We discuss how the frequency dependent detective quantum efficiency [DQE(f)] in a well-designed amorphous silicon flat panel detector is affected by several phenomena that reduce the DQE in other types of medical imaging detectors. The detector examined employs a CsI(Tl) scintillator and is designed for general diagnostic x-ray imaging applications. We consider DQE degradation due to incomplete x-ray absorption, secondary quantum noise, Swank factor, Lubberts effect, spatial variation in gain, noise aliasing, and additive system noise. The influences of detector design parameters on the frequency- and exposure-dependent DQE are also examined. We find that the DQE does not depend directly on MTF and that DQE is independent of exposure within the detector's operating range, except at the lowest exposures. Likewise the signal per absorbed x-ray, which contains the fill factor as one of several multiplicative components, does not affect DQE except at the lowest exposures. A methodology for determining DQE(f) from measurements of MTF(f), noise power spectrum (NPS), average signal, and x-ray exposure is presented. We find that it is important to incorporate several corrections in the NPS measurement procedure in order to obtain accurate results. These include corrections for lag, non-linearity, response variation from pixel to pixel, and use of a finite number of flat-field images. MTF, NPS, and DQE results are presented for a 41 X 41-cm² flat panel detector designed for radiographic applications.

This paper introduces the key design optimizations which have been carried out recently in Trixell in order to prepare the future family of large area, combined static (Radiography) and dynamic (Fluoroscopy, Cardio...) digital X-ray detectors based on a-Si:H/CsI:Tl flat panel technology. These optimizations have been carried out on a 16' X 12' prototype that has been designed and built in a product-oriented way. We describe the detector technology and give some of its main characteristics, as well as some preliminary measurement results.The heart of the new prototype is a Cesium Iodide scintillating screen, directly evaporated onto a 2 K X 2.5 K pixel, array of amorphous silicon photodiodes and TFTs deposited on a glass substrate. The pixel pitch is 155 micrometer. The detective flat panel is connected to dedicated electronics which provides line addressing, low-noise column readout and multiplexing into a serial electrical signal. This signal is digitized over 14 bits to provide a direct digital image output, available for the host radiology system via an optical fiber. This type of detector (flat panel + electronics) is built into a light and thin (less than 100 mm) packaging which can be easily integrated in various x-ray equipment such as R&F tables, Angiography systems (incl. Cardiology), and mobile C-arm systems.

The performance of the new generation of high fill factor two- dimensional imagers with high spatial resolution and low data line capacitance is described. These arrays have a continuous a-Si:H sensor layer deposited over the whole imager to improve sensitivity. We have studied charge collection and lateral leakage in the gap region in between two neighboring pixels. Experiments demonstrate that a 10 micrometer gap between pixels leads to an effective fill factor of approximately 92% and can be fabricated in a way to reduce the charge leakage between pixels to a very low level. We have also studied the capacitance of the data lines that can lead to increased electronic noise, degrading the imager performance. Experimental determination of the actual capacitance for different insulator materials are compared with numerical simulations, to identify the optimum structure. Based on these results, the new imager generation could be manufactured with a total parasitic capacitance of about 6 fF/pixel. Finally, we report measurements of the high fill factor imager under light and X-ray exposures.

Described are two types of direct-detection flat-panel X-ray detectors utilizing amorphous selenium (a-Se) and cadmium telluride (CdTe). The a-Se detector is fabricated using direct deposition onto a thin film transistor (TFT) substrate, whereas the CdTe detector is fabricated using a novel hybrid method, in which CdTe is pre-deposited onto a glass substrate and then connected to a TFT substrate. The detector array format is 512 X 512 with a pixel pitch of 150 micrometer. The imaging properties of both detectors have been evaluated with respect to X-ray sensitivity, lag, spatial resolution, and detective quantum efficiency (DQE). The modulation transfer functions (MTFs) measured at 1 lp/mm were 0.96 for a- Se and 0.65 for CdTe. The imaging lags after 33 ms were about 4% for a-Se and 22% for CdTe. The DQE values measured at zero spatial frequency were 0.75 for a-Se and 0.22 for CdTe. The results indicate that the a-Se and CdTe detectors have high potential as new digital X-ray imaging devices for both radiography and fluoroscopy.

A first image of some tiny screws were obtained for the first time with polycrystalline HgI₂ acting as the photoconductor material deposited on a-Si direct conversion X- ray image sensors, produced by Xerox -- Palo Alto Research Center. The initial results are very promising and show a high X-ray sensitivity and low leakage current. The response of these detectors to a radiological X-ray generator of 65 kVp has been studied using the current integration mode. Already its sensitivity expressed in (mu) C/R*cm², is very high, values of 20 (mu) C/R*cm² have been measured for films of 100 - 250 microns thickness and bias of 50 - 200 volts respectively, which is superior to the published data for competing materials such as polycrystalline PbI₂ and a-Se detectors. The fabrication and characterization measurements of the Polycrystalline HgI₂ thick film detectors will be given. The characterization data which will be reported here consists of: (a) sensitivity, (b) dark currents, (c) stability of sensitivity dependence on the number of exposure, (d) X-ray response dependence on dose energy and (e) signal decay dependence on the number of exposures.

In the near future, it will be possible to perform coincidence detection on a gamma camera with three heads, which increases the geometric sensitivity of the system. Different geometric configurations are possible, and each configuration yields a different geometric sensitivity. The purpose of this work was to calculate the sensitivities for different three-headed configurations as a function of the position in the field of view, the dimensions of the detector heads and the distance of the heads from the center of the field of view. The configurations that were compared are: a regular two headed configuration (180 deg. opposed), a triple-headed configuration with the three heads in an equilateral triangle (120 deg.), and a triple-headed configuration with two heads in a regular two headed configuration, and the third perpendicular between the first two, which makes a U-shaped configuration. An expression was derived for any planar detector configuration to calculate the geometric sensitivity for each Line Of Response (LOR). This sensitivity was integrated to get the sensitivity profile, which gives the geometric sensitivity at a certain distance from the center of rotation. We found that the triangular configuration gave the best sensitivities when placed very near to each other (nearly full ring configuration), but for larger fields of view, the U-shaped configuration performed better.

The new high performance radioisotopic imager IRIS we recently developed has an intrinsic spatial resolution of about 1.7 mm. To obtain a good sensitivity, it is necessary to use a collimator with holes greater than this spatial resolution. However, this induces artifacts in images. In this paper, we describe a new concept: a moving collimator. Deterministic and Monte Carlo simulations demonstrate that it is possible to obtain simultaneously global high spatial resolution and good sensitivity in radioisotopic imaging.

We apply task-based assessment of image quality to optical tomography imaging systems. In particular, we studied the task of detecting a signal, specified as a change in scattering and absorption coefficients, when its shape and location were known. The detectability was quantified using the optimal linear (Hotelling) observer. The non-linearity of the problem was no impediment in computing the Hotelling observer using a hybrid approach that combines knowledge of the measurement statistics with sampling to account for anatomical variation. We compared the observer performance on the raw data in uniform and structured backgrounds for several data and signal types. Two of the data types studied were the total number of photons (total counts) collected for each source-detector pair and their respective mean time of arrival. Results show that the spatial variation of detectability was different for the total counts than for the mean time. The performance of the total counts and its relative performance to the mean time varied significantly with both signal type and background variations.

We are developing a novel concept for portal imaging that would allow for on-line control and verification of the radiation treatment of cancer patients both at diagnostic and therapeutic energies. This device will consist of two consecutive detectors confided in one gas chamber: a KeV- photon detector, which can visualize the internal soft tissue of the patient, and an MeV-photon detector, which will measure the absolute intensity of the therapeutic beam and its position with respect to the tumor and normal tissues. Both detectors are based on gas and solid photon to electron converters combined with recently invented gas electron multipliers. The device will have a common charge collecting pad-type readout plate equipped with ASIC-based electronics for both detectors. A first simplified prototype device has recently been built and extensively tested. Special efforts were made to find conditions for a safe and reliable operation of the readout electronics that can be damaged by plasma-type discharge effects induced specially at high dose rates. Results obtained so far indicate that our new detector concept may satisfy all requirements on advanced therapy beam monitoring systems.

The European Synchrotron Radiation Facility Medical Research Beamline is now fully operational. One of the primary programs is the development of dual-energy transvenous coronary angiography for in vivo human research protocols. Previous work at this and other synchrotrons has been entirely devoted to the use of the dual-energy digital subtraction technique at the iodine k-absorption edge at 33.17 keV. The images are recorded in a line scan mode following venous injection of the contrast agent. Considerations of the patient dose, the dilution of the contrast agent in the pulmonary system and the arteries overlying the filled ventricles have limited the image quality. The ESRF facility was designed to allow dual- energy imaging at higher energies, for example at the gadolinium k-absorption edge at 50.24 keV. The advantages have been theoretically known for many years, with the higher energy promising higher image quality with less radiation dose. During the commissioning phase of the ESRF angiography program, the opportunity presented itself to image adult pigs in vivo with Gd contrast agent. This paper presents some initial results of the image quality in the Gd studies in comparison with iodine contrast agent studies, also carried out in adult pigs at the ESRF.

