Abstract:A three-dimensional (3D) skeletonization algorithm extracts the skeleton of a 3D model and provides it for many applications, such as 3D model classification and identification. There are three major skeletonization methodologies used in the literature, distance transform field-based methods, Voronoi diagram-based methods, and thinning-based methods. However, the existing algorithms cannot preserve the connectivity of the skeletons of the 3D mesh models. In this paper, we propose a 3D skeletonization algorithm for 3D mesh models using a partial parallel thinning algorithm and a 3D skeleton correcting algorithm. The proposed algorithm uses pre-defined removing and recovering templates. The partial parallel 3D thinning algorithm separates 62 symmetrical removing templates into two groups based on symmetry. It thins a model with the templates of each group in each thinning procedure. The 3D skeleton correcting algorithm uses six correcting templates to inspect the disconnected voxels in the skeleton and corrects them. The experimental results show several comparisons of skeletons extracted by different skeletonization algorithms. The proposed algorithm can extract the skeleton of each branch of a model and preserve the connectivity.

No abstract

Abstract

No abstract

Abstract:In this paper, we propose a 3 dimensional (3D) model identification method based on weighted implicit shape representation (WISR) and panoramic view. The WISR is used for 3D shape normalization. The 3D shape normalization method normalizes a 3D model by scaling, translation, and rotation with respect to the scale factor, center, and principal axes. The major advantage of the WISR is reduction of the influences caused by shape deformation and partial removal. The well-known scale-invariant feature transform descriptors are extracted from the panoramic view of the 3D model for feature matching. The panoramic view is a range image obtained by projecting a 3D model to the surface of a cylinder which is parallel to a principal axis determined by the 3D shape normalization. Because of using only one range image, the proposed method can provide small size of features and fast matching speed. The precision of the identification is 92% with 1200 models that consist of 24 deformed versions of 50 classes. The average feature size and matching time are 4.1 KB and 1.9 s.