Performance evaluation of reconstruction with fan-beam algorithm with no backprojection weight in truncated projection completion using linear prediction

The work presented in this paper examines the performance of reconstruction from truncated projections using linear prediction for projection completion and recently proposed fan beam algorithm having no backprojection weight. The truncation in the tomographic projections can occur due to compact nature of the detector, outsized imaging object and region of interest (ROI) reconstruction applications. The reconstruction of the object function from truncated projections suffers from undesirable truncation artifacts and incorrect reconstructed density values. In this work, the truncated projection data is estimated by using the linear projection technique. The object is then reconstructed from the completed projections by using Hilbert filter based fan-beam algorithm with no backprojection weight. The resulting reconstruction improves the image quality in terms of noise performance characteristics. Simulation results are presented to show the improvements in image quality and compared with the reconstruction obtained using ramp filtered fan-beam algorithm.