Direct two-dimensional reconstruction algorithm for an inverse-geometry CT system

Abstract

An inverse-geometry computed tomography (IGCT) system uses a large source array opposite a smaller detector array. A previously described IGCT reconstruction algorithm uses gridding, but this gridding step produces blurring in the reconstructed image. In this article, the authors describe a two-dimensional (2D) IGCT reconstruction algorithm without gridding. In the transverse direction, the raw data of the IGCT system can be viewed as being composed of many fan beams. Because the spacing between source spots is larger than the spot width, each fan beam has undersampled projection data, but the missing samples are effectively provided by other undersampled fan beam views. In the proposed method, a direct fan beam reconstruction algorithm is used to process each undersampled fan beam. Initial images with this method showed ring artifacts caused by nonuniform sampling in the radial direction as compared to an ideal fan beam. A new method for correcting this effect was developed. With this correction, high quality images were obtained. The noise performance of the proposed 2D IGCT reconstruction algorithm was investigated, and it was comparable to that of the fan beam system. A MTF study showed that the proposed method achieves better resolution than the gridding method.