## Shape-Similarity Search of 3D Models by using Enhanced Shape Functions (2003)

### Cached

### Download Links

- [www.kki.yamanashi.ac.jp]
- [www.kki.yamanashi.ac.jp]
- [www.kki.yamanashi.ac.jp]
- [www.kki.yamanashi.ac.jp]
- DBLP

### Other Repositories/Bibliography

Citations: | 42 - 5 self |

### BibTeX

@INPROCEEDINGS{Ohbuchi03shape-similaritysearch,

author = {Ryutarou Ohbuchi and Takahiro Minamitani and Tsuyoshi Takei},

title = {Shape-Similarity Search of 3D Models by using Enhanced Shape Functions},

booktitle = {},

year = {2003},

pages = {97--104}

}

### Years of Citing Articles

### OpenURL

### Abstract

shape-similarity search of 3D polygonal-mesh models. The shape features are extension of the D2 shape functions proposed by Osada et al. [OsadaO1, Osada02]. Our proposed shape features are tolerant of topological variations and geometrical degeneracies. Our shape feature is also invariant to similarity transformation. Experiments showed that, with only modest increase in computational cost, our shape feature achieved significant performance improvement over Osada's D2.

### Citations

302 | Topology matching for fully automatic similarity estimation of 3d shapes - Hilaga, Shinagawa, et al. - 2001 |

228 | D.: A search engine for 3D models
- FUNKHOUSER, MIN, et al.
- 2003
(Show Context)
Citation Context ...antly outperformed D2 in our evaluation experiment. While the AD and AAD might have the performance lower than the more elaborate shape features, such as the spherical harmonics based feature used in =-=[Funkhouser03]. -=-However, the computational costs of AD and AD2 estimated to be significantly lower than that of the sophisticated shape feature used in [Funkhouser03]. As the future work, we’d like to improve our p... |

171 | D.: Matching 3d models with shape distributions
- OSADA, FUNKHOUSER, et al.
(Show Context)
Citation Context ...nce interval. It is robust against variation in tessellation, not sensitive to connectivity of surfaces nor surface orientation. To generate points on triangles, we adopted the method by Osada et al. =-=[Osada01], whic-=-h uses the following formula to generate a point P at a random location on a triangle.s( 1 r1 ) 1 r1 ( 1 r2) 2 r1 ( r2 3) P = − t + − t + ⋅t . (1) In the formula, t 1 , t 2 , and t 3 are vertice... |

140 | 3D Shape Histograms for Similarity Search and Classification
- Ankerst, Kastenmüller, et al.
- 1999
(Show Context)
Citation Context ...ir pose normalization. The points used to generate the covariance matrix may be the original vertices or they may be generated uniformly on the surfaces for the purpose of pose normalization. Ankerst =-=[12]-=- proposed one of the first 3D shape similarity matching algorithms that targeted 3D molecular databases. One of their shape descriptor parameterizes the 3D space using concentric spherical shells, mak... |

94 | Content Based Retrieval of VRML Objects: An Iterative and
- Elad, Tal, et al.
- 2001
(Show Context)
Citation Context ...ession words and 3D shape by means of multidimensional scaling. Zaharia [9] employed a 3D Hough transformation computed from distribution of vertices of a model as its shape feature, while Elad et al =-=[10]-=- computed various moments from the points generated randomly on the surface. Elad’s tried to match the human notion of shape similarity with that of mechanical distance by employing a learning classif... |

91 | Shape matching: similarity measures and algorithms - Veltkamp - 2001 |

70 | Tools for 3D Object Retrieval: Karhunen-Loeve Transform and Spherical Harmonics - Vranic, Saupe, et al. - 2001 |

49 | Polyhedral model retrieval using weighted point sets, in: Shape Modeling - Tangelder, Veltkamp |

34 | A similarity retrieval of 3D polygonal models using rotation invariant shape descriptors - Suzuki, Kato, et al. - 2000 |

