Dimensionality Reduction for the Visualization of Fuzzy Data
Dimensionality is one of the main obstacles for the visualization and classification of remotely sensed data. Here we address the problem when multi-spectral data is used in association with fuzzy verification data. We utilize the framework of exploratory projection pursuit, a statistical method that seeks interesting projections of a high dimensional space into a lower one, according to a measure of interestingness called projection index and a numerical optimization algorithm. It is a flexible and convenient framework that can include current approaches to the dimensionality reduction problem as well as fresh approaches. Two-dimensional and three dimensional projection pursuit algorithms have been developed for the visualization of multi-spectral data with projection indices, which maximize class separability. Fuzzy data contributes to further complication and each pixel is not only associated with with several spectral values but with several class membership values as well.
Bibliographic Reference: Paper presented: 99 Earth Observation from Data to Information, Cardiff (GB), 8-10th September (1999)
Availability: Available from Public Relations and Publications Unit, Ispra (IT)
Record Number: 199911583 / Last updated on: 1999-11-13
Original language: en
Available languages: en