Alignment of tomographic projections using an incomplete set of fiducial markers

Reconstruction of three-dimensional images using tomography requires that the projections be aligned along a common rotational axis. We present here a solution to the problem of alignment for single-axis tomography using fiducial markers. The algorithm is based on iterative linearization of the projection equations and is least-squares-optimized by a linear least-squares solution instead of a gradient search. The algorithm does not require markers to be available in every projection, and initial estimates are unnecessary. Program execution is robust, fast, and can quickly align large data sets containing 256 or more projections.