This PDF file contains the front matter associated with SPIE Proceedings Volume 9594 including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

To obtain two tomograms with two different photon energy ranges simultaneously, we have performed dual-energy Xray photon counting using a cadmium telluride (CdTe) detector, two comparators, two frequency-voltage converters (FVCs), and an analog digital converter (ADC). X-ray photons are detected using the CdTe detector with an energy resolution of 1% at 122 keV, and the event pulses from a shaping amplifier are sent to two comparators simultaneously to regulate two thresholds of photon energy. The logical pulses from a comparator are sent to an FVC consisting of two integrators, a microcomputer, and a voltage-voltage amplifier. The smoothed outputs from the two FVCs are input to the ADC to carry out dual-energy imaging. To observe contrast variations with changes in threshold energy, we performed energy-dispersive computed tomography utilizing the dual-energy photon counting at a tube voltage of 100 kV and a current of 8.7 μA. Two tomograms were obtained simultaneously at two energy ranges of 34.0-50.2 keV and 50.2-100 keV. The photon-count subtraction was carried out using a computer program. The maximum count rate was 5.4 kilocounts per second with energies of 10.0-100 keV, and the exposure time for tomography was 10 min.

Although there have been great strides in object recognition with optical images (photographs), there has been comparatively little research into object recognition for X-ray radiographs. Our exploratory work contributes to this area by creating an object recognition system designed to recognize components from a related database of radiographs. Object recognition for radiographs must be approached differently than for optical images, because radiographs have much less color-based information to distinguish objects, and they exhibit transmission overlap that alters perceived object shapes. The dataset used in this work contained more than 55,000 intermixed radiographs and photographs, all in a compressed JPEG form and with multiple ways of describing pixel information. For this work, a robust and efficient system is needed to combat problems presented by properties of the X-ray imaging modality, the large size of the given database, and the quality of the images contained in said database. We have explored various pre-processing techniques to clean the cluttered and low-quality images in the database, and we have developed our object recognition system by combining multiple object detection and feature extraction methods. We present the preliminary results of the still-evolving hybrid object recognition system.

Quasi-monochromatic photon counting was performed using a cadmium telluride detector and an energy-selecting device, consisting of two comparators and a microcomputer (MC). The two threshold energies are determined using low and high-energy comparators, respectively. The MC produces a single logical pulse when only a logical pulse from a low-energy comparator is input to the MC. Next, the MC never produces the pulse when two pulses from low and high-energy comparators are input to the MC, simultaneously. The logical pulses from the MC are input to a frequency-voltage converter (FVC) to convert count rates into voltages; the rate is proportional to the voltage. The output voltage from the FVC is sent to a personal computer through an analog-digital converter to reconstruct tomograms. The X-ray projection curves for tomography are obtained by repeated linear scans and rotations of the object at a tube voltage of 70 kV and a current of 12 μA. Iodine (I) K-edge CT was performed using contrast media and X-ray photons with a count rate of 2.2 kilocounts per second and energies ranging from 34 to 50 keV, since these photons with energies beyond I-K-edge energy 33.2 keV are absorbed effectively by I atoms.

We report on the development of silicon strip detectors for energy-resolved clinical mammography. Typically, X-ray integrating detectors based on scintillating cesium iodide CsI(Tl) or amorphous selenium (a-Se) are used in most commercial systems. Recently, mammography instrumentation has been introduced based on photon counting Si strip detectors. The required performance for mammography in terms of the output count rate, spatial resolution, and dynamic range must be obtained with sufficient field of view for the application, thus requiring the tiling of pixel arrays and particular scanning techniques. Room temperature Si strip detector, operating as direct conversion x-ray sensors, can provide the required speed when connected to application specific integrated circuits (ASICs) operating at fast peaking times with multiple fixed thresholds per pixel, provided that the sensors are designed for rapid signal formation across the X-ray energy ranges of the application. We present our methods and results from the optimization of Si-strip detectors for contrast enhanced spectral mammography. We describe the method being developed for quantifying iodine contrast using the energy-resolved detector with fixed thresholds. We demonstrate the feasibility of the method by scanning an iodine phantom with clinically relevant contrast levels.

