A comparison of methods for estimating differential properties of triangulation models

Estimation of differential properties of a discrete geometric model is an important stage of numerous algorithms in geometric modeling, computer vision, computer graphics, computer aided design and reverse engineering. This paper presents a review about different methods for curvature estimation of local geometric properties on triangulation models. In order to compare different methods of curvature estimation, some experiments are designed and carried out with several triangulation models as samples. We give the statistical measure and point out excellent methods for the estimation of Gaussian and mean curvatures on triangulation models. An application about feature detecting and quality checking is shown.