A Metric Learning Method for Image-based 3D Shape Retrieval
Authors: 
Panpan Mu, Sanyuan Zhang, Xiuzi YeAuthors Info & Claims
Article No.: 17, Pages 1 - 5
Abstract
With the advent of new technology in 3D reconstruction and capturing, the number of 3D shapes is growing rapidly. We already have enormous images, thus it is very meaningful to retrieve 3D shapes using images as the queries. This problem can be solved by the following metric learning framework. First, we render each 3D shape to a view-images set by a predefined camera setup, and construct the view-images set as a covariance matrix. It is a point lying on a Riemannian manifold. Second, we construct each query image as a Euclidean point, and design a framework to map the (symmetric positive definite matrices) SPDs and Euclidean points to the same high-dimensional Hilbert space, in which distance computation is fast and convenient. At last, we transform this problem to an optimization problem and solve it by an iterative algorithm. The final experiments proves the effectiveness of our method.
References
[1]
Bai, S., Bai, X., Zhou, Z., Zhang, Z., and Jan Latecki, L., 2016. Gift: A real-time and scalable 3d shape search engine. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 5023--5032.
[2]
Su, H., Maji, S., Kalogerakis, E., and Learned-Miller, E., 2015. Multi-view convolutional neural networks for 3d shape recognition. In Proceedings of the IEEE International Conference on Computer Vision, 945--953.
[3]
Bai, X., Bai, S., Zhu, Z., and Latecki, L.J., 2015. 3d shape matching via two layer coding. IEEE Transactions on Pattern Analysis and Machine Intelligence 37, 12, 2361--2373.
[4]
Tabia, H., Laga, H., Picard, D., and Gosselin, P.-H., 2014. Covariance descriptors for 3D shape matching and retrieval. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 4185--4192.
[5]
Huang, Z., Wang, R., Shan, S., and Chen, X., 2014. Learning euclidean-to-riemannian metric for point-to-set classification. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1677--1684.
[6]
Wu, Z., Song, S., Khosla, A., Yu, F., Zhang, L., Tang, X., and Xiao, J., 2015. 3d shapenets: A deep representation for volumetric shapes. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 1912--1920.
[7]
Chatfield, K., Simonyan, K., Vedaldi, A., and Zisserman, A., 2014. Return of the devil in the details: Delving deep into convolutional nets. arXiv preprint arXiv:1405.3531.
[8]
Kazhdan, M., Funkhouser, T., and Rusinkiewicz, S., 2003. Rotation invariant spherical harmonic representation of 3 d shape descriptors. In Symposium on geometry processing, 156--164.
[9]
Chen, D.Y., Tian, X.P., Shen, Y.T., and Ming, O., 2003. On Visual Similarity Based 3D Model Retrieval. Computer Graphics Forum 22, 3, 223--232.
[10]
Papadakis, P., Pratikakis, I., Theoharis, T., and Perantonis, S., 2010. PANORAMA: A 3D Shape Descriptor Based on Panoramic Views for Unsupervised 3D Object Retrieval. International Journal of Computer Vision 89, 2--3 (2010/09/01), 177--192.
[11]
Bai, S., Bai, X., Zhou, Z., Zhang, Z., and Latecki, L.J., 2016. Gift: A real-time and scalable 3d shape search engine. arXiv preprint arXiv:1604.01879.
Index Terms
- A Metric Learning Method for Image-based 3D Shape Retrieval
Recommendations
Set-to-Set Distance Metric Learning on SPD Manifolds
Pattern Recognition and Computer VisionAbstractThe Symmetric Positive Definite (SPD) matrix on the Riemannian manifold has become a prevalent representation in many computer vision tasks. However, learning a proper distance metric between two SPD matrices is still a challenging problem. ...
Feature extraction by learning Lorentzian metric tensor and its extensions
We develop a supervised dimensionality reduction method, called Lorentzian discriminant projection (LDP), for feature extraction and classification. Our method represents the structures of sample data by a manifold, which is furnished with a Lorentzian ...
Kernel-based distance metric learning for content-based image retrieval
For a specific set of features chosen for representing images, the performance of a content-based image retrieval (CBIR) system depends critically on the similarity or dissimilarity measure used. Instead of manually choosing a distance function in ...
Comments
Please enable JavaScript to view thecomments powered by Disqus.Information & Contributors
Information
Published In
Copyright © 2017 ACM.
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]
In-Cooperation
- The Hong Kong Polytechnic: The Hong Kong Polytechnic University
- HKBU: Hong Kong Baptist University
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Published: 25 May 2017
Check for updates
Author Tags
Qualifiers
- Research-article
- Research
- Refereed limited
Funding Sources
Conference
DMCIT '17
DMCIT '17: 2017 International Conference on Data Mining, Communications and Information Technology
May 25 - 27, 2017
Phuket, Thailand
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 123Total Downloads
- Downloads (Last 12 months)0
- Downloads (Last 6 weeks)0
Reflects downloads up to 07 Mar 2025
Other Metrics
Citations
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in