Mixtures of probabilistic principal component analyzers
Neural Computation
Laplacian Eigenmaps for dimensionality reduction and data representation
Neural Computation
Shape Matching: Similarity Measures and Algorithms
SMI '01 Proceedings of the International Conference on Shape Modeling & Applications
3D model metrieval based on volumetric extended gaussian image and hierarchical self organizing map
MULTIMEDIA '06 Proceedings of the 14th annual ACM international conference on Multimedia
ACM SIGGRAPH 2002 conference abstracts and applications
| Hi-index | 0.00 |
One of the most challenging issues in object based 3D model retrieval research is how to extract efficient feature vector from the original model with translation, scaling and rotation independence. Spherical harmonic transform based methods can maintain the property of rotation independence. But there is a pre-requisite for performing spherical harmonic transform: the surface sphere should be uniformly sampled. However, spherical harmonic transform based methods divide the spherical surface by sampling the latitude and longitude direction uniformly which causes singularities in the two poles. In other words, the poles are over-sampled while the equator parts are under-sampled. In this paper, we introduce an improvement approach to fix this problem through incorporating a variant principal component analysis into the Volumetric Extended Gaussian Image (VEGI) shape representation proposed by J. Zhang et al. We evaluate this adaptive VEGI shape representation on the Princeton Shape Benchmark database and a public online 3D model retrieval system we developed. The experimental results show that our proposed retrieval approach enhance the VEGI method.