Share Email Print
cover

Proceedings Paper

Nonlinear feature extraction and Bayesian mixture model approaches to target classification using MMW ISAR imagery: a preliminary study
Author(s): Adrian Britton; Keith D. Copsey; Guy T. Maskall; Andrew R. Webb; Karl West
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The problem we are addressing is one of generalization: given training data characterizing a set of targets (in specific configurations), how can we design a classifier that is robust to changes in target configuration and can generalize to other targets of the same generic class? The specific problem is identifying land vehicles from an inverse synthetic aperture radar image of the target. Issues in data modeling, experimental design and exploratory data analysis are discussed. Two complementary approaches are described: one that seeks to capture structure in the high- dimensional data space by projecting the data nonlinearly to a reduced dimensional feature space prior to classification; and a second that models the data in the data space using a Bayesian mixture model approach. Preliminary results for the mixture model approach are presented.

Paper Details

Date Published: 19 July 2000
PDF: 12 pages
Proc. SPIE 4033, Radar Sensor Technology V, (19 July 2000); doi: 10.1117/12.391844
Show Author Affiliations
Adrian Britton, Defence Evaluation and Research Agency Malvern (United Kingdom)
Keith D. Copsey, Defence Evaluation and Research Agency Malvern (United Kingdom)
Guy T. Maskall, Defence Evaluation and Research Agency Malvern (United Kingdom)
Andrew R. Webb, Defence Evaluation and Research Agency Malvern (United Kingdom)
Karl West, Defence Evaluation and Research Agency Malvern (United Kingdom)


Published in SPIE Proceedings Vol. 4033:
Radar Sensor Technology V
Robert Trebits; James L. Kurtz, Editor(s)

© SPIE. Terms of Use
Back to Top