We are investigating active matrix flat panel x-ray detectors for real-time operation in the fluoroscopic exposure range. The typical exposure range for fluoroscopy is low, (0.1 - 10 (mu) R/frame). Hence the flat panel must be very sensitive to produce quantum noise limited images. The application of avalanche multiplication in amorphus selenium, ((alpha) -Se) is examined. Avalanche multiplication, M, can be used to increase signal size, potentially eliminating the quantum sink at low exposure levels. However, M greater than 1 also causes an overall degradation of DQE due to the addition of a new source of gain fluctuation noise. Using a linear cascaded systems model, this noise can be expressed as an additional avalanche Swank factor A_(alpha ) in the expression for DQE(0) equals (eta) A_sA_(alpha ) where (eta) is the quantum efficiency, A_S is the conventional conversion gain Swank factor. Depending upon the parameters, the value of A_(alpha ) can vary from unity to less than 0.2. Our model was in agreement with experimentally observed values of A_(alpha ) obtained using an imaging system (HARP) with (alpha) -Se layers capable of avalanche multiplication. The results indicate that a balance must be maintained between the improvement from avalanche multiplication to the amplifier noise limited part of the image, and the degradation effect it has on the quantum noise limited parts of the image. It also suggests that proper engineering of the avalanche layer can minimize, and perhaps eliminate, the additional noise fluctuations arising from avalanche multiplication of x-ray signals.

One of the issues in (alpha) -Si:H X-ray detectors is signal to noise ratio for low dose fluoroscopic applications. An optimized sensitivity of the X-ray detection system together with low and isotropic system noise characteristics are primary pre-conditions needed for maximum image quality. However, in spite of high DQE numbers of this Flat Detector technology in radiological and fluoroscopic application areas, a SNR for low dose fluoroscopy is found, which is inferior to that found with Image Intensifier-TV based systems. The problem area is a small dose range, producing gray levels just above absolute dark. Except for the dark level, these levels can (dependent on the application area) contain clinically relevant information. Since scatter affects the darker parts of the relevant image areas there will be noise in those areas, caused by X-ray quantum statistics and readout noise. The objective of the simulations is to investigate the influence of the various system noise components on the image quality. A level of system noise can be found where the subjective image quality is mainly determined by the X-ray quantum statistics and where the readout noise does not necessarily have to be invisible in totally dark parts. The simulation concerns a threshold contrast detail detectability (TCDD) observation test, where observers score discs of various diameter and absorption in an image sequence (being a fixed scene of the test object with (temporal) X-ray noise and system noise). The dynamic sequence is based upon total simulation, i.e. the test object as well as the X-ray noise and the system noise components were simulated. To verify the simulations also an image sequence was acquired on a Flat Detector system. The observations are done at various dose levels, with and without post processing to obtain noise reduction like it is used in clinical practice for this kind of system. The sequences are observed on a medical CRT display.

The use of selenium alloys for direct conversion fluoroscopy flat-panel detectors has been underestimated. The purpose of this paper is to demonstrate the salient features of a selenium-based detector designed for R&F applications. The detector has an active area of 30 cm X 27 cm and comprises 2048 X 1792 pixels at 150 micrometer pitch. The geometric fill factor is 66%, but experimental evidence supports the fact that internal electric field bending leads to an effective fill factor approaching unity. The detector is designed to support full resolution images at 15 frames/second, and 896 X 1024 resolution at 30 frames/second. The detector is coated with a simple coplanar 'p-i-n' selenium diode structure which has a dark current less than 100 pA/cm². The thickness of this structure is 1000 micrometer to absorb 77% of a NEMA standard fluoroscopy beam. Measurements show we have obtained an x-ray sensitivity of 4400pC/mR/cm², which translates to 1212 collected charges per absorbed x-ray. Resolution was measured to be near the theoretically predicted values, with a modulation of 63% at the Nyquist limit of 3.33 lp/mm. Phantom images were obtained at a frame rate of 15 frames per second, and negligible lag was observed in this image sequence.

We evaluate a fluoroscopy mode for a CCD-based digital radiography system. It is a modular system based on known technologies (scintillator screen, lenses, fast and high resolution CCD cameras) and optimized for medical purpose. The detector can provide different acquisition modes: a large field (43 cm) and high resolution (pixel size 105 micrometer) mode, but also fluoroscopy modes: acquisition field size 21 cm, pixel size 210 micrometer or 420 micrometer, image rate of 6 or 10 images/s. We investigate the image quality for a fluoroscopy mode (pixel size 210 micrometer) through two criteria: Modulation Transfer Function (MTF) and Detection Quantum Efficiency (DQE) as function of spatial frequency. These results show that it is possible to have a fluoroscopy mode at a low frame rate on a CCD-based sensor with an interesting image quality. Our detector will be the first CCD- based sensor that will provide a fluoroscopy mode in addition to a high resolution acquisition mode.

A new flat X-ray image intensifier (Flat II) system, using a large-area electron multiplier for applications in real-time fluoroscopy imaging, is under development by the authors. The Flat II system, mainly consists of two devices; the image processing equipment and the image acquisition system by charge-coupled-device (CCD) camera. The image processing is performed as follows. The X-ray is converted to visual light by cesium iodide scintillator and to electrons by photo- cathode. A large-area electron multiplier is located adjacent to the photo-cathode, and amplifies electron current up to a few hundred times. Amplified electrons are again converted to bright visual image by the output phosphor screen. The bright visual image is subsequently detected with a CCD camera system. The electron multiplier used for present work is that of metal-dynode array construction. The principle of metal- dynode electron multipliers is well known. Since it is technically difficult to make the electron multiplier with a fine pitch, sufficient resolution could not be obtained until now. However, we could manufacture that of 6-inches (15 cm) size with 0.3 mm and 0.2 mm pitch, and manufactured the prototype Flat II system using the 0.3 mm pitch multiplier as an experiment. We acquired the fundamental characteristics and the image quality of prototype Flat II system. In this paper, the physical characteristics such as modulation transfer function (MTF), signal-to-noise ratio (SNR), etc. are discussed.

Digital mammography has demanding imaging requirements, including very high spatial resolution (50 micrometer) and SNR. To make efficient use of the radiation dose, it is also desirable that the DQE of the image receptor is high. To achieve these requirements, a prototype CCD read-out has been designed, which is hybridized to a semiconductor array to form a direct conversion detector that can be employed in a slot- scanned digital x-ray imaging system. The image quality of the detectors in which the CCD is hybridized to either a silicon photodiode array or a CdZnTe photoconductor array has been measured. A 1 mm thick silicon device has shown a DQE(f) of 0.64 at 0 mm^-1, falling to 0.14 at 10 mm^-1 in the slot direction (20 keV). The CdZnTe hybrid device is very thin (150 micrometer) and has a theoretical DQE in excess of 0.9 at 20 keV. The resolution of the CdZnTe device is excellent, with an experimental MTF that is limited only by the detector element size, and the TDI scanning technique. However, the experimental DQE is lower than predicted, believed to be due to crystal non-uniformity, and excessive carrier trapping. Future work will investigate the improvement in image quality obtainable by using a very high purity single-crystal CdZnTe device.

Pixel size is an important parameter in digital mammography because it directly influences both the image quality and the cost of the imaging system. We have investigated the effects of pixel/aperture sizes on image properties in digital mammography. Studies were made with a small field digital mammography unit (SenoVision by GE Medical System, Milwaukee, WI) which provides a 30 micrometer X 30 micrometer pixel/aperture size. Pixel-averaging was used to increase both sampling distance and aperture size. Sub-sampling was used to increase the sampling distance without altering the aperture size. The effective pixel size was increased by pixel- averaging or sub-sampling. A tilted slit camera was employed to measure the presampling MTF. Uniform exposure images were used to measure SNRs and NEQs for various pixel sizes. Simulated microcalcifications of various sizes were imaged to evaluate the low contrast performance as well.

CCD image sensors have been utilized as the detectors in digital spot mammography for needle biopsy systems. The requirements of a detector for this application are large sensing area, large pixel size, moderate resolution, and low noise. A large area full-frame CCD image sensor has been developed that is 50 mm X 50 mm with 2084 X 2084 pixels. It is manufactured using a true two-phase, transparent gate, buried channel CCD process. The transparent gate process improves the quantum efficiency from 40 percent to 55 percent at 540 nm. The true two-phase CCD operates in accumulation mode for low dark current without compromising charge capacity. The sensor has two different output amplifiers optimized for either low signal applications or high dynamic range applications. At a 2 MHz pixel rate the dynamic range is 84 dB or 88 dB depending on which output amplifier is selected. The large sensor format allows the elimination of the demagnification requirement in digital spot mammography applications and may be useful for mid-field digital diagnostic systems.

