| X-ray computed tomography (CT) imaging of small animals suffers from limited capability to resolve low-contrast targets due to the tradeoff between voxel size and image noise. As a result there has been great interest in employing energy-selective techniques (e.g., dual-energy CT) to extract additional information about the energy dependence of x-ray attenuation to improve target identification and differentiation. However, conventional energy-selective instrumentation has the drawbacks of lack of spectral separations and the need of multiple x-ray exposures.;The goal of this work was to investigate the potential advantages and challenges resulting from the use of novel energy resolved photon-counting x-ray detectors on imaging small animal sized targets. We first developed a set of simulation tools to model planar and tomography energy-selective x-ray imaging based on this detector. With the simulation tool we investigated the benefits of energy resolved photon-counting detectors in terms of basis material decomposition, energy-dependent contrast, and material separation. The results showed that the performance of energy resolved photon-counting was comparable, and often exceeded that of conventional energy-selective instrumentation. Next we developed a table-top small animal CT system based on an energy resolved photon-counting detector prototype, and performed experiments using mouse-sized phantoms and target materials resembling soft tissue, bone and contrast agents. Results confirmed that most findings in the simulation were observed on a physical system, while several limitations were also revealed that existed on the current detector prototype that could significantly degrade image quality and hamper the efficacy of the system. We investigated charge sharing and ring artifacts, two of the major limitations, and proposed solutions to model or/and compensate for these effects. Results indicated that the proposed charge sharing model was able to reproduce the spectral distortions observed on the physical system, and ring artifacts were significantly reduced by the proposed pre-acquisition tuning and post-acquisition correction algorithm. The results of this dissertation work demonstrate that energy resolved photon-counting detectors can improve the performance of small animal CT in terms of improved image contrast and better separation between different materials, but that advances in hardware and software correction algorithms are needed to fully exploit their capabilities. |
