 |
General: 663 |
P2: Nonlinear Dimensionality Reduction for Hyperspectral Image Classification
Author: Tim Doster , Advisors: John J. Benedetto (Mathematics & Norbert Wiener Center), Wojciech Czaja (Mathematics and Norbert Wiener Center)
Problem Statement Presentation
Project Proposal
Abstract
Today with sensors becoming more complex and cost no longer a deterrent to storing large amounts of data, analysts need methods to reduce the volume of stored data and reveal its important facets. Dimensionality reduction, particularly non-linear dimensionality reduction, is a solution to this problem. In this paper, we will look at two nonlinear dimensionality reduction algorithms, Local Linear Embedding and Isomap. These algorithms both have been shown to work well with artificial and real world data sets, but are computationally expensive to execute. We solve this problem for both algorithms by applying landmarks or out of sample extensions. Finally, we will apply these algorithms first to artificial data sets for validation and then to hyperspectral images for the application of classification.
MidYear Progress Report and
Presentation
Final Presentation ,
Final Report
