An algorithm of multi-structure based on Riemannian manifold learning

Wei Wang; Du-yan Bi; Lei Xiong

doi:10.1117/12.866982

20 August 2010 An algorithm of multi-structure based on Riemannian manifold learning

Wei Wang, Du-yan Bi, Lei Xiong

Proceedings Volume 7820, International Conference on Image Processing and Pattern Recognition in Industrial Engineering; 78202J (2010) https://doi.org/10.1117/12.866982
Event: International Conference on Image Processing and Pattern Recognition in Industrial Engineering, 2010, Xi'an, China

Abstract

Riemannian Manifold Learning (RML) is a global algorithm proposed recently, so it can't preserve the local geometry property of neighboring data well. An algorithm of multi-structure based on RML is proposed in order to solve the problem. In the algorithm, all points were projected by PCA firstly so as to the extracted character is irrelevant, then constructed a neighbor graph. The most important step was that all data points were classified to two parts, for the k - NN of a base point, it adopted a weight which can preserve local property of the base point and neighboring nods to get the low-dimensional embedding coordinates. As for the other points, it still used the RML algorithm. Thus the new algorithm can both preserve the metrics at all scales and keep the geometrical property of local neighbor to the maximum. Experimental results on synthetic data and MNIST data set demonstrate that the new algorithm can reflect the intrinsic property better than the other manifold learning algorithms.

Citation Download Citation

Wei Wang, Du-yan Bi, and Lei Xiong "An algorithm of multi-structure based on Riemannian manifold learning", Proc. SPIE 7820, International Conference on Image Processing and Pattern Recognition in Industrial Engineering, 78202J (20 August 2010); https://doi.org/10.1117/12.866982

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
6 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Detection and tracking algorithms

Principal component analysis

3D image processing

Error analysis

Bismuth

Data modeling

Vector spaces

Show All Keywords

Keywords/Phrases

Search In:

Publication Years