A nonlinear helical reconstruction algorithm for multi-slice CT

With the recent introduction of multi-slice CT scanners, image reconstruction techniques for multi-slice helical CT become an active research topic. The majority of the previously proposed algorithms limits the plane-of-reconstruction (POR) to a flat plane (either tilted or non-tilted). Additionally, linear interpolation is typically employed to estimate the projections at the POR from the measured conjugate samples. In this paper, we propose a new helical reconstruction algorithm for multi-slice helical CT. To avoid discontinuity in the helical weighting function, a double cone shaped region is selected as the POR. All available projection samples that are located within a predefined distance from the POR are utilized to estimate a set of projections at the POR. In addition, a weighted interpolation-extrapolation approach is proposed to replace linear interpolation. The proposed scheme has the advantage of improved slice sensitivity profile due to a better preservation of high-frequency contents present in the projection data set. It also exhibits better performance in terms of helical artifact suppression. Phantom experiments have been conducted to demonstrate the efficacy of our approach