Local Tomography and the Motion Estimation Problem

In this paper we study local tomography (LT) in the motion contaminated case. It is shown that microlocally, away from some critical directions, LT is equivalent to a pseudodifferential operator of order one. LT also produces nonlocal artifacts that are of the same strength as useful singularities. If motion is not accurately known, singularities inside the object $f$ being scanned spread in different directions. A single edge can become a double edge. In such a case the image of $f$ looks cluttered. Based on this observation we propose an algorithm for motion estimation. We propose an empiric measure of image clutter, which we call edge entropy. By minimizing edge entropy we find the motion model. The algorithm is quite flexible and is also used for solving the misalignment correction problem. The results of numerical experiments on motion estimation and misalignment correction are very encouraging.