The development of a new high spatial resolution x-ray detector system is described. The prototype detector is based on a patented detector technology that utilizes selenium for the x-ray conversion material, charge storage capacitors, and a thin film transistor (TFT) array for reading out the charge image. This experimental detector consists of a 512 X 512 matrix with a pixel pitch of 70 microns. The selenium layer deposited on the TFT array is 250 microns thick. With a low absorption entrance window the system is optimized for an energy range of 10 - 30 keV, and is designed for applications that require high spatial resolution and low noise. This presentation describes the imaging performance of the detector using the DQE and MTF metrics. Example images of phantoms are shown. Previously, we demonstrated a practical flat-panel self-scanned digital radiography system based on amorphous selenium and TFT technology. This system is being used clinically for chest radiography and general musculoskeletal imaging, and in industrial applications. The current work demonstrates the feasibility of adapting this technology for applications requiring higher spatial resolution.

We have applied a synchrotron radiation computed tomography (SRCT) system to the lung specimens and evaluated its resolving power compared with the histopathologic appearances, precisely. An SRCT system has been constructed in the bending magnet beamline at the SPring-8. The system consists of a double-crystal monochromator, a rotating sample stage, a fluorescent screen, and a charge-coupled device (CCD) array detector (1024 X 1024 pixels with 12 X 12 micrometers ² pixel size). The energy of the x-ray beam was tuned to 9 - 12 keV. The lungs obtained at autopsy were inflated and fixed by Heitzman's method. A cylindrical specimen (diameter, approximately 8 mm; height, 15 - 25 mm) was rotated in the plane parallel to the beam. The detected signal was transferred to a workstation; then, SRCT images (matrix size, 800 X 800 pixels) were reconstructed by a filtered back- projection. Finally, 6 - 12 micrometer-thick microscopic sections were obtained and stained with hematoxylin and eosin for histopathologic examination. SRCT images well depicted the terminal bronchiole, respiratory bronchiole, alveolar duct, alveolar sac, and alveolar septum. Different pathologic processes (alveolar hemorrhage, alveolitis) demonstrated on SRCT images were well correlated with the histopathologic appearances.

A technique called Variable-Resolution X-ray (VRX) detection that greatly increases the spatial resolution in computed tomography (CT) and digital radiography (DR) is presented. The technique is based on a principle called 'projective compression' that allows the resolution element of a CT detector to scale with the subject or field size. For very large (40 - 50 cm) field sizes, resolution exceeding 2 cy/mm is possible and for very small fields, microscopy is attainable with resolution exceeding 100 cy/mm. Preliminary results from a 576-channel solid-state detector are presented. The detector has a dual-arm geometry and is comprised of CdWO₄ scintillator crystals arranged in 24 modules of 24 channels/module. The scintillators are 0.85 mm wide and placed on 1 mm centers. Measurements of signal level, MTF and SNR, all versus detector angle, are presented.

Mammographic density (MD) has been shown to be a strong risk predictor for breast cancer. Compared to subjective assessment by a radiologist, computer-aided analysis of digitized mammograms provides a quantitative and more reproducible method for assessing breast density. However, the current methods of estimating breast density based on the area of bright signal in a mammogram do not reflect the true, volumetric quantity of dense tissue in the breast. A computerized method to estimate the amount of radiographically dense tissue in the overall volume of the breast has been developed to provide an automatic, user-independent tool for breast cancer risk assessment. The procedure for volumetric density estimation consists of first correcting the image for inhomogeneity, then performing a volume density calculation. First, optical sensitometry is used to convert all images to the logarithm of relative exposure (LRE), in order to simplify the image correction operations. The field non-uniformity correction, which takes into account heel effect, inverse square law, path obliquity and intrinsic field and grid non- uniformity is obtained by imaging a spherical section PMMA phantom. The processed LRE image of the phantom is then used as a correction offset for actual mammograms. From information about the thickness and placement of the breast, as well as the parameters of a breast-like calibration step wedge placed in the mammogram, MD of the breast is calculated. Post processing and a simple calibration phantom enable user- independent, reliable and repeatable volumetric estimation of density in breast-equivalent phantoms. Initial results obtained on known density phantoms show the estimation to vary less than 5% in MD from the actual value. This can be compared to estimated mammographic density differences of 30% between the true and non-corrected values. Since a more simplistic breast density measurement based on the projected area has been shown to be a strong indicator of breast cancer risk (RR equals 4), it is believed that the current volumetric technique will provide an even better indicator. Such an indicator can be used in determination of the method and frequency of breast cancer screening, and might prove useful in measuring the effect of intervention measures such as drug therapy or dietary change on breast cancer risk.

There is a lot of effort to develop digital detectors for mammography, for example for the screening programs. For this development it would be very helpful to know, which structure sizes have to be reproduced. Besides that, the information content of images produced by digital systems may be influenced by aliasing artifacts, if there are frequency components in the incoming signal higher than the Nyquist- frequency of the detector. Therefore the spatially modulated pattern of the X-ray intensity in a mammogram has to be known, but so far only little information is available. The method to measure and analyze the X-ray intensity pattern of a radiograph of the thorax in the detector plane which was presented at 'Medical Imaging' conference last year has been further developed in a way, that it meets the requirements for determining the intensity pattern in mammography.

Scintimammography is a promising technique for breast cancer detection. Scintimammography uses radiotracer containing ^99mTc that emits 140 keV gamma photons. We developed a small field of view gamma ray imaging probe called IRIS. A possible application of this probe is scintimammography. IRIS is composed by a single NaI(Tl) scintillator coupled to a 5 inch round PSPMT. In order to optimize compromise between resolution and detection efficiency, we developed a Monte Carlo code modeling light transport in NaI crystals. The thickness of the scintillator (4 mm) was optimized for ^99mTc imaging. We also designed a high-resolution collimator with a 35 mm thickness and 1.7 mm hole diameter. Detection efficiency of the crystal is 65% at 140 keV. IRIS shows a 2.5 mm global spatial resolution in contact. Energy and spatial corrections allow a +/- 5% uniformity and an energy resolution better than 10% at 140 keV. IRIS has a 10 cm field of view and a 13 cm external diameter at the entrance face. The small size of the detector head allows placing the detector close to the breast, improving global spatial resolution. The high-resolution gamma ray imaging probe IRIS shows physical characteristics well suited for ^99mTc breast imaging.

We are developing a dose-efficient system for digital mammography. The system incorporates a refractive x-ray lens combined with a photon counting silicon sensor system. The system matches the digital nature of x-rays and maximum information can be extracted from the transmitted x-ray beam. The photon counting sensor system opens the possibility to weight each photon according to its information content. High- energy photons with little contrast information can be weighted low while low energy photons are weighted high. At the same time, using the refractive x-ray lens, the incident x-ray spectrum can be tuned to the optimum energy for a certain breast thickness. The performance of the system is expected to be close to the theoretical limit in terms of dose efficiency and is entirely limited by quantum noise. Since the standard use of the term DQE does not take into account the incident x-ray energy spectrum, scattered x-rays or the difference in information content from low to high-energy photons an alternative measure of performance is used. We refer to this measure as dose-efficiency and in this case the system is related to an ideal x-ray imaging system.

A theoretical investigation of the system performance of active matrix, flat-panel imagers (AMFPIs) under mammographic imaging conditions is reported. These imagers employ either indirect or direct detection of the incident radiation. The x- ray converter materials assumed in these studies were Gd₂O₂S:Tb and CsI:Tl scintillators for indirect detection and a-Se and PbI₂ photoconductors for direct detection. A model based on cascaded systems formalism was used to predict the detective quantum efficiency (DQE) of various AMFPI designs incorporating these converters. The functional dependence of DQE performance on pixel-to-pixel pitch, collection fill factor and additive electronics noise was investigated under mammographic conditions. Incorporation of a continuous photodiode structure in indirect detection AMFPI arrays is necessary in order to achieve reasonable DQE performance for devices with pixel pitches significantly below 100 micrometer. In the case of CsI:Tl, a-Se, and PbI₂, through optimization of the converter thickness, the DQE performance of these advanced AMFPI designs is predicted to exceed that of an AMFPI system incorporating a conventional Gd₂O₂S:Tb mammographic screen. Finally, the model predicts that the incorporation of a high gain converter such as PbI₂ leads to a high value of DQE which is independent of exposure over the mammographic exposure range at all spatial frequencies, even at 50 micrometer pitch.

