A surface reconstruction approach based on multi-resolution methods and T-surfaces framework

Abstract

In this paper we present a new approach which integrates the T-Surfaces framework and a multi-resolution method in a unified methodology for segmentation and surface reconstruction. For noise images, we can improve the result by anisotropic diffusion. Despite this improvement, some manual intervention may be required to complete the reconstruction. Thus, we take advantage of the topological capabilities of T-Surfaces to enable the user to modify the topology of a surface. We test the method for synthetic and medical volumes images.