Jason M. Jaworski

Maximum-Likelihood Deconvolution in the Spatial and Spatial-Energy Domain for Events With Any Number of Interactions

In previous works, maximum-likelihood expectation-maximization deconvolution for two-interaction events within a single CdZnTe detector with dimensions of was implemented. This deconvolution method is capable of estimating the source image for each energy range as well as the incident spectrum for each direction around the detector. To improve the detection efficiency and the image resolution, we have built a four-detector array system; each detector has dimensions of . Using this detector-array system, from a Co-60 measurement, 41.5% of recorded events in the energy window from 1100 keV to 1200 keV are two-interaction events. The goal of this work is to increase the efficiency of this deconvolution algorithm by extending the calculation of the system response functions to events with other number of interactions. We first analytically extend the system response function calculation to three-interaction events by deriving the probability density function, considering the measurement noise, and integrating over the digitization and pixelation volume. The system response function is then simplified, modularized, and extrapolated to events with other numbers of interactions from an array system. By including events with any number of interactions in the system model, imaging makes use of all recorded events, and the angular resolution is improved. This deconvolution algorithm is applicable to any gamma-ray detector system that has the capability of recording 3D interaction location and energy deposition for each interaction.

Results From Testing of 145 3D Position-Sensitive, Pixelated CdZnTe Detectors

Testing of 20 × 20 × 15 mm3 3D position sensitive, CdZnTe detectors grown by Redlen Technologies Inc. has shown that 98 out of the 145 detectors analyzed achieved sub-1% FWHM at 662 keV for single-pixel events. Improvements in the detector performance over time reflect the improvements in detector growth and fabrication over the past several years. In addition to the spectroscopic performance of the detectors, the imaging performance of each detector was also quantified. The results show that imaging is weakly correlated to the spectroscopic performance, and more strongly correlated to the accuracy of the depth reconstruction. Other detector issues such as gain deficit and gain variation are also discussed.

Model-based reconstruction of spectral and spatial source distribution from objects with known motion

Radiation imaging has many applications ranging from health care to homeland security and defense, and source motion is present in many of these applications. When the motion profile of the source is known or otherwise estimated, one can use motion-compensation techniques to reduce blur in the reconstructed image. In this paper, we present a model-based source-intensity reconstruction in the energy and spatial domains using list-mode data. The model includes separate parameterization for objects moving with known motion that is independent of the stationary backdrop. This approach corrects for object motion without smearing stationary sources in the backdrop space. The goal is to simultaneously obtain an estimate of the incident energy and spatial distribution of the radiation field for the stationary backdrop and for each moving object. Experimental Compton-imaging results using an 18-detector array of 3-D-position-sensitive CdZnTe detectors show that the method can successfully reconstruct the source intensity of moving objects while also revealing stationary sources in the backdrop. Also, by modeling the possibility of partial photon energy deposition in the detector, the incident energy spectrum is reconstructed more accurately.

Experimental demonstration of coded aperture imaging using thick 3D-position-sensitive CdZnTe detectors

3D-position-sensitive CdZnTe semiconductor detectors have demonstrated 4π Compton imaging capability and excellent energy resolution at room-temperature operation. However, Compton gamma-ray imaging is not feasible at low energies due to the small Compton-scatter cross-section. This work extends the current imaging capabilities to lower energies by utilizing coded aperture masks. Multiple coded aperture masks are applied to a single detector system of four 20mm×20mm×15mm CdZnTe detectors. Near-4n coded aperture imaging has been demonstrated through Monte Carlo simulation. The correct source direction is consistently identified using measured data with one mask above the cathode side and another mask above the non-cathode side of the detector. Challenges related to electric field distortion due to space charge in the detector are discussed. The focus of this research is to image near-4n field of view using coded apertures, ultimately, combining both Compton and coded aperture imaging techniques to expand the range of gamma-ray imaging.

UMImaging: A Software Package for Image Reconstruction From 3D-Position-Sensitive Gamma-Ray Detectors

In order to support analysis of data from 3D-position-sensitive gamma-ray detectors, a software package, UMImaging, has been designed and implemented. UMImaging is designed to apply to a wide range of hardware configurations and analysis needs. Its framework centers around reconstruction methods that perform analysis including spectroscopy, imaging, and detection. These methods are supported by classes that supply events, describe geometry, set options, store and manipulate results, and interact with the user. The ability to bin and display highly multidimensional data is presented, drawing numerous examples from the analysis of data from a CdZnTe array system. Parallel computing integrated into the structure of UMImaging is shown to provide almost a seven-fold speedup on a six-core computer with hyper-threading, and the parallel implementation of iterative reconstruction is described. Design choices that make UMImaging platform independent and easily extendable are also discussed.

3D Compton image reconstruction using a moving 3D-position-sensitive room-temperature CdZnTe detector array

The ability to do real-time 4π Compton imaging is potentially useful when searching for radioactive sources with appropriate gamma-ray energy emissions. Stationary 4π Compton imagers have the ability to reconstruct the correct direction to gamma-ray point sources; however, if the source is positioned far from the detector compared to the dimensions of the detector, the distance from the source to the detector is uncertain. If the same detector moves and the position of the detector as a function of time is known, a 3D radiation image can be made. This work describes the use of room-temperature 3D-position-sensing CdZnTe detectors in motion to create a 3D radiation map of sources in an unknown environment. Experimental results show that an array composed of 18 CdZnTe detectors, can accurately map out a room of sources. Finally, a pair of fisheye cameras are used to correlate the reconstructed hotspot in 3D Cartesian space to a physical location in the optical image.