This paper introduces a new multislice helical weighting method. When used in conjunction with rebining from fan-beam to parallel projections, this approach leads to a single- filtering algorithm, that is, each projection needs to be filtered only once independently of the number of image reconstructions performed. This is particularly useful when overlapping images are reconstructed to fully leverage the z- resolution available from the projection data, or when performing CT fluoroscopy, where up to twelve or more images are reconstructed per 360-degree source rotation. A variable helical contribution width method is presented, which allows helical reconstruction for any selected pitch. The algorithm differs significantly from the current published helical weighting methods. In particular it can be applied at any helical pitch in a relevant continuum and to a system with any number of rows (in as much as cone-angles may be ignored). The interpolation method is chosen in such a way that the number of projections contributing to an image is adjustable. This results in user-selectable image thickness as well as user- selectable pitch. Further the minimum number of projections to be used is significantly smaller than that used with the current algorithms.

In this paper, we will introduce a resampling method for in vivo projection reconstruction (PR) magnetic resonance (MR) signals. We will describe the physical processes causing the inaccurate sampling of these signals. Based on the theoretical properties of the signal, a technique to reduce the influence of this effect on the signals will be proposed. The method will be validated using simulations and in vivo MR signals. The corrected signals will be shown to be a better approximation of the signals that would be expected on a theoretical basis.

We present two new approaches to single-slice helical computed tomography image reconstruction that exploit the fast Fourier transform and the Fourier shift theorem to generate from the helical projection data a set of fan-beam sinograms corresponding to equispaced transverse slices. Slice-by-slice reconstruction is then performed by use of two-dimensional fan-beam algorithms. The first approach, called 360 FT, makes us only of the directly measured projection data, but an extension called 180 FT exploits the redundancy of fan-beam data acquired over 360 degree to generate a second set of longitudinal samples at each projection angle and bin. The proposed approaches are compared to their counterparts based on the use of linear interpolation -- the 360 LI and 180 LI approaches. Particular attention is paid to the spatially variant aliasing that has recently been shown to be an important consideration in helical CT, and which affects the 360 LI and 180 LI approaches equally despite the latter's use of the additional fan-beam redundancy samples. Remarkably, because of the way it makes use of the fan-beam redundancy data, the proposed 180 FT approach is found to eliminate this spatially variant aliasing under certain conditions. This is a clear advantage of the approach and represents a step toward the desirable goal of achieving isotropic resolution in reconstructed helical CT volumes.

The fan-beam filtered backprojection (FFBP) algorithm has been widely used for image reconstruction in fan-beam computed tomography (CT). We have also developed fullscan-hybrid algorithms for image reconstruction in fan-beam CT. Numerical evaluation indicates that such fullscan-hybrid algorithms possess computational and noise properties superior to those of the FFBP algorithm. However, because these fullscan-hybrid algorithms involve the calculation of the Fourier series expansion of the fan-beam data, they require fullscan data acquired over an angular range of 2(pi) . Therefore, these fullscan-hybrid algorithms cannot be applied directly to a halfscan fan-beam data that are acquired only over an angular range of (pi) plus the fan angle. In this work, we present new reconstruction algorithms -- the halfscan-hybrid algorithms -- for image reconstruction in halfscan CT. These halfscan-hybrid algorithms have properties that are similar to those of the full-scan hybrid algorithms and that are superior to those of the widely used halfscan-FFBP algorithm. Such properties of the halfscan-hybrid algorithms may have significant implications to the accuracy and precision of lesion detection and parameter estimation in noisy CT images without increasing the radiation dose to the patient.

Many studies have been conducted on the utilization of solid state detectors for computed tomography (CT). One of the important performance parameters for the solid state detector has been shown to be the primary speed and afterglow. In this paper, we present a detailed investigation on the signal decay characteristics of the HiLight^TM scintillating detector. We first develop an analytical model to fully characterize the detector impulse response. The model sensitivity to x-ray photon energy, detector aging, and radiation exposure is then established and analyzed. The impact of various decay time constants on CT image quality is subsequently illustrated with computer simulations and phantom experiments. Finally, a recursive correction approach is derived and evaluated.

Microwave imaging is a relatively new modality to perform non- invasive diagnostic of biological tissues. The potential of these techniques results from the dependence of the dielectric properties of these tissues with respect to quantities of practical and clinical relevance such as water content, blood flow rate, temperature, etc. Initiated in the 80s, these techniques have suffered from the complexity of the interaction mechanisms between microwave beams and biological structures. The compensation of strong diffraction effects by high dielectric contrast structures requires efficient reconstruction algorithms. Recently, major advances have been achieved for improving the time resolution and the quantification of microwave images. This paper is focused on the first aspect. It reports some unique results obtained, for the first time, with a microwave camera providing qualitative images of biological targets at the rate of 15 images per second. Image reconstruction is performed by means of spectral diffraction tomography algorithms. The camera consists of a 2D array of 32 X 32 sensors, covering a square area of approximately 22 cm X 22 cm. The camera is operated at 2.45 GHz, according to the Modulated Scattering Technology (MST). The biological target is immersed in water (or can be inserted between bolus) and illuminated by a lens-compensated horn antenna. From the amplitude/phase measurement of the field scattered by the target, microwave images can be reconstructed, thanks to numerical focusing, in any plane located between the transmitting antenna and the camera. Typically, the investigation depth is 25 cm, and the spatial resolution is about 5 mm. The capabilities of this microwave camera will be illustrated by means of a short VHS video tape showing quick motions of living structures. Expected improvements of the camera performances are discussed and possible clinical applications are analyzed.

Experimental and theoretical studies of image quality using iodinated contrast agent and x-ray spectra generated by laser- based x-ray source were performed. A TableTop Terawatt (T³) laser (intensity: 10¹⁷ - 10¹⁹ W/cm^-2, pulse duration: 150 fs or 450 fs, with or without controlled pre-pulse) was used to crate x-ray source. Infrared and/or green beams were utilized. Ba, La, Ce, Nd, and Gd laser targets were used. For each target, a number of suitable filters was utilized to produce optimized x-ray spectra for a specific imaging task. The MTF function due to the focal spot was obtained. A simple theoretical model of x-ray detector response was developed. An index of image quality (Detective Image Quality) as well as a figure of merit for dual energy imaging FOM(DESA) were defined and optimized via x-ray spectrum manipulation. The optimum, for a specific imaging task, technique parameters such as: target/filter combination, focal spot size, laser-light wavelength and surface power density, laser pulse duration, pre-pulse delay and contrast ratio, and hot electrons temperature were obtained experimentally and confirmed theoretically. We found that an optimized laser-based x-ray source can outperform conventional x-ray tube-based source in application to vascular imaging in terms of contrast resolution and spatial resolution.

A crystal diffraction lens for focusing energetic gamma rays has been developed at Argonne National Laboratory for use in medical imaging of radioactivity in the human body. A common method for locating possible cancerous growths in the body is to inject radioactivity into the blood stream of the patient and then look for any concentration of radioactivity that could be associated with the fast growing cancer cells. Often there are borderline indications of possible cancers that could be due to statistical functions in the measured counting rates. In order to determine if these indications are false or real, one must resort to surgical means and take tissue samples in the suspect area. We are developing a system of crystal diffraction lenses that will be incorporated into a 3- D imaging system with better sensitivity (factors of 10 to 20) and better spatial resolution (a few mm in both vertical and horizontal directions) than most systems presently in use. The use of this new imaging system will allow one to eliminate 90 percent of the false indications and both locate and determine the size of the cancer with mm precision. The lens consists of 900 single crystals of copper, 4 mm X 4 mm on a side and 2 - 4 mm thick, mounted in 13 concentric rings.

Bone tissue consists primarily of calcium hydroxyapatite crystals (bone mineral) and collagen fibrils. Bone mineral density (BMD) is commonly used as an indicator of bone health. Techniques available at present for assessing bone health provide a measure of BMD, but do not provide information about the degree of mineralization of the bone tissue. This may be adequate for assessing diseases in which the collagen-mineral ratio remains constant, as assumed in osteoporosis, but is insufficient when the mineralization state is known to change, as in osteomalacia. No tool exists for the in situ examination of collagen and hydroxyapatite density distributions independently. Coherent-scatter computed tomography (CSCT) is a technique we are developing that produces images of the low- angle scatter properties of tissue. These depend on the molecular structure of the scatterer making it possible to produce material-specific maps of each component in a conglomerate. After corrections to compensate for exposure fluctuations, self-attenuation of scatter and the temporal response of the image intensifier, material-specific images of mineral, collagen, fat and water distributions are obtained. The gray-level in these images provides the volumetric density of each component independently.

