An automatic framework for 3D objects-parts learning
This paper describes an approach for 3D triangular meshes segmentation. The method that we adopt for the 3D mesh segmentation is based on the unbiased hierarchical queue which has been used for 2D images. Our method uses the hierarchical transformation on a connected faces structure. The approach that we propose is based on the principal curvature to calculate the faces curvature. The results show the effectiveness of the proposed method. Indeed our approach permits to get a fast and efficient segmentation of 3D mesh models.