32 | A geometric approach to 3D object comparison - Novotni, Klein - 2001 |

29 | Efficient geometry-based similarity search of 3d spatial databases - Keim - 1999 |

27 | Nefertiti: a Query by Content Software for ThreeDimensional Models Database Management - Paquet, Rioux |

26 | Description of shape information for - Paquet, Murching, et al. - 2000 |

23 | Managing Digital Libraries for Computer-Aided Design, Computer Aided Design - Regli, Cicirello |

23 | F.: Shape-based retrieval of 3D mesh models - Zaharia, Preteux - 2002 |

22 | Three-Dimensional Shape-Based Retrieval within the MPEG-7 Framework - Zaharia, Prteux - 2001 |

22 |
Ming Ouhyoung: On visual similarity based 3D model retrieval
- Chen, Tian, et al.
- 2003
(Show Context)
Citation Context ... spherical harmonics shape feature is combined with the normalization of translation and uniform scaling to achieve invariance to all the 7 DOF of similarity transformation. The methods by Chen et al =-=[14]-=- and Ohbuchi et al [15] compares 3D shape by using a set of 2D images taken from multiple orientations. These two methods first normalize position and scale DOF to place a normalized model at the coor... |

15 | Transformation invariant shape similarity comparison of solid models - McWherter, Peabody, et al. - 2001 |

8 | 3D Object Retrieval based on Subjective Measures - Suzuki, Kato, et al. - 1998 |

7 | An Algorithm for Deciding Similarities of 3-D Objects - Mukai, Furukawa, et al. - 2002 |

4 |
Press et al., Numerical Recipes in C-The Art
- H
- 1992
(Show Context)
Citation Context ...he point generation. In our implementation of the D2, which we call modified D2 (mD2) to distinguish from Osada’s, we used the low-discrepancy sequence or quasi-random number sequence (QRNS) by Sobo=-=l [Press92], in-=-stead of the PRNS. Compared to PRNS, the Sobol’s QRNS produces feature vectors that are more consistent, i.e., low-variance. That is, using the Sobol’s sequence, generated points are more spatiall... |

4 |
Takei, Shape-Similarity Search of Three-Dimensional Models Using Parameterized Statistics
- Ohbuchi, Otagiri, et al.
- 2002
(Show Context)
Citation Context ...omly on the surface. Elad’s tried to match the human notion of shape similarity with that of mechanical distance by employing a learning classifier support vector machine. The method by Ohbuchi et al =-=[11]-=- normalizes pose, then computes moment of inertia, average and variance of distance from the three principal axes to the model surface. These methods employed principal component analysis of the covar... |

2 | Directed Search in A 3D Objects Database Using - Elad, Tal, et al. - 2000 |

2 | Tomo Otagiri, Ryutarou Ohbuchi, Shape-similarity Search of Three-Dimensional Models Based on Subjective - Ibato - 2002 |

1 |
Masatoshi Nakazawa, Tsuyoshi Takei, Retrieving 3D Shapes Based On Their Appearance
- Ohbuchi
- 2003
(Show Context)
Citation Context ...ape feature is combined with the normalization of translation and uniform scaling to achieve invariance to all the 7 DOF of similarity transformation. The methods by Chen et al [14] and Ohbuchi et al =-=[15]-=- compares 3D shape by using a set of 2D images taken from multiple orientations. These two methods first normalize position and scale DOF to place a normalized model at the coordinate origin. Then, 2 ... |

1 |
Numerical Recipes in C-The Art of Scientific Programming, 2nd Ed
- Press
- 1992
(Show Context)
Citation Context ...f Computer Applications in Technology (IJCAT), pp.70-85, Vol.23, No. 2/3/4, 2005 version the modified D2 (mD2). Instead of a PRNS used by Osada, we used a Quasi-Random Number Sequence (QRNS) by Sobol =-=[16]-=- for the r 1 , r 2 above. The spatially uniformity of distribution of points on polygons are better if a QRNS is used, compared to the case in which a PRNS is used (See Figure 3.) That is, feature vec... |