Three-dimensional imaging properties of rotation-free square and hexagonal micro-CT systems

We study the 3-D imaging properties of a rotation-free micro-computed tomography (CT) system with square and hexagonal geometries. These systems use linear arrays of carbon-nanotube-based X-ray sources that are individually addressable. The source arrays and area detectors in the square and the hexagonal geometries form the sides of a polygon. The tomographic angular sampling for both geometries requires no motion of the sources, the detectors, or the subject. We demonstrate that the hexagonal geometry has improved angular coverage as compared to the square geometry. The ordered-subset convex iterative algorithm is implemented in both geometries for reconstructions from cone-beam projection data. Simulation studies show that both geometries can be effectively reconstructed with polychromatic or monochromatic source spectra. As a result of the incomplete tomographic sampling of the two geometries, some streaking artifacts appear in the reconstructed images. The hexagonal geometry, in general, produces fewer streaking artifacts than the square geometry. On the other hand, the two geometries perform quite similarly in resolution-noise trade-off, so we conclude that the proposed geometries are comparably effective for the rotation-free micro-CT and the hexagonal geometry is superior in reducing streaking artifacts.