A 3D dynamic computed-tomography (CT) scanner was developed for imaging objects undergoing periodic motion. The scanner system has high spatial and sufficient temporal resolution to produce quantitative tomographic/volume images of objects such as excised arterial samples perfused under physiological pressure conditions and enables the measurements of the local dynamic elastic modulus (E_dyn) of the arteries in the axial and longitudinal directions. The system was comprised of a high resolution modified x-ray image intensifier (XRII) based computed tomographic system and a computer-controlled cardiac flow simulator. A standard NTSC CCD camera with a macro lens was coupled to the electro-optically zoomed XRII to acquire dynamic volumetric images. Through prospective cardiac gating and computer synchronized control, a time-resolved sequence of 20 mm thick high resolution volume images of porcine aortic specimens during one simulated cardiac cycle were obtained. Performance evaluation of the scanners illustrated that tomographic images can be obtained with resolution as high as 3.2 mm^-1 with only a 9% decrease in the resolution for objects moving at velocities of 1 cm/s in 2D mode and static spatial resolution of 3.55 mm^-1 with only a 14% decrease in the resolution in 3D mode for objects moving at a velocity of 10 cm/s. Application of the system for imaging of intact excised arterial specimens under simulated physiological flow/pressure conditions enabled measurements of the E_dyn of the arteries with a precision of +/- kPa for the 3D scanner. Evaluation of the E_dyn in the axial and longitudinal direction produced values of 428 +/- 35 kPa and 728 +/- 71 kPa, demonstrating the isotropic and homogeneous viscoelastic nature of the vascular specimens. These values obtained from the Dynamic CT systems were not statistically different (p less than 0.05) from the values obtained by standard uniaxial tensile testing and volumetric measurements.

The improved image quality and characteristics of new flat- panel x-ray detectors have renewed interest in advanced algorithms such as tomosynthesis. Digital tomosynthesis is a method of acquiring and reconstructing a three-dimensional data set with limited-angle tube movement. Historically, conventional tomosynthesis reconstruction has suffered contamination of the planes of interest by blurred out-of- plane structures. This paper focuses on a Matrix Inversion Tomosynthesis (MITS) algorithm to remove unwanted blur from adjacent planes. The algorithm uses a set of coupled equations to solve for the blurring function in each reconstructed plane. This paper demonstrates the use of the MITS algorithm in three imaging applications: small animal microscopy, chest radiography, and orthopedics. The results of the MITS reconstruction process demonstrate an improved reduction of blur from out-of-plane structures when compared to conventional tomosynthesis. We conclude that the MITS algorithm holds potential in a variety of applications to improve three-dimensional image reconstruction.

Typically in three-dimensional (3D) computed tomography (CT) imaging, hundreds or thousands of x-ray projection images are recorded. The image-collection time and patient dose required rule out conventional CT as a tool for screening mammography. We have developed a CT method that overcomes these limitations by using (1) a novel image collection geometry, (2) new digital electronic x-ray detector technology, and (3) modern image reconstruction procedures. The method, which we call Computed Planar Mammography (CPM), is made possible by the full-field, low-noise, high-resolution CCD-based detector design that we have previously developed. With this method, we need to record only a limited number (10 - 50) of low-dose x- ray images of the breast. The resulting 3D full breast image has a resolution in two orientations equal to the full detector resolution (47 microns), and a lower, variable resolution (0.5 - 10 mm) in the third orientation. This 3D reconstructed image can then be viewed as a series of cross- sectional layers, or planes, each at the full detector resolution. Features due to overlapping tissue, which could not be differentiated in a conventional mammogram, are separated into layers at different depths. We demonstrate the features and capabilities of this method by presenting reconstructed images of phantoms and mastectomy specimens. Finally, we discuss outstanding issues related to the further development of this procedure, as well as considerations for its clinical implementation.

The purpose of this study is to characterize a real time flat panel detector (FPD)-based imaging system for cone beam volume tomographic digital angiography (CBVTDA) applications. A prototype FPD-based imaging system has been designed and constructed on a modified GE 8800 CT scanner. This system is evaluated for CBVTDA using two phantoms. The system is first characterized in terms of linearity and dynamic range of the detector, the effect of image lag and scatter on the image quality, low contrast resolution and high contrast spatial resolution. The results indicate that the FPD-based imaging system can achieve 2lp/mm spatial resolution and provide appropriate low contrast resolution for intravenous CBVTDA angiography with clinically acceptable entrance exposure level.

Application of flat-panel imagers (FPIs) in cone-beam computed tomography (CBCT) offers a promising new modality for full three-dimensional (3-D) x-ray imaging. Understanding the potential performance and fundamental limitations of such technology, however, requires knowledge of the noise characteristics of the 3-D imaging system. The noise performance of a prototype flat-panel cone-beam CT (FPI-CBCT) system is investigated empirically and theoretically in terms of voxel noise, noise-power spectrum (NPS), and detective quantum efficiency (DQE). Methods for NPS analysis common in characterizing 2-D imagers are extended to the fully 3-D case, and a general framework for NPS analysis in n dimensions is presented, including the important considerations of NPS convergence and normalization within the constraints of system linearity and stationarity. Factors affecting imager noise and NPS are numerous, including exposure, number of views, image blur, additive noise, reconstruction filter, and sampling matrix. The applicability of existing theoretical descriptions of CT voxel noise is examined. A theoretical cascaded systems model that accurately predicts the 2-D noise characteristics of the FPI is extended to describe the signal and noise transfer characteristics of the fully 3-D FPI-CBCT system employing filtered back-projection. Analysis of the fully 3-D NPS reveals features of the 3-D imaging system that might otherwise be missed and shows the significant effect of 3-D noise aliasing. Furthermore, it quantifies the performance of various FPI designs in CBCT (e.g., direct and indirect detectors, and variations therein) and provides a guide for the development of high-performance FPI-CBCT systems suited to specific clinical objectives.

We report on a-Si direct detection x-ray image sensors with polycrystalline PbI₂, and more recently with HgI₂. The arrays have 100 micron pixel size and, we study those aspects of the detectors that mainly determine the DQE, such as sensitivity, effective fill factor, dark current noise, noise power spectrum, and x-ray absorption. Line spread function data show that in the PbI₂ arrays, most of the signal in the gap between pixels is collected, which is important for high,DQE. The leakage current noise agrees with the expected shot noise value with only a small enhancement at high bias voltages. The noise power spectrum under x-ray exposure is reported and compared to the spatial resolution information. The MTF is close to the ideal sinc function, but is reduced by the contribution of K-fluorescence in the PbI₂ film for which we provide new experimental evidence. The role of noise power aliasing in the DQE and the effect of slight image spreading are discussed. Initial studies of HgI₂ as the photoconductor material show very promising results with high x-ray sensitivity and low leakage current.

The x-ray flat-panel detector (FPD) will be a key component of the coming generation of x-ray imaging systems. FPD systems applicable to both fluoroscopy and radiography especially, will be the prime candidate to replace current image intensifier x-ray (IIXR-TV) systems. Nevertheless, IIXR-TV systems which have recently been improved by the addition of CCD cameras, have established themselves over time by offering good image quality which in most cases clinicians appear to be satisfied with. It will thus take a substantial improvement in image quality combined with a new ease of use due to reduced physical size for new FPDs to replace those systems that have evolved over many decades. Our group has been developing a selenium-based FPD which has superior spatial resolution characteristics. The purpose of this research is to elucidate the FPD's potential to replace IIXR-TV systems by offering improved image quality. Detailed measurements of physical characteristics were made and extensive in vivo animal studies were conducted. It can be concluded that the FPD's demonstrated superior image quality appears to have the potential to improve clinical performance.

The Imix radiography system (Qy Imix Ab, Finland)consists of an intensifying screen, optics, and a CCD camera. An upgrade of this system (Imix 2000) with a red-emitting screen and new optics has recently been released. The image quality of Imix (original version), Imix 200, and two storage-phosphor systems, Fuji FCR 9501 and Agfa ADC70 was evaluated in physical terms (DQE) and with visual grading of the visibility of anatomical structures in clinical images (141 kV). PA chest images of 50 healthy volunteers were evaluated by experienced radiologists. All images were evaluated on Siemens Simomed monitors, using the European Quality Criteria. The maximum DQE values for Imix, Imix 2000, Agfa and Fuji were 11%, 14%, 17% and 19%, respectively (141kV, 5μGy). Using the visual grading, the observers rated the systems in the following descending order. Fuji, Imix 2000, Agfa, and Imix. Thus, the upgrade to Imix 2000 resulted in higher DQE values and a significant improvement in clinical image quality. The visual grading agrees reasonably well with the DQE results; however, Imix 2000 received a better score than what could be expected from the DQE measurements. Keywords: CCD Technique, Chest Imaging, Digital Radiography, DQE, Image Quality, Visual Grading Analysis

Large area flat panel solid-state detectors are being studied for digital radiography and fluoroscopy. Such systems use active matrix arrays to readout latent charge images created either by direct conversion of x-ray energy to charge in a photoconductor or indirectly using a phosphor and individual photodiodes on the active matrix array. Our work has utilized the direct conversion method because of its simplicity and the higher resolution possible due to the electrostatic collection of secondary quanta. Aliasing of noise occurs in current designs of direct detectors based on amorphous selenium ((alpha) -Se) because of its high intrinsic resolution. This aliasing leads to a decrease in detective quantum efficiency (DQE) as frequency increases. It has been predicted, using a previously developed model of the complete imaging system, that appropriately controlled spatial filtration can reduce this aliased noise and hence increase DQE at the Nyquist frequency, f_NY. Our purpose is to experimentally verify this concept by implementing presampling filtration in a practical flat panel system. An (alpha) -Se based flat panel imager is modified by incorporating an insulating layer between the active matrix and the (alpha) -Se layer to introduce a predetermined amount of presampling burring. The modified imager is evaluated using standard linear analysis tools, modulation transfer function (MTF), noise power spectra (NPS) and DQE(f), and the results are compared to theoretical predictions.

