Feature Extraction Using an Information Theoretic Framework

Abstract

This report addresses the rejection of confusers, the last piece of the work conducted under the contract F33615-97-1-1019. The performance of the information theoretic feature extraction is elucidated and compared with the traditional (perceptrons and template matchers) classifiers in the Moving and Stationary Target Acquisition and Recognition (MSTAR) database. But no performance evaluation would he complete without assessing the quality of the new classifier in rejection to confusers. Therefore, the MSTAR database and the previous classifiers were utilized as the basis of comparison. Results are that the information theoretic feature extraction works at the same performance level as the very sophisticated support vector machine (SVM) for both misclassification error and rejection to confusers. The method should be more widely applied since its use transcends classification: it is a general method to create features that preserve as much information as possible with respect to a given response. It has also been shown that the same principle can be applied to classification and pose estimation, which shows the wide applicability of the technique.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1999
Accession Number
ADA397483

Entities

People

  • José Príncipe

Organizations

  • University of Florida

Tags

Communities of Interest

  • Energy and Power Technologies
  • Engineered Resilient Systems
  • Sensors

DTIC Thesaurus Topics

  • Air Force Research Laboratories
  • Artificial Intelligence
  • Classification
  • Databases
  • Detection
  • Detectors
  • Dimensionality Reduction
  • Feature Extraction
  • Information Science
  • Information Theory
  • Machine Learning
  • Pattern Recognition
  • Recognition
  • Signal Processing
  • Supervised Machine Learning
  • Synthetic Aperture Radar
  • Target Recognition

Fields of Study

  • Computer science

Readers

  • Computer Vision.
  • Neural Network Machine Learning.
  • Systems Analysis and Design

Technology Areas

  • AI & ML