Three-dimensional images with an improvement of resolution and contrast

摘要

A method for reconstructing a three-dimensional image of an object (I3D) defined by values taken by a property of the object (1) in a network of voxels of the three-dimensional image of the object, wherein measured projections (PM) of the property are determined, and the three-dimensional image of the object is iteratively reconstructed by (a) reconstructing a first three-dimensional image of the object; (b) estimating local parameters defining an anisotropic model (ML) of the object according to the reconstructed three-dimensional image; (c) deducing calculated projections (PC) of the object through at least one voxel; (d) comparing (E1) at least one calculated projection (PC) with the measured projection (PM) to deduce a projection difference; (e) determining a weight coefficient; (f) updating (E2) each voxel; and (g) repeating at least steps (c) through (f). The method is useful for medical imaging.