Several x-ray parameters must be optimized to deliver exceptional fluoroscopic and radiographic x-ray Image Quality (IQ) for the large variety of clinical procedures and patient sizes performed on a cardiac/vascular x-ray system. The optimal choice varies as a function of the objective of the medical exam, the patient size, local regulatory requirements, and the operational range of the system. As a result, many distinct combinations are required to successfully operate the x-ray system and meet the clinical imaging requirements. Presented here, is a new, configurable and automatic method to perform x-ray technique and IQ optimization using an x-ray spectral model based simulation of the x-ray generation and detection system. This method incorporates many aspects/requirements of the clinical environment, and a complete description of the specific x-ray system. First, the algorithm requires specific inputs: clinically relevant performance objectives, system hardware configuration, and system operational range. Second, the optimization is performed for a Primary Optimization Strategy versus patient thickness, e.g. maximum contrast. Finally, in the case where there are multiple operating points, which meet the Primary Optimization Strategy, a Secondary Optimization Strategy, e.g. to minimize patient dose, is utilized to determine the final set of optimal x-ray techniques.

MTFs and focal spot intensity profiles were modeled as Gaussian functions. The overall resolution limits of the MTF were derived as a function of the magnification factor, detector resolution limit and focal spot size. The MTF and NPS for a small field digital mammography system was measured and used to compute NEQ for various magnification factors. Computation of DQE is discussed. Images of simulated microcalcification cluster were acquired and used to demonstrate the improvement of low contrast detectability in magnification imaging. It was shown that MTF improves with magnification when the detector MTF is low. The improvement decreases as the detector MTF increases. It was observed that at low to medium frequencies, the MTF improvement would be limited by the focal spot blurring effect while at high frequencies, the MTF improves with all magnification factors. The NEQ was found to improve with magnification factor. The microcalcification detectability was also found to improve as the magnification is increased.

Digital imaging systems require offset and gain calibration to normalize the behavior of individual pixels. This normalization corrects for imperfections in the system and also external variables that have effects on uniformity. Imaging metrics like Detective Quantum Efficiency (DQE) and Modulation Transfer Function (MTF) define how sensitivity and resolution are transferred through the system. Gain calibration can result in a loss of DQE due to the noise associated with its application. The typical technique to minimize this noise is to average several gain calibration procedures so that the introduced noise is minimized. This paper discuses the effects of gain calibration on DQE. It measures DQE as a function of the number of gain calibration procedures averaged and contrasts it with a novel technique that uses a single filtered gain calibration. It demonstrates that noise filter techniques, applied to a single gain calibration, regains the loss in DQE without any degradation in resolution. This paper also compares imaging performance of a system using a filtered gain map against a system that has many gain calibrations averaged. The technique is demonstrated using a Thin-Film-Transistor (TFT)-based large area medical imaging system.

Photon interaction cross sections for a few biological materials are calculated at 60 keV using non-relativistic form factors and the incoherent scattering function approximation. Compton, Rayleigh and total scattering cross sections were estimated in the momentum transfer region 0 to 5A^o-1 for simulation and compilation purposes. Rayleigh scattering cross sections for water are estimated using the molecular form factors of Morin at low photon energies. Computed tomographic images of a few soft materials are obtained using a tomographic system based on an image intensifier. It consists of a charged coupled device camera and an acquisition board. The charge coupled device and the acquisition board allows image processing, filtration and restoration. A reconstruction program, written in PASCAL is able to give the reconstruction matrix of the linear attenuation coefficients, and simulates the matrix and the related tomography. X-ray imaging is a well established technique of detecting strongly attenuating materials as distinguished from weakly attenuating materials. The image contrast between the weakly attenuating materials can be enhanced by optimum selection of the X-ray energy and improving the spatial resolution.

In this paper a new technique of time-domain 2-D fluorescence lifetime microscopic measurement is presented. A fluorescence lifetime imaging microscope system based on a synchroscan streak camera is developed, and a data processing method of this technique is described. Reference sample is used to test the system. Experimental results show that, when magnification of the microscope is 100X, temporal resolution of the system is 2ps and spatial resolution is 8 micrometer. When a finer aperture mask and a micro-shift-mechanical system are used, spatial resolution of less than 1 micrometer can be obtained.

We investigated the influence of the scattered x-rays on the object sharpness of radiographs for tube voltages of 50 to 100 kV. For the purpose, using a phantom of the polymethyl methacrylate (PMMA) with an aluminum plate of low contrast object, the edge images including the blurs due to the x-ray focal spot, the screen-film system, and the scattered x-rays were produced. However, the shape of the overall modulation transfer function (MTF) derived from the edge image is considered not to be influenced by the scattered x-rays. As a result of comparing the overall MTFs with the product of the MTFs of the x-ray focal spot and the screen-film system, we found that the scattered x-rays gave different influence on the contrast at each spatial frequency for each tube voltage. To investigate the influence of the scattered x-ray on the overall MTF in detail, considering the MTF of the x-ray focal spot, we developed a method for determining the primary and scatter components of the overall MTF, whose components are proportional to the MTFs of the screen-film system and the scattered x-rays, respectively. Using the method, it was found that (1) the overall MTF depends on the MTF of scattered x- rays as well as the MTFs of the focal spot and screen-film system; and (2) at frequencies near 0 mm^-1, the influence of the blur due to the scattered x-rays on the object sharpness was greater than that occurred in the intensifying screen and increased with the tube voltage.

Cone beam CT has a capability of 3-dimensional imaging of large volumes with isotropic resolutions, and has a potentiality of 4-dimensional imaging (dynamic volume imaging) because cone beam CT acquires a large volume data with one rotation of X-ray tube-detector pair. However, one of the potential drawbacks with cone beam CT is larger amount of scattered X-rays. These X-rays may enhance the noise in reconstructed images, and thus affect low contrast detectability. The aim of this work was to estimate scatter fractions and effects of scatter on image noise, and was to seek methods of improving image quality in cone beam CT. First we derived a relationship between the noises in reconstructed image and in X-ray intensity measurement. Then we estimated scatter to primary ratios in X-ray measurements, using a Monte Carlo simulation. From these we estimated image noise in clinical relevant conditions. The results showed that scatter radiation made a substantial contribution to the image noise. However focused collimators improved it, because they decreased scatter radiation drastically while keeping the primary radiation nearly the same level. A conventional grid also improved image noise though the improvement was less than that of focused collimators.

A microangiography system using monochromatized synchrotron radiation has been investigated as a diagnostic tool for circulatory disorders and early stage malignant tumors. The monochromatized X-rays with energies just above the contrast agent K-absorption edge energy can produce the highest contrast image of the contrast agent in small blood vessels. At SPring-8, digital microradiography with 6 - 24 micrometer pixel sizes has been carried out using two types of detectors designed for X-ray indirect and direct detection. The indirect-sensing detectors are fluorescent-screen optical-lens coupling systems using a high-sensitivity pickup-tube camera and a CCD camera. An X-ray image on the fluorescent screen is focused on the photoconductive layer of the pickup tube and the photosensitive area of the CCD by a small F number lens. The direct-sensing detector consists of an X-ray direct- sensing pickup tube with a beryllium faceplate for X-ray incidence to the photoconductive layer. Absorbed X-rays in the photoconductive layer are directly converted to photoelectrons and then signal charges are readout by electron beam scanning. The direct-sensing detector was expected to have higher spatial resolution in comparison with the indict-sensing detectors. Performance of the X-ray image detectors was examined at the bending magnet beamline BL20B2 using monochromatized X-ray at SPring-8. Image signals from the camera are converted into digital format by an analog-to- digital converter and stored in a frame memory with image format of 1024 X 1024 pixels. In preliminary experiments, tumor vessel specimens using barium contrast agent were prepared for taking static images. The growth pattern of tumor-induced vessels was clearly visualized. Heart muscle specimens were prepared for imaging of 3-dimensional microtomography using the fluorescent-screen CCD camera system. The complex structure of small blood vessels with diameters of 30 - 40 micrometer was visualized as a 3- dimensional CT image.

