a slight improvement on the quantum mottle, and also increases the contrast of the high-frequency patterns. The combination improves visualization at high spatial frequencies. Interpretation of the QAD is assisted by a direct comparison with the corresponding Monte Carlo images. It is concluded that the secondary quantum sink results in a visible deterioration of image quality at any specified frequency when the QAD at that frequency is less than approximately five times the primary quantum sink QAD value.

A commercial prototype electronic intraoral dental x-ray imaging system employing a direct sensing CCD array has been developed. Image quality parameters were measured using x-ray sources at the National Institute of Standard and Technology radiation physical department in Gaithersburg, MD. Detector response to x-rays in the 10 to 70 keV energy range was measured. The beam hardening effects of human anatomy on a typical 70 kVp spectra was measured using a tissue-equivalent dental phantom.

This paper describes the requirements and constraints involved in the design of a digital scanning mammography system. The SNR model of Muntz is used to assess the influence of detector size and detected scatter-to-primary ratio on the detectability of microcalcifications in a breast phantom, and hence to consider the design of an optimized system. The effects of patient dose, exposure time and various technological constraints, such as x-ray tube power limits, maximum scanning speed and system MTF, are also discussed. It is demonstrated that the combined use of an air gap between the patient and the detector and a wider scanning slot can give the same scatter-to-primary ratio as a significantly smaller slot, while reducing the x- ray tube load. It is also demonstrated that, to some extent, the tube power constraints can be overcome by using higher kVp, less added filtration and a smaller source to patient distance.

A large-area electronic x-ray imaging cassette is proposed that has the potential to replace screened film and real-time imaging modalities in diagnostic radiology applications.

This paper presents results of experiments performed to find parameters affecting signal, noise and Detective Quantum Efficiency of x-ray imaging systems with potential for use in mammography.

In 1990 a new high resolution Plumbicon camera tube development was presented at SPIE Medical Imaging IV. At that time, image processors were not readily available to take advantage of the performance capabilities of this imaging device. This paper presents the result obtained with the latest imaging system capable of utilizing the performance advantages of high resolution camera tubes.

We have investigated the number of bits necessary for digitizing fluoroscopic images. An A/D with sufficient bits must be used to prevent degradation of the Signal-to-Noise Ratio (SNR) by quantization noise. The maximum achievable SNR with a 8-bit A/D is 887 assuming that the input signal probability density distribution is constant within a quantization bin width. In special cases of low light level video imaging where images are quasi-stationary (for example in electronic portal imaging), one can increase the inherent image SNR either by integrating the image on the camera target for extended periods or by averaging a number of real time (30 frames per second) images into a frame buffer. An A/D with higher number of bits will be required in the first case since the SNR of the analog signal is increased due to target integration. The same SNR can be achieved with a relatively low resolution A/D by averaging an equivalent number of frames if one neglects readout noise. We have used theoretical reasoning and experimental measurement with real time portal imaging to show that one can achieve an SNR higher than 887 with an 8 bit A/D with frame averaging if the input analog signal has certain amount of random noise.

Monte Carlo methods play an important role in medical imaging research. Direct analog Monte Carlo simulations can be very accurate but require considerable computational resources. Variance reduction techniques may offer a solution to this problem. In this paper we present a comparison of expected values of standard quantities of interest for SPECT using these two simulation methods. The effect of variance reduction on the statistical characteristics of the simulated data is also investigated.

A medical imaging camera with an avalanche-type image pickup tube (HARPICON) was developed for digital radiography. The camera obtains high quality images even at low doses because its sensitivity is as much as 32 times higher than that of a conventional pickup tube camera. The camera also has a wide dynamic range and a high signal-to-noise ratio because of HARPICON's gamma characteristic.

X-ray imaging devices for dental applications must have an image area of at least 3 X 4 cm². MOS CCDs cover an area of about only 1.5 X 2.5 cm², in this case the x-ray image should be reduced to these smaller dimensions. This can be realized by tapered optical fibers after the x-radiation has been converted into light. This procedure results in a rather bulky construction. Junction CCDs can be realized in the desired image area of 3 X 4 cm², thus making the taper construction unnecessary. Experiments on a linear array Junction CCD covered with a phosphorescent layer have been carried out. These experiments show, that only 3.2 mR for one exposure is needed. This results in a very low radiation load for the patient. The radiation load for the sensor itself is very low too, which makes a high number of exposure possible during the lifetime of the device.

A method for spatial-velocity imaging of nonrigid moving targets, such as the human heart, which utilizes a quasi-monostatic phased array or a bistatic phased array that illuminates the target scene with a continuous-wave large-bandwidth signal and simultaneously records the Doppler-shifted echoed signals is presented. The imaging system utilizes an inverse solution for phased array imaging that incorporates the spectral distribution of an element's radiation pattern and it does not put any restrictions on the irradiating signal and the target scene. The system is applicable in near field as well as far field imaging problems of diagnostic medicine involving the human cardiovascular system.

This paper presents a theoretical framework of a new time-domain method for estimating the slope of ultrasonic attenuation ((beta) ) in reflection. The proposed method has two main advantages: high accuracy and technical simplicity.

