Sparse-view Image Reconstruction in Prospectively Gated Micro-CT for Fast and Low-dose Imaging

Abstract

We conducted a feasibility study using a total-variation minimization algorithm for image reconstruction in prospectively gated micro computed tomography (micro-CT). The total-variation (TV) minimization algorithm exploits the sparseness of the image's gradient magnitude and can successfully reconstruct CT images from undersampled data for which conventional analytic reconstruction algorithms fail. We implemented the algorithm and applied it to sparsely-sampled data for a mouse by using a prospectively gated micro-CT system. The images were successfully reconstructed, and an image similarity index was quantitatively calculated with respect to the reference images reconstructed from fully-sampled data. Compared to a conventional image reconstruction algorithm, the TV-minimization algorithm substantially reduced image inaccuracy related to the image artifacts.