A new high spatial resolution micro-angiographic camera will enable routine viewing within a region of interest of detailed vascular structure unable to be seen with current full field of view (FOV) angiographic detectors. Such details include perforator vessels, vessel contractility or compliance, and condition and location of 50 micron or smaller stent wires. Although the basic CsI(Tl) phosphor-optical taper-CCD design of the new ROI micro-angiographic camera is essentially the same as that of the pre-clinical prototype, many of the physical parameters are much improved. The FOV is 5 cm X 5 cm vs. the previous 1 cm X 1 cm; the phosphor thickness is 350 - 400 micron vs. the previous 100 micron; the taper ratio is now 1.8 rather than 3.0 (2.8X improvement in light collection). The pixel size is either 25 or 50 micron. Additionally, detector noise may now be carefully considered in the camera design as may mechanical supporting mechanisms, methods to synchronize image acquisition with exposure and the effects of other physical factors such as exposure parameters, tube loading, focal spot size and geometric unsharpness. It is expected that this new capability should allow improved treatments and further development of smaller interventional devices and catheter delivery systems.

Cascade gamma-ray tomography depends on the measurement of two or more gamma-rays emitted from a radionuclide in coincidence. In using suitable detectors of appropriate timing resolution it may also be possible to obtain information about the 'chemical environment' of the radionuclide i.e. the binding site in the material to which it is attached by carrying out, simultaneously, time differential perturbed angular correlation (TDPAC) measurements. A dual block bismuth germanate (BGO) detector system, designed for use in positron emission tomography (PET) has been employed to perform cascade gamma-ray tomography with a point ⁶⁰Co source (1173 and 1332 keV). It is shown that despite the poor energy resolution of the coincidence system, the point source can be imaged with a multi-hole collimator, on one or both detectors, with a resolution (full width at half maximum) of 3.8 and 2.4 +/- 0.2 mm respectively, at the expense of significantly reduced sensitivity. Practical constraints limit usefulness of the system which new scintillation systems, with respect to time and energy resolution, should overcome.

Simulated mass lesions were superimposed onto an anthropomorphic breast phantom and x-rayed using a small field of view digital mammography system. Digital radiographs were acquired at a range of x-ray tube potential (constant detector exposure), and a range of x-ray tube current-time product values (constant x-ray tube potential). Twelve readers assessed the probability of a simulated mass being present in specified regions of interest, with the resultant data used to determine the area under the receiver operating characteristic curve (A_z). The mean A_z value obtained at 28 kVp/72 mAs was 0.69. At the same x-ray tube potential, the A_z score fell to 0.63 (p less than 0.05) at 32 mAs, whereas the mean A_z score of 0.71 was not significantly different for the image acquired at 144 mAs. At a constant detector exposure, reducing the x-ray tube potential to 22 kVp (320 mAs) resulted in a mean A_z value of 0.72, whereas increasing the x-ray tube potential to 34 kVp (28 mAs) resulted in a mean A_z value of 0.69. For the detection of simulated mass lesions in an anthropomorphic breast phantom, changing the kVP at a constant detector exposure had no significant effect on imaging performance, whereas halving the mAs at a constant kVp reduced the A_z value by approximately 10%.

A method has been developed to utilize a 3D B₀ fieldmap, with a multi-volume-of-interest segmentation map, to quantify and correct geometric distortions in echo-planar images. The purpose is to provide accurate co-registration of anatomical MRI to functional MRI time course sequences. A data structure capable of extracting and reporting the necessary information forms a central part of the solution. Images were obtained from a 1.5 Tesla scanner with an experimental y-gradient insert coil. Two 3D-gradient echo sequences supply the data needed to calculate the B₀ map across the volume. Segmentation of the volume into brain/background produces the data needed for the phase unwrapping and volume(s) of interest generation, from which the global B₀ variation map is obtained. Subsequent EPI acquisition yields the fMRI time- course information. Tests were carried out on a phantom and a human volunteer engaged in a motor task (finger-tapping). Strong distortions were measured, and subsequently corrected, particularly near the petrous bone/mastoid air cells and in the frontal and maxillary sinuses. Additionally, a strong eddy current resulting from the unshielded y-gradient was detected. The method facilitates geometric distortion correction through an imaging volume, containing multiple regions of interest within a slice, starting from a single starting point.

A 'variable resolution x-ray detector' (VRX) capable of resolving beyond 100 cycles/main a single dimension has been proposed by DiBianca, et al. The use of detectors of this design for computed-tomography (CT) imaging requires novel preprocessing of data to correct for the detector's non- uniform imaging characteristics over its range of view. This paper describes algorithms developed specifically to adjust VRX data for varying magnification, source-to-detector range and beam obliquity and to sharpen reconstructions by deconvolving the ray impulse function. The preprocessing also incorporates nonlinear interpolation of VRX raw data into canonical CT sinogram formats.

In order to evaluate and characterize the spatial resolution properties of a direct digital radiography (DDR) system based on a four CCD array detector, we measure the modulation transfer function (MTF) of the imaging system from its edge spread function (ESF). Different from the traditional algorithm for fine sampling of the ESF of an edge or a slit image with a slight angle to the image coordinate, in this paper we propose a new iterative algorithm. The novel algorithm finely samples the ESF along the column almost parallel to the edge, but not the row perpendicular to the edge. The samples of the first column and the second column are combined and averaged to get a segment of fine-sampled ESF. Then the segment of fine-sampled ESF is further combined with the next column and averaged to get a slight longer segment of ESF, and so forth. A complete fine-sampled ESF can be obtained after repeating the procedure for all columns of the edge image. An adaptive filtering is used for smoothing the ESF so that the sharpness at the edge of ESF can be protected. The MTF of the DDR is obtained from a series of image processing and data processing. The image processing includes the automatic determination of the angle of the edge to the pixel array of the edge image and fine sampling of edge spread function (ESF); data processing includes smooth filtering of ESF, numerical differential and fast Fourier transformation (FFT).

We developed a high resolution Imaging Plate (IP) with a transparent support and a reading system that can detect emissions from dual sides of the IP to improve the image quality of the CR mammography. And we proposed an addition method using the filtering processing in real space in order to achieve optimum addition in all the spatial frequencies. The image quality of the system was evaluated by NEQ and CD- MAM phantom. Furthermore we separated the noise component that consisted of X-ray photon noise, light photon noise, and the structural noise of the IP utilizing laser power dependency of the amount of emissions and the Wiener Spectrum. By using the above reading system and the addition method, NEQ of the system was improved by 40% - 50% compared to the latest CR mammography system. We confirmed by use of CD-MAM that the detectability of the image in this reading system was remarkably improved. The noise analysis showed that the ratios of the light photon noise at 1c/mm of the front side image and the back side image were about 15% and 40%, respectively, and according as the spatial frequency became higher the ratio of the light photon noise increased.

X-ray scattering in megavoltage portal imaging becomes more of an issue when quantitative results are needed. This is the case in megavoltage computed tomography (MVCT) and transit dosimetry, where the absorbed dose delivered to the patient is to be reconstructed. Although sensor arrays based on amorphous silicon (a-Si) photodiodes show promising results for this application, the scatter problem has so far not been examined. In this paper portal scatter distributions are calculated by means of Monte-Carlo (MC) simulations for typical clinical parameters. The aim of the MC simulations is to design a detector which is able to reject photons and electrons scattered by the phantom. As expected the analysis of the spectrum shows that multiply scattered photons can be differentiated from singly scattered photons by means of their energy. The MC results indicate that by using a detector with a high-Z conversion plate combined with a moderately thick phosphor screen a significant fraction of low energy scattered photons and most electrons can be rejected. However, to reduce the scatter signal further a software correction method based on a dedicated scatter model is still necessary.

The effects of fat content on single- and dual-energy CT measurements of bone mineral were quantified using a set of tissue-mimicking phantoms which more accurately represents the in-vivo situation than previous phantoms. The key to performing these measurements in CT is to have a mixture of tissue types within each image voxel, a condition which is not satisfied with standard phantoms. The phantoms used in these studies were solid materials which mimicked 17 different homogeneous mixtures of bone, muscle, and fat. The concept of creating phantoms to mimic different mixtures of these tissues is new. The materials are epoxy-resin based and have different mixtures of phenolic microspheres, polyethylene, and calcium carbonate suspended in them. Single- and dual-energy CT were used to image the phantom materials, and the effects of fat content on bone-mineral measurements were determined. The single-energy CT measurements show how fat content causes an underestimation of the amount of bone mineral present in a specimen, with the underestimation increasing as a function of fat content. With 25% and 50% fat by volume, the single-energy measurements underestimated bone volume percentage by 2.7% and 3.6% respectively. With dual-energy CT, fat content has no effect on the measurement of bone mineral. These results are not surprising. In fact, the effects of fat content on single- and dual-energy CT measurements have been studied many times previously. However, a system of accurately measuring these effects using a set of phantom measurements with physiologically accurate tissue-mimicking materials has not been developed previously. Using these phantoms, dual-energy CT measurements can be accurately calibrated for measurements of bone mineral while the errors possible while measuring bone mineral with single-energy CT can be quantified for any given imaging parameters.

