Diffusion Maps and Geometric Harmonics for Automatic Target Recognition (ATR). Volume 2. Appendices

Abstract

Geometric harmonics provides a framework for taking data in high-dimensional measurement spaces and embedding them in low dimensional Euclidean space according to a similarity measure. Euclidean coordinates then characterize the "manifold" on (or near) which the data live. Our goal in this project is to develop this manifold as a mechanism for integrating data from different sensors to facilitate automatic recognition. During the tenure of this research grant, we were able to formulate and complete the first series of experiments on embedded fusion. The resulting experiment on integrating voice and audio streams was extremely successful. This definitely revealed the potential for this approach and set the stage for further experiments. The problem formulation has been completed and confirmed with an experiment on the integration of audio and video streams. Researchers at AFRL, Wright-Patterson Air Force Base, have received a first version of the software, and are attempting to apply it to radar signal interpretation.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 2007
Accession Number
ADA476152

Entities

People

  • Ronald Coifman
  • Steven W. Zucker

Organizations

  • Yale University

Tags

Communities of Interest

  • C4I
  • Energy and Power Technologies
  • Sensors

DTIC Thesaurus Topics

  • Air Force
  • Cognitive Science
  • Computational Science
  • Computer Vision
  • Data Mining
  • Data Processing
  • Detectors
  • Dimensionality Reduction
  • Information Processing
  • Information Science
  • Machine Learning
  • Network Science
  • Self Organizing Systems
  • Simplex Method
  • Target Recognition
  • Three Dimensional
  • Two Dimensional

Readers

  • Neural Network Machine Learning.
  • Technical Research and Report Writing.
  • Theoretical Analysis.

Technology Areas

  • Space
  • Space - Space Objects