Share Email Print

Proceedings Paper

Informative representation learning for automatic target recognition
Author(s): Charles F. Hester; Kelly K. D. Risko
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

Informative representations are those representations that do more than reconstruct the data; they have information embedded implicitly in them and are compressive for utilization in real-time Automatic Target Recognition. In this paper we create methods for embedding information in subspace bases through sparsity and information theoretic measures. We present a theory of informative bases and demonstrate some practical examples of basis learning using infrared imagery. We will employ sparsity and entropy measures to drive the learning process to extract the most informative representation and will draw relations between informative representations and the quadratic correlation filter.

Paper Details

Date Published: 19 May 2011
PDF: 10 pages
Proc. SPIE 8049, Automatic Target Recognition XXI, 80490A (19 May 2011); doi: 10.1117/12.885031
Show Author Affiliations
Charles F. Hester, U.S. Army Research, Development, and Engineering Command (United States)
Kelly K. D. Risko, U.S. Army Research, Development, and Engineering Command (United States)

Published in SPIE Proceedings Vol. 8049:
Automatic Target Recognition XXI
Firooz A. Sadjadi; Abhijit Mahalanobis, Editor(s)

© SPIE. Terms of Use
Back to Top