The purpose of this paper is to analyze the image quality of a selenium-based flat panel detector suited for digital interventional mammography. To characterize the image quality, the DQE was measured at various x-ray exposures. The results indicate that when the detector is quantum noise limited, the DQE is independent of the exposure. A measurement of the quantum detection efficiency of 90% indicates that an electrostatic field shaping effect within the selenium layer gives a greater collection efficiency than that predicted simply by the geometric fill factor of each pixel collection electrode. Measurements were also conducted to determine the relative strength of ghost images on the detector. An image of a high contrast object using an exposure of 183 mR was acquired, followed by a low exposure 6 mR flat field image. No visual indication of a ghost could be found in the latter image even after appropriate windowing and leveling of the image was performed. A subjective comparison of image quality between film/screen and the detector was conducted by acquiring images of the ACR phantom under various exposure conditions. The digital images were printed on film using optimally adjusted LUT's. The resulting images were randomly presented to 15 non-trained observers, who assessed a score for each image. The comparison results show that the image quality obtained with the digital detector is superior to the images acquired with film/screen.

A distortion correction table compression method based on polynomial fitting has been developed for implementation in a commercial volume-CT system. To achieve the fastest processing rates, distortion correction tables must fit into the limited memory present in hardware. The number of elements in raw lookup tables is approximately 2 X N_i X N_j X N_(theta ), where N_i X N_i is the image dimensions in pixels, and N_(theta ) is the number of frames. Two- dimensional (2D) compression fits 4th-order polynomials to columns and rows of the raw table, reducing table size to 2 X 5 X 5 X N_f. Three-dimensional (3D) compression further compresses 2D tables in the angle dimension; reducing table size to 2 X 5 X 5 X 5. Tradeoffs between table size, accuracy, speed, and amount of distortion were investigated with data acquired from 7', 9', 10', 12', 14', and 16' IIs. The mean error was approximately 0.11, 0.20, and 0.20 pixels for raw table, 2D and 3D corrected data; with standard deviations of 0.08, 0.12, and 0.12 pixels.

The purpose of this work was to determine the effects of various incident x-ray beam spectra on the measured noise power spectrum (NPS) of a computed radiography (CR) image acquisition device. A CR phosphor was uniformly exposed to 1 mR with x-ray beams whose peak tube potentials were 70, 95, and 120 kVp that were filtered by various thicknesses of a 'patient equivalent phantom' (PEP; 2% aluminum, 98% acrylic by thickness), aluminum, and copper. From the uniform exposure images, NPS curves were calculated and their integral values were computed. The integral noise values were found to vary substantially as a function of x-ray beam spectral content. A simple x-ray beam and filter model that accounted for the shape of the filtered x-ray spectra and the mass energy absorption coefficient of the storage phosphor verified the qualitative behavior of the integral NPS values corresponding to changes in the incident x-ray beam used. The x-ray beam and filter combinations of 70 kVp and 10.1 cm of PEP filtration and 120 kVp and 20.2 cm of PEP filtration were chosen as standard techniques for evaluating clinical imaging systems. These two combinations represent a relatively low, clinically relevant CR noise (integral NPS equals 2.6 X 10^-6 mm) technique and a relatively high, clinically relevant CR noise (integral NPS equals 3.3 X 10^-6 mm) technique.

This communication presents a new method for measuring the thickness of thin membranes with pulse-echo ultrasound. The method is based on a consideration of the structure of the interference of the two echoes from the sides of the membrane. Essentially, it is the interference between these two echoes that gives rise to poor thickness estimation. The new method, which is explained in detail, proceeds from a consideration of the zeros of the echo signal in the complex frequency domain. Measurements with the new technique are compared with two other methods: the time-separation of echo envelopes, and a cross-correlation method. The analysis is presented with both simulated and real (laboratory) data. The effect of noise is taken into account in the laboratory data. This new method is shown to be capable of measuring sub-wavelength membrane thicknesses with excellent precision. The ultrasound rf signal is required, but a substantial improvement over existing techniques is gained.

Creatv MicroTech is developing two-dimensional, air-core, anti-scatter grids that have the potential to significantly reduce scatter-to-primary ratio and increase primary transmission in mammography. The fabrication method uses x-ray lithography and electroplating, which allows the fabrication of high aspect ratio metal parts. Two unfocused nickel grids were fabricated, one 1.5 cm X 1.5 cm and the other 1.44 cm X 1.44 cm. The grids have 20 micron thick walls and a period of 300 microns. Monte Carlo simulations were performed to predict their performance. The x-ray source was a 30 kVp Mo-anode spectrum and 30 microns of added Mo filtration. Preliminary calculations for a 2 mm-high grid and a 4 cm lucite phantom indicate that a scatter-to-primary ratio less than 3% can be achieved even at 3 cm from the center of the grid. Experiments to test the performance of the grids have been conducted at FDA using a Mo target, 30 micron Mo filter at 30 kVp and a 4 cm thick lucite phantom. A germanium detector was used. Data from a mammographic grid made by Smit Rontgen was taken as a reference. These Ni grids with grid ratios of 6.4 and 7.1 reduce scatter and increase primary transmission compared to the conventional reference grid. This fabrication method is capable of producing focused grids. The demonstration of larger, focused grids is the next step.

ROI imaging techniques can improve image quality and reduce radiation dose to patient and staff when the optimal combination of filter material and image receptor is used. An investigation was conducted to evaluate the effect of thickness of a-Se (0.5 mm and 1.0 mm) on image contrast, contrast to noise ratio (CNR), and figure of merit (FOM) with or without ROI filters (various thickness' of Gd and Cu) and to compare the results with the corresponding values obtained using a 0.4 mm thick cesium iodide (CsI) image receptor. Simulated x-ray spectra and published attenuation coefficients were used to calculate the x-ray transmission through a broad range of thickness' of various contrast materials for beams of 50 to 100 kVp. The results indicate that a-Se provides substantially better contrast compared to CsI for barium and iodinated contrast media for all cases, especially when the combination of Gd filter and the thinner a-Se is used. Moreover, during ROI procedures, the thicker a-Se generally increases image contrast, CNR, and FOM compared to CsI. Although, the thinner a-Se provides the highest image contrast for all cases, its combination with Cu results in lower CNR and FOM at higher kVp's compared to CsI.

This paper presents a new method for the measurement of modulation transfer function (MTF) using Wiener filtering. The method, unlike conventional methods using slit or edge devices, allows the direct determination of the MTF in all directions at one step. An image containing a precise circular region is acquired and its Fourier transform is calculated. In the absence of any blurring, the Fourier transform yields a simple Bessel function. Because of the symmetry in the convolution theorem, the roles of the blurring function and object can be interchanged, allowing the blurring function to be recovered using a Wiener filter. We simulated this process to understand the effects of attenuation, signal-to-noise ratio, and circle size. Images were simulated containing a circular region and degraded by spatial domain blurring with a Gaussian convolution kernel and by additive Poisson noise. The determined MTF matches the expected MTF except for a slight high frequency overestimate due to noise aliasing, which can be compensated. This method allows one to easily measure the two-dimensional MTF, particularly in systems which have an asymmetrical point spread function such as computed radiography. The method can be used as a tool for quality assurance and for comparing the resolution characteristics of various digital radiography systems.

The photorefractive polymer based holographic memories provide the alternative of achieving terabytes of digital data storage with gigabits per seconds of transfer. The stored information in the medium is in the form of electromagnetic distribution and it is governed by the second order inhomogeneous Helmholtz equations with no known exact solution. A numerical approximation using finite element method is employed to model the wave propagation in the polymer. The model is used to approximate the Helmholtz equation for a spatially varying index of refraction medium subject to input of polarized planar waves. The initial results from the mathematical formulation conform to the theoretical expectation of wave propagation on the non-conducting and non-magnetic materials. In addition, assuming periodicity of the inhomogeneous material, the reflective waves are computed and conservation of energy on the medium is verified. The model is used to simulate electromagnetic wave propagation through a two- dimensionally varying index of refraction poly (N- vynilcarbazol) (PVK) polymer. The transmitive waves are computed at various angles of incidents. The stored hologram represented as transmitive wave is reconstructed in reading process.

In MRI, the intensity distortion usually does not greatly affect the clinical relevance of the images, but its correction is an essential step for image segmentation and intensity measurements. In our MR scanner, the main sources of nonuniformity are the RF transmitting system, the receiver coil and the analogue receiver filter. The filter response was isolated and measured from images with no signal. The Homomorphic Unsharp Masking (HUM) was used to correct the RF related distortion. The HUM method is used to remove the image low-frequency components. In addition to its relatively long processing time (large windows are needed), the HUM method has some other disadvantages: it can affect the image contrast and generate artifacts at sharp edges. In order to overcome the limitations of the standard HUM approach, in this work, the correction maps were obtained from minimal contrast images (intermediate-echo images from a multiecho sequence). The correction maps obtained from these images were used to correct the corresponding high contrast images obtained from this sequence. This way, optimal results were obtained using small windows (9 X 9 - 13 X 13). The computation time was reduced and the image contrast was not affected. The background pixels were easier to handled and no artifacts were generated.