The resolving power of X -ray microtomography systems used for evaluating bone specimens.

Abstract

This thesis addresses the characterization of a 3-D microtomography system's overall resolution and those factors which contribute to system blur: focal spot blur, geometric instabilities, detector blur, and reconstruction blur. Using novel experimental and computational tools, the resolution response of three components of the full system blur are investigated for several clinical systems. Additionally, a term describing the overall system resolution response, the resolving power, is introduced and experimental methods for assessing this value are implemented. The first chapter of this thesis provides an in-depth review of the history of bone microtomography and closes with a discussion of the methods employed by researchers in characterizing the performance of their own systems, noting the absence of any standard method for determining total system performance. The second chapter then provides a summary of the traditional methods for characterizing resolution blur and introduces the concept of resolving power. In the third chapter, a method to characterize the resolution blur associated with a particular cone-beam reconstruction algorithm is reported. The fourth chapter describes the characterization of the resolution blur associated with the geometric instability of the components of a microtomography system. In the fifth chapter, an edge-based method to assess the resolution blur of digital detectors used for imaging soft x-rays is reported. The sixth chapter of the thesis reports on a method to experimentally measure the resolving power of microtomography systems. The method relies on obtaining projection views through a full rotation of an optical-quality ruby ball and the full reconstruction of the data under the system's standard operating conditions. The reconstructed volumetric data is scanned determining each voxel's distance from the center of the sphere and reprojecting that data into a one-dimensional vector representing the surface spread function. The plane spread function and the modulation transfer function representing the overall system resolution blur are derived from this surface spread function. Results from the application of this method to the evaluation of two systems are reported along with a discussion on the estimation of the system resolution response from the combination of the resolution blur measures performed on each of the system components.