Skip navigation.

UCSD Computer Vision

Locally Smooth Manifold Learning

Piotr Dollár, Vincent Rabaud, Serge Belongie


Synopsis

LSML is a method for determining a warping from a point on a manifold to its neighbors on the manifold. A direct application of this method is performed on video sequences where the ways of moving on the underlying manifold are learned and then used to move within and out of the training set. The warping is also applied to an unseen frame in order to transfer the transformations.



Downloads

Please, ask pdollar@caltech.edu or vrabaud@cs.ucsd.edu for the password in order to download our toolbox. The LSML code requires Piotr's Image Processing Toolbox, version 2.20 or higher. Updated as of Mar. 06, 2009, see readme.

user:
password:


Related Publications

2008
Rabaud V., Belongie S., "Re-Thinking Non-Rigid Structure From Motion", IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2008. [BibTex][pdf]
2007
Dollár P., Rabaud V., Belongie S., "Learning to Traverse Image Manifolds", Tech Report, no. CS2007-087: UCSD CSE, 2007. [BibTex][pdf]
Dollár P., Rabaud V., Belongie S., "Non-Isometric Manifold Learning: Analysis and an Algorithm", International Conference On Machine Learning (ICML), June, 2007. [BibTex][pdf]
2006
Dollár P., Rabaud V., Belongie S., "Learning to Traverse Image Manifolds", Neural Information Processing Systems Conference (NIPS), Dec., 2006. [BibTex][pdf]