Three-dimensional regularized binary image reconstruction from three two-dimensional projections using a randomized ICM algorithm
✍ Scribed by Florent Retraint; Françoise Peyrin; Jean Marc Dinten
- Publisher
- John Wiley and Sons
- Year
- 1998
- Tongue
- English
- Weight
- 341 KB
- Volume
- 9
- Category
- Article
- ISSN
- 0899-9457
No coin nor oath required. For personal study only.
✦ Synopsis
In this article, we address the problem of fully threethe reconstruction algorithm must be reconsidered, since the indimensional (3D) binary image reconstruction from three projections.
version problem is underdetermined. Such an ill-posed problem
The regularization approach relies on the use of a 3D Ising model may be regularized by limiting the solution space to a reduced set.
which is a particular Gibbs prior. The problem is then equivalent to For instance, constraining the solution to be binary is a possibility.
the minimization of a nonconvex energy function on discrete values.
The reconstruction of binary objects arises in many applica-
After an evaluation of standard optimization methods such as simutions. For instance, in medicine, the problem of reconstructing a lated annealing or its deterministic version ICM, we propose a new vascular tree from a small number of radiographs using a convenalgorithm called parallel randomized ICM. It improves the reconstructional angiography device is still of interest. The acquisition is tion quality compared to ICM solutions while keeping reasonable reconstruction time. Its performances are evaluated from simulated pro-performed before and after injecting a contrast agent in the arterjections for different 3D test images. Its application to real data is ies, which may be assumed to provide a homogeneous medium.