| Paper: | SS-10.3 | ||
| Session: | Manifolds and Geometry in Signal Processing | ||
| Time: | Friday, May 21, 10:10 - 10:30 | ||
| Presentation: | Special Session Lecture | ||
| Topic: | Special Sessions: Manifolds and Geometry in Signal Processing | ||
| Title: | MANIFOLD LEARNING USING EUCLIDEAN k-NEAREST NEIGHBOR GRAPHS | ||
| Authors: | Jose Costa; University of Michigan | ||
| Alfred O. Hero III; University of Michigan | |||
| Abstract: | In the manifold learning problem one seeks to discover a smooth low dimensional surface, i.e., a manifold embedded in a higher dimensional linear vector space, based on a set of n measured sample points on the surface. In this paper we consider the closely related problem of estimating the manifold's intrinsic dimension and the intrinsic entropy of the sample points. Specifically, we view the sample points as realizations of an unknown multivariate density supported on an unknown smooth manifold. In previous work we introduced a geometric probability method called Geodesic Minimal Spanning Tree (GMST) to obtain asymptotically consistent estimates of manifold dimension and entropy. In this paper we present a simpler method based on the k-nearest neighbor (k-NN) graph that does not require estimation of geodesic distances on the manifold. The algorithm is applied to standard synthetic manifolds as well as real data sets consisting of images of faces. | ||
| Back | |||
Home -||-
Organizing Committee -||-
Technical Committee -||-
Technical Program -||-
Plenaries
Paper Submission -||-
Special Sessions -||-
ITT -||-
Paper Review -||-
Exhibits -||-
Tutorials
Information -||-
Registration -||-
Travel Insurance -||-
Housing -||-
Workshops