Surface reconstruction from tomography data

Abstract

We propose a method to reconstruct the surface of a three-dimensional (3D) object from its volume. We use an active surface model which is a 3D generalization of the basic two-dimensional (2D) model. The model is an elastic surface that is deformed under the action of internal and external forces. The internal forces model the smoothness constraints while the external forces model the image constraints. In our case, the image constraints are 3D detected edges obtained using the Zucker and Hummel (1981) operator. The finite difference method is used to solve the energy-minimization problem for a surface. The proposed method is used for the segmentation of the patient body from the tomography reconstruction in nuclear medicine. The body surface in nuclear medicine is very important in order to model the attenuation of the photons.