Fast Generation of Multiresolution Surfaces from Contours

Abstract

Surface reconstruction from contours is an important problem especially in medical applications. Other uses include reconstruction from topographic data, or isosurface generation in general. The drawback of existing reconstruction algorithms from contours is, that they are relatively complicated and often have numerical problems. Furthermore, algorithms to generate multiresolution surface models do not exploit the special situation having contours. In this paper we describe a new robust and fast reconstruction algorithm from contours that delivers a multiresolution surface with controlled distance from the original contours. Supporting selective refinement in areas of interest, this multiresolution model can be handled interactively without giving up accuracy.