Optimal interpolation in helical scan 3D computerized tomography

A data acquisition scheme combining the axial translation of the object with rotation of the imaging gantry is introduced for three-dimensional computerized tomography (CT), and a reconstruction algorithm for its implementation is explored. A computationally attractive interpolation scheme that converts the data to standard CT format and permits reconstruction by existing algorithms is proposed. The approach is to convert by interpolation the helical-scan projection data into standard planar scan format and then use a standard 2-D reconstruction technique, such as convolution backprojection, on each slice. Optimum interpolation coefficients are derived, minimizing a worst case normalized maximum error magnitude criterion. The theoretical predictions are illustrated by simulation results.<
>