Curvature-constrained Skeleton Graph Extraction And Its Applications

In recent decade 3D model skeleton extraction is a worldwide important research issue. Many scientific organization, including Siggraph, and scholar have taken a deep research on this issue, making it improved fast.Skeletons represented the topological structure of 3D models; they have been widely used for similarity comparison, animation and data compression areas recently.Reeb graphs are compact shape descriptors which convey topological information related to the level sets of a function defined on the shape. But it is not capable of dealing with local geometric details. This article addressed an optimized skeleton extraction approach by using discrete Gaussian curvature based on MRG method. It first calculated the Gaussian curvature of each vertex which implied the convex and concave feature of the local surfaces, and it then searched the curvature extrema (concave point) as constrained points. Finally, new nodes denoting the joint features are extracted and used for adaptively updating the original Reeb graph. The advantage of the enhanced feature graph is to provide an affine-invariant and visually meaningful skeleton of arbitrary topological shape in a reasonable execution time.A series of experiments with applications in animation and shape matching have been implemented and shown the robustness and efficiency.This article could be separated into 3 parts:The first part introduces related work, comparison and its application.The second part would be the main part which expatiates the algorithm of this article.The last part shows some application of the skeleton extracted from this article including skeleton deforming and 3D models retrieval.