Fan-Beam Reconstruction Under Motion and Data Truncation: Comparing Analytic and Iterative Approaches
Résumé
In this paper we propose comparisons and correlations between analytic and iterative fan-beam reconstruction approaches when object's rigid motion and data truncation occur during a circular scan. Based on our recent work presenting an exact analytic reconstruction method, we are able to predict the field of theoretically reconstructible points and transform the problem from a dynamic to a static point a view where the source trajectory is virtually modified taking into account the known rigid motion. We implement the iterative reconstruction as the convex minimization of a data-fidelity term under non-negativity constraint and regularization to solve this virtually static inverse problem. We compare the reconstructed field of view by the two methods on 2D fan-beam Shepp-Logan phantom simulations. Our results show that both methods validate the predicted reconstructible zone and are in good correlation in terms of reconstruction quality. The iterative reconstruction also demonstrates that in certain cases it is possible to recover structures beyond the analytic strict frontier of reconstructibilty.
Origine | Fichiers éditeurs autorisés sur une archive ouverte |
---|