SU-C-201-03: Coded Aperture Gamma-Ray Imaging Using Pixelated Semiconductor Detectors

Purpose: Improved localization of gamma-ray emissions from radiotracers is essential to the progress of nuclear medicine. Polaris is a portable, room-temperature operated gamma-ray imaging spectrometer composed of two 3×3 arrays of thick CdZnTe (CZT) detectors, which detect gammas between 30keV and 3MeV with energy resolution of <1% FWHM at 662keV. Compton imaging is used to map out source distributions in 4-pi space; however, is only effective above 300keV where Compton scatter is dominant. This work extends imaging to photoelectric energies (<300keV) using coded aperture imaging (CAI), which is essential for localization of Tc-99m (140keV). Methods: CAI, similar to the pinhole camera, relies on an attenuating mask, with open/closed elements, placed between the source and position-sensitive detectors. Partial attenuation of the source results in a “shadow” or count distribution that closely matches a portion of the mask pattern. Ideally, each source direction corresponds to a unique count distribution. Using backprojection reconstruction, the source direction is determined within the field of view. The knowledge of 3D position of interaction results in improved image quality. Results: Using a single array of detectors, a coded aperture mask, and multiple Co-57 (122keV) point sources, image reconstruction is performed in real-time, on an event-by-event basis, resulting in images with an angular resolution of ∼6 degrees. Although material nonuniformities contribute to image degradation, the superposition of images from individual detectors results in improved SNR. CAI was integrated with Compton imaging for a seamless transition between energy regimes. Conclusion: For the first time, CAI has been applied to thick, 3D position sensitive CZT detectors. Real-time, combined CAI and Compton imaging is performed using two 3×3 detector arrays, resulting in a source distribution in space. This system has been commercialized by H3D, Inc. and is being acquired for various applications worldwide, including proton therapy imaging R&D.

Polaris-H measurements and performance

Polaris-H, a compact Compton camera designed by H3D, Inc. has been characterized and tested in field measurements at nuclear power plants. Polaris-H integrates a 3D-position-sensitive pixelated CZT detector (20 mm × 20 mm × 15 mm), associated readout electronics, an embedded computer, a 5-hour battery, and an optical camera all within a portable water-resistant case. The total mass is about 4 kg. Start-up time is 2 minutes. Additionally, there is a connection for a tablet, which displays a real-time gamma-ray spectrum (near 1% FWHM at 662 keV) and isotope-specific images of the radiation distribution in all 4π in real time. The newest design of Polaris-H features the capability to control data acquisition without using a tablet interface, a higher energy dynamic range, and a higher maximum count rate. Absolute efficiency, imaging efficiency, energy resolution, and detection and identification performance are presented. Measurements are shown for applications of decontamination and detection of previously unknown contamination regions.

Source motion compensated coded aperture imaging using thick 3D-position-sensitive CdZnTe detectors

The 18 - detector array system consists of eighteen 20mm × 20mm × 15mm CdZnTe detectors. In addition to spectroscopic capabilities, it has the ability to image gamma-ray sources with energies above 200keV in 4-pi space via Compton imaging. To extend the imaging capabilities to lower energies, two coded aperture masks have been applied to the system. In a practical scenario, the source could be moving in relation to the detector system. The resultant coded aperture images formed will be blurred and distorted unless this source movement is considered in the system model. This motion can be used to our advantage in the case of an imperfect detector and mask configuration with image artifacts. By averaging the images from multiple positions, the image intensity in the source direction will be unaffected, while the artifacts will be reduced significantly. In theory, this motion compensated image will not only reduce the fluctuations due to non-uniformity of the detector material properties, but will also flatten the image side-lobes caused by the random mask design, thus improving the signal-to-noise ratio of the source peak to background in the reconstructed image. The focus of this research is to validate this claim through measurement and simulation, and to ultimately automate this process given the coordinates and direction of the source.

Applications of the energy-imaging integrated deconvolution algorithm for source chatracterization

The energy-imaging integrated deconvolution (EIID) algorithm is capable of deconvolving the source image at any specific energy, as well as the incident spectrum from any direction. This spectrum-image reconstruction method takes place in integrated spatial and energy domain using the maximum likelihood expectation maximization (MLEM) algorithm. This technique has been demonstrated for combined two-, three- and four-interaction Compton events from an array of 2 cm × 2 cm × 1.5 cm detectors in our previous work. By including Compton continuum events in the system model and using a sensitivity image to correct efficiency effects, the MLEM solution estimates the true incident gamma-ray spectrum with correct branching ratio. For each direction, we can then identify the source isotope based on the peak energy and relative peak area. After identifying the source isotope, the source activity can be estimated. From the reconstructed image and change of the relative peak areas from the branching ratio, the presence of shielding and the material and thickness of shielding can be estimated.