We propose a new MRI acquisition method where data are collected from multiple directions in k-space. Its principal advantages are the reduction in acquisition time and the improvement of the signal to noise ratio. The potential disadvantage of the method is a loss in resolution. We have implemented the method and compared it with conventional truncated Fourier imaging using simulated objects, physical phantoms and human brains. The results show noise reduction ranging 25 - 40%. For brain images the loss of resolution is difficult to detect. The use of the method is recommended in applications where the signal to noise is poor, such as image subtraction in functional imaging.

The authors attempted laser therapies for squamous cell carcinoma combined with the chemotherapy in bleomycin (BLEO) intending to reinforce the therapeutic effect. Based on the result of the spectral analysis on the target tissues, argon ion laser (Ark laser) was selected for this photocheinotherapy. Following the intravenous administration of BLEO, the target tissue was exposed to the laser beam spot at low power density, and the laser treatment was repeated on all over the affected site including the surrounding normal part. The photochemotherapy regimen was repeated at every therapeutic session with twice the interval of administration of BLEO as usual. In spite of the reduction of the usual dosage of BLEO by half, remarkable therapeutic effect was assured at an early stage in the photochemotherapy session, and the tumor was reduced with increase in the therapeutic sessions. In addition to the successful therapeutic result, the following interesting phenomenon was noticed in the photochemotherapy regimen. When the target tissue was observed during exposure to the laser beam with reduction of the hyperbrightness of the reflection of the laser beam through an appropriate ND filter, the reflective image of the laser beam in the tumor site showed comparatively darker than that of normal part. The reflection image data on the targets to the laser beam were processed to the brightness profiles using a computer, and the results assured that the brightness contrast in the reflective image between the tumor site and the normal part was sufficient for discriminating the cancerous site from the normal part. The advantages of this diagnostic method are suninarized as follows : New method for visual capture of cancerous area, ® Simplicity in technique and examination system, ® Easy operation, J Non-invasive and complete safety. The result of this experiment hopefully proposes a new imaging diagnostic method for imilignant tumors depending on the optical characteristics.

The authors spectral analyses for each dental structure proposed the capability that the photo thermal reaction on teeth to laser exposure might be applicable for the detection of dental caries. This experiment was designed intending to realize the prospect on the fundamental experiment. Nd:YAG laser was selected for the appropriate wavelength for this experiment. When the occlusal surface of an extracted human tooth was exposed to a fine shot of the laser beam, the temperature of the occlusal surface was measured with an infrared thermometer and a real time infrared thermocamera, and the relative intensity of the transillumination of the laser beam at the lateral surface of the subject tooth was evaluated as a thermal data using a thermocouple. As the result of this experiment, the photo-thermal reaction of carious teeth following exposure to the Nd:YAG laser beam was different so contrastively from that of intact teeth that this method was successfully applied to the detection of dental caries, and the carious site was apparently discriminated as a thermal pattern of high temperature on the real time thermogram. The detectability for dental caries in this method was so sensitive as to identify even a fine dental anomaly which was hardly discriminated from a normal tooth in the usual roentgenography. The advantages in this method are summarized as follows : Much more sensitive for detection of dental caries than roentgenography, Complete safety, Simplicity in technique, ® Not only the existence of dental caries but also the decayed area can be detected, ® In addition to these, the results of this experiment hopefully propose the expectation that this method might lead to an epoch-making technique for evaluating biochemical quality of teeth.

Ultrasound images obtained with a simple sector scan show a granular appearance, called `speckle'. The speckle is the useless property of the ultrasound introskopic images as it mask all small differences of the images. The possibility of the speckle noise reduction by special created filter is analyzed. The computer processing results of ultrasound introskopic thyroid gland images by such filter are presented.

A new gamma camera using a-Si:H photodetectors has been designed for the imaging of heart and other small organs. In this new design the photomultiplier tubes and the position sensing circuitry are replaced by 2D array of a-Si:H p-i-n pixel photodetectors and readout circuitry which are built on a substrate. Without the photomultiplier tubes this camera is light weight, hence can be made portable. To predict the characteristics and the performance of this new gamma camera we did Monte Carlo simulations. In the simulations 128 X 128 imaging array of various pixel sizes were used. ^99mTc (140 keV) and ²⁰¹Tl (70 keV) were used as radiation sources. From the simulations we could obtain the resolution of the camera and the overall system, and the blurring effects due to scattering in the phantom. Using the Wiener filter for image processing, restoration of the blurred image could be achieved. Simulation results of a-Si:H based gamma camera were compared with those of a conventional gamma camera.

Scatter to primary radiation ratios at detector positions were calculated for spherical water phantoms in the cone-beam x-ray 3D CT scanner using the Monte Carlo simulation method. In the Monte Carlo simulation, 1 approximately 4 X 10⁶ photons of cone-beamed several monoenergetic x-rays have been traced individually as these photons interact within the phantoms with diameters 10 and 15 cm. In case of the incident photon energies are 72.86 and 68.79 keV, photon spectra at the center of the planar detector system are depicted. The decrease in CT number due to scatter was found to be about 100. The scatter reduces CT numbers in the whole region of images and the intensity of scattered rays is nearly uniform, so that rounded-off image is reconstructed.