Development of selective tomography algorithm for CT image reconstruction

A CT reconstruction algorithm to obtain selective tomography from multiple CT slices for localized 3D reconstruction is developed.

The proposed method uses Radon transform to obtain projections for a region of interest. Image reconstruction is performed using filtered back projection technique. The property that distinguishes the developed algorithm from previous algorithms is the ability to track the region of interest for every projection angle, computing projections along the lines intersecting the region of interest, and performing reconstruction using purely local data. Parallel beam projections are used and region of interest is defined by a square or circular region.

We conclude that using our algorithms, we can reconstruct the part of a cross section and thereby use less X-ray exposure for image reconstruction.