Skip to Main Content
A novel efficient skeleton extraction algorithm is proposed, which is based on feature points extraction and Reeb graph theories. Because of the topological facility of feature points, a model can be divided into several branches according to them. One feature point can present one branch. So we just extract the other point of the branch - that could be skeleton point, connect these points with their corresponding feature points, then we can get all branch skeletons. Finally, connecting branch skeletons through connecting skeleton points according the topological relationship of each skeleton point preserving, then 3D modelpsilas skeleton can be extracted.Without pre-processing stages and without input parameters, this algorithm can automatically extract the skeleton of 3D models. Theoretical analyses and experimental results show that our method has a lower computing complexity, and meets the requirement of extracting nice-looking and affine-invariant skeletons efficiently.