The adaptive single-photon emission computed tomography (SPECT) system studied here acquires an initial scout image to obtain preliminary information about the object. Then the configuration is adjusted by selecting the size of the pinhole and the magnification that optimize system performance on an ensemble of virtual objects generated to be consistent with the scout data. In this study the object is a lumpy background that contains a Gaussian signal with a variable width and amplitude. The virtual objects in the ensemble are imaged by all of the available configurations and the subsequent images are evaluated with the scanning linear estimator to obtain an estimate of the signal width and amplitude. The ensemble mean squared error (EMSE) on the virtual ensemble between the estimated and the true parameters serves as the performance figure of merit for selecting the optimum configuration. The results indicate that variability in the original object background, noise and signal parameters leads to a specific optimum configuration in each case. A statistical study carried out for a number of objects show that the adaptive system on average performs better than its nonadaptive counterpart.

In PET imaging, the information needed to form an image is obtained from the detection of pairs of gamma-ray photons emitted by electron-positron annihilations. An optimal timing resolution allows the system to include time-of-flight (TOF) information, which improves image quality. The two methods to approach timing estimation are analog processing and digital captured waveform analysis. In digital analysis, there is a trade-off between the amount of data acquired and the timing resolution of a detector. In order to develop an efficient data acquisition system, we want to minimize the number of digital samples by acquiring the samples that contains the most information for timing estimation. We developed a simulation package to perform Fisher information analysis on the waveform samples in order to quantify the timing information conveyed by segments of the waveform. The diagonal components of the inverse of the Fisher information matrix set the bound that establishes the Cramér-Rao inequality on the variance of an unbiased estimator. The Maximum-Likelihood (ML) estimator is unbiased and asymptotically achieves the Cramér-Rao lower bound; for this reason, the ML estimator is ideal for performing timing estimation and extracting information as described by Fisher information analysis. This document explains the simulation of the waveforms, ML estimation method, Fisher information analysis and the calculation of the Cramér-Rao lower bound, for different lengths of the pulse. The results show that the timing resolution approaches a limit using just a segment of the waveform and there are parts of the pulse that are redundant information. The yields of this work will be used to build an efficient data acquisition (DAQ) system that will acquire less amount of data, and therefore, the complexity and cost of the DAQ system will be reduced.

Cerium activated mixed lutetium/gadolinium- and aluminum/gallium-based garnets have great potential as host scintillators for medical imaging applications. (Gd,Lu)₃(Al,Ga)₅O₁₂:Ce and denoted as GLuGAG feature high effective atomic number and good light yield, which make it particularly attractive for Positron Emission Tomography (PET) and other γ-ray detection applications. For PET application, rapid decay and good timing resolution are extremely important. Most Ce-doped mixed garnet materials such as GLuGAG:Ce, have their main decay component at around 80 ns. However, it has been reported that the decays of some single crystal scintillators (e.g., LSO and GGAG) can be effectively accelerated by codoping with selected additives such as Ca, Mg and B. In this study, transparent polycrystalline (Gd,Lu)₃(Al,Ga)₅O₁₂:Ce ceramics codoped with Ca or Mg or additional Ce, were fabricated by the sinter-HIP approach. It was found the transmission of the ceramics are closely related to the microstructure of the ceramics. As the co-dopant levels increase, 2^nd phase occurs in the ceramic and thus transparency of the ceramic decreases. Ca and Mg co-doping in GLuGAG:Ce ceramic effectively accelerate decays of GLuGAG:Ce ceramics at a cost of light output. However, additional Ce doping in the GLuGAG:Ce has no benefit on improving decay time but, on the other hand, reduces transmission, light output. The mechanism under the different scintillation behaviors with Mg, Ca and Ce dopants are discussed. The results suggest that decay time of GLuGAG:Ce ceramics can be effectively tailored by co-doping GLuGAG:Ce ceramic with Mg and Ca for applications with optimal timing resolution.

Scintillators are important functional parts in x-ray and Υ-radiation medical imaging instruments, while the high refractive index of scintillation materials significantly reduced the light yield from the scintillators to the detectors, which limited acquired image quality. In this paper, we reviewed two ways to improve the light yield of scintillators via nano photonic devices based on different scintillation materials and integrated nano structures.

