Abstract:This paper presents a novel 3D shape descriptor "the generalized shape distributions" for effective shape matching and analysis, by taking advantage of both local and global shape signatures. We start this process by generating spin images on meshes. These local shape descriptors are then quantized via k-means clustering. The key contribution of this paper is to represent a global 3D shape as the spatial configuration of a set of specific local shapes. We achieve this goal by computing the distributions of the Euclidean distance of pairs of local shape clusters. Because of the spatial, sparse distribution of local shapes defined over a 3D model, an indexing data structure is adopted to reduce the space complexity of the proposed shape descriptor. The technical merits of our new approach are at least two-fold: (1) it is robust to non-trivial shape occlusions and deformations, since there are statistically a large number of chances that some local shape signatures and their spatial layouts are unchanged and users can easily identify those unchanged parts; (2) it is more discriminative than a simple collection of local shape signatures, since the spatial layouts of a global shape are explicitly computed. Our preliminary experiments have shown the effectiveness of this new approach for shape comparison and analysis

New Spherical Analysis Method For 3d Shape Matching

3D Shape Matching Using Collinearity Constraint

Shape Representation and Recognition in High Dimensional Feature Space

Radius-normal Histogram and Hybrid Strategy for 3D Shape Retrieval

A 3D Shape Retrieval Method Based on Continuous Spherical Wavelet Transform

Spherical Hybrid Curvature Images of 3d Shapes and Its Applications

A New 3D Shape Retrieval Method Using Spherical Healpix.

The Generalized Shape Distributions for Shape Matching and Analysis

Superquadric Matching Computation with Spherical Harmonics and 3D Moments

A Shape Distributions Retrieval Algorithm of 3D CAD Models Based on Normal Direction

A Metric Learning Method for Image-based 3D Shape Retrieval

3D Shape Similarity Comparison Using Multi-level Spherical Moments

A New Algorithm for Non-Rigid Shape Matching with Anisotropic-Scaling Transformation Parameters

Non-rigid 3D Shape Retrieval Using Multidimensional Scaling and Bag-of-Features

3d Shape Inner-Distance Computing For Shape Matching Based On Adaptive Volume Representation

A Novel 3D Shape Retrieval Method Using Point Spatial Distributions along Principal Axis

Multiresolution Wavelet Analysis of Shape Orientation for 3d Shape Retrieval

Evaluation of Spherical Parameterization Methods for Three-Dimensional Reconstruction of Medical Images

3D Shape Matching Based on Gromov-Wasserstein Distance

Matching contour of 3-D objects based on B-splines curve representation

3D partial surface matching using differential geometry and statistical approaches