Affiliations: College of Health Science Engineering, University of
Tennessee Health Science Center, Memphis, TN 38163, USA
Note: [] Corresponding author: Lawrence M. Jordan, Ph.D., College of
Health Science Engineering, University of Tennessee Health Science Center, 920
Madison Avenue, Room 1005, Memphis, TN, 38163, USA. Tel.: +1 901 448 7343; Fax:
+1 901 448 7387; E-mail: [email protected]
Abstract: Efforts to improve the spatial resolution of CT scanners have
focused mainly on reducing the source and detector element sizes, ignoring
losses from the size of the secondary-ionization charge "clouds" created by
the detected x-ray photons, i.e., the "physics limit." This paper focuses on
implementing a technique called "projective compression," which allows
further reduction in effective cell size while overcoming the physics limit as
well. Projective compression signifies detector geometries in which the
apparent cell size is smaller than the physical cell size, allowing large
resolution boosts. A realization of this technique has been developed with a
dual-arm "variable-resolution x-ray" (VRX) detector. Accurate values of the
geometrical parameters are needed to convert VRX outputs to formats ready for
optimal image reconstruction by standard CT techniques. The required
calibrating data are obtained by scanning a rotating pin and fitting a
theoretical parametric curve (using a multi-parameter minimization algorithm)
to the resulting pin sinogram. Excellent fits are obtained for both
detector-arm sections with an average (maximum) fit deviation of
∼0.05 (0.1) detector cell width. Fit convergence and
sensitivity to starting conditions are considered. Pre- and post-optimization
reconstructions of the alignment pin and a biological subject reconstruction
after calibration are shown.