We have developed microstructured Lu₂O₃:Eu scintillator films that provide spatial resolution on the order of micrometers for hard X-ray imaging. In addition to their outstanding resolution, Lu2O3:Eu films also exhibits both high absorption efficiency for 20 to 100 keV X-rays, and bright 610 nm emission whose intensity rivals that of the brightest known scintillators. At present, high spatial resolution of such a magnitude is achieved using ultra-thin scintillators measuring only about 1 to 5 μm in thickness, which limits absorption efficiency to ~3% for 12 keV X-rays and less than 0.1% for 20 to 100 keV X-rays; this results in excessive measurement time and exposure to the specimen. But the absorption efficiency of Lu2O3:Eu (99.9% @12 keV and 30% @ 70 keV) is much greater, significantly decreasing measurement time and radiation exposure. Our Lu₂O₃:Eu scintillator material, fabricated by our electron-beam physical vapor deposition (EB-PVD) process, combines superior density of 9.5 g/cm³, a microcolumnar structure for higher spatial resolution, and a bright emission (48000 photons/MeV) whose wavelength is an ideal match for the underlying CCD detector array. We grew thin films of this material on a variety of matching substrates, measuring some 5–10μm in thickness and covering areas up to 1 x 1 cm², which can be a suitable basis for microtomography, digital radiography as well as CT and hard X-ray Micro-Tomography (XMT). The microstructure and optical transparency of such screens was optimized, and their imaging performance was evaluated in the Argonne National Laboratory’s Advanced Photon Source. Spatial resolution and efficiency were also characterized.

Dual-energy photon counting was performed using an energy-selecting device (ESD) and a detector, consisting of a Lu₂(SiO₄)O [LSO)] crystal and a multipixel photon counter (MPPC). The ESD is used to determine a low-energychannel range for CT and consists of two comparators and a microcomputer (MC). The two threshold channels in proportion to energies are determined using low and high-energy comparators, respectively. The MC in the ESD produces a single logical pulse when only a logical pulse from the low-energy comparator is input to the MC. To determine the high-energy-channel range for CT, logical pulses from the high-energy comparator are input to the MC outside the ESD. Logical pulses from the two MCs are input to frequency-voltage converters (FVCs) to convert count rates into voltages. The output voltages from the two FVCs are sent to a personal computer through an analog-digital converter to reconstruct tomograms. Dual-energy computed tomography was accomplished at a tube voltage of 70 kV and a maximum count rate of 14.3 kilocounts per second, and two-different-energy tomograms were obtained simultaneously.

Scientific computing is rapidly advancing due to the introduction of powerful new computing hardware, such as graphics processing units (GPUs). Affordable thanks to mass production, GPU processors enable the transition to efficient parallel computing by bringing the performance of a supercomputer to a workstation. We elaborate on some of the capabilities and benefits that GPU technology offers to the field of biomedical imaging. As practical examples, we consider a GPU algorithm for the estimation of position of interaction from photomultiplier (PMT) tube data, as well as a GPU implementation of the MLEM algorithm for iterative image reconstruction.

This position paper describes a potential implementation of a large-scale grating-based X-ray Phase Contrast Imaging System (XPCI) simulation tool along with the associated challenges in its implementation. This work proposes an implementation based off of an implementation by Peterzol et. al. where each grating is treated as an object imaged in the field-of-view. Two main challenges exist; the first, is the required sampling and information management in object space due to the micron-scale periods of each grating propagating over significant distances. The second is maintaining algorithmic numerical stability for imaging systems relevant to industrial applications. We present preliminary results for a numerical stability study using a simplified algorithm that performs Talbot imaging in a big-data context

The difference of spectral distribution between lesions of epithelial cells and normal cells after excited fluorescence is one of methods for the cancer diagnosis. In our previous work, we developed a portable LED Induced autofluorescence (LIAF) imager contained the multiple wavelength of LED excitation light and multiple filters to capture ex-vivo oral tissue autofluorescence images. Our portable system for detection of oral cancer has a probe in front of the lens for fixing the object distance. The shape of the probe is cone, and it is not convenient for doctor to capture the oral image under an appropriate view angle in front of the probe. Therefore, a probe of L shape containing a mirror is proposed for doctors to capture the images with the right angles, and the subjects do not need to open their mouse constrainedly. Besides, a glass plate is placed in probe to prevent the liquid entering in the body, but the light reflected from the glass plate directly causes the light spots inside the images. We set the glass plate in front of LED to avoiding the light spots. When the distance between the glasses plate and the LED model plane is less than the critical value, then we can prevent the light spots caused from the glasses plate. The experiments show that the image captured with the new probe that the glasses plate placed in the back-end of the probe has no light spots inside the image.