3D model similarity measurement with geometric feature map based on phase-encoded range image

Authors:
Donghui Wang;Chenyang Cui
Affiliations:
Dept. of Computer Science, Zhejiang University, Hangzhou, Zhejiang, P.R.China;National Key Lab. Of CAD&CG, Zhejiang University, Hangzhou, Zhejiang, P.R.China
Venue:
PCM'04 Proceedings of the 5th Pacific Rim conference on Advances in Multimedia Information Processing - Volume Part III
Year:
2004

Citing 6
Cited 0

Efficient Data Structures for Model-Based 3-D Object Recognition and Localization from Range Images

IEEE Transactions on Pattern Analysis and Machine Intelligence
Shape feature determination usiang the curvature region representation

SMA '97 Proceedings of the fourth ACM symposium on Solid modeling and applications
Topology matching for fully automatic similarity estimation of 3D shapes

Proceedings of the 28th annual conference on Computer graphics and interactive techniques
The Complex EGI: A New Representation for 3-D Pose Determination

IEEE Transactions on Pattern Analysis and Machine Intelligence
Shape-Similarity Search of Three-Dimensional Models Using Parameterized Statistics

PG '02 Proceedings of the 10th Pacific Conference on Computer Graphics and Applications
Matching 3D Models with Shape Distributions

SMI '01 Proceedings of the International Conference on Shape Modeling & Applications

Quantified Score

Hi-index	0.00

Visualization

Abstract

Measuring the similarity between 3D models is a very important problem in 3D model retrieval. A challenge aspect of this problem is to find a suitable shape descriptor that can grasp the feature of 3D model. In this paper, Geometric Feature Map of 3D model based on phase-encoded range-image is proposed. This map contains the information about the surface normal of the model and the area proportion of the planar surface of the models .From the local map of model at every possible rotation, we can obtain a global map of the model. The similarity calculation between 3D models is processed using a coarse-to-fine strategy,the similarity calculation is fast and efficient. The experimentation result shows that our method is invariant to translation , rotation and scaling of the model and agree with general human intuition, and particularly useful for classification of 3D models.