Initial Experience in Image Reconstruction from Limited-angle C-arm CBCT Data

C-arm cone-beam CT (CBCT) often acquires data over a limited angular range. Image reconstruction from limited-angle data (i.e., data less than a short scan) becomes mathematically an ill-posed problem, and the inversion of the X-ray transform by use of conventional analytic formulas can introduce significant artifacts. In this work, using the adaptive steepest descent and projection-onto-convex-sets (ASD-POCS) algorithm, we investigate image reconstructions from limited-angle simulation cone-beam data and real patient data acquired with a C-arm system. The results show that the algorithm can reconstruct potentially useful images from data collected over limited angular range.