Digital Repository at the University of Maryland (DRUM)  >
Theses and Dissertations from UMD  >
UMD Theses and Dissertations 

Please use this identifier to cite or link to this item:

Title: A Study of the Relationship Between Spectrum and Geometry Through Fourier Frames and Laplacian Eigenmaps
Authors: Duke, Kevin W.
Advisors: Benedetto, John J.
Department/Program: Mathematics
Type: Dissertation
Sponsors: Digital Repository at the University of Maryland
University of Maryland (College Park, Md.)
Subjects: Mathematics
Keywords: Cantor measures
Fourier frames
hyperspectral imagery
Laplacian Eigenmaps
sampling theory
Issue Date: 2012
Abstract: This thesis has two parts. The first part is a study of Fourier frames. We follow the development of the theory, beginning with its classical roots in non-uniform sampling in Paley-Wiener Spaces, to its current state, the study of the spectral properties of finite measures on locally compact abelian groups. The aim of our study is to understand the relationship between the geometry of the supporting set of a measure and the spectral properties it exhibits. In the second part, we study extensions of the Laplacian Eigenmaps algorithm and their uses in hyperspectral image analysis. In particular, we show that there is a natural way of including spatial information in the analysis that improves classification results. We also provide evidence supporting the use of Schrödinger Eigenmaps as a semisupervised tool for feature extraction. Finally, we show that Schrödinger Eigenmaps provides a platform for fusing Laplacian Eigenmaps with other clustering techniques, such as kmeans clustering.
Appears in Collections:Mathematics Theses and Dissertations
UMD Theses and Dissertations

Files in This Item:

File Description SizeFormatNo. of Downloads
Duke_umd_0117E_13600.pdf3.79 MBAdobe PDF328View/Open

All items in DRUM are protected by copyright, with all rights reserved.


DRUM is brought to you by the University of Maryland Libraries
University of Maryland, College Park, MD 20742-7011 (301)314-1328.
Please send us your comments