Implicit Surface Reconstruction and uniform Sampling by Striped Marching Triangles

Abstract

The synthesis of computer models from real objects is a usual procedure in medical image processing and reverse engineering. In both cases, the main issues are to reach a topologically consistent and geometrically precise object reconstruction. Implicit surfaces provide a solution to this problem. They define a smooth surface surrounding objects which must be polygonized for an efficient visualization. Due to the increasing size of the application generated data, it becomes necessary to use precise and efficient methods. This work presents a modified method of implicit surface polygonization based on “Marching Triangles”. This modification assures the topological consistency with the initial dataset during the progressive reconstruction of the surface, avoiding overlapping triangles. In addition, a comparison of our method with the marching cubes and the adaptive skeleton climbing algorithms is provided.