Share Email Print
cover

Proceedings Paper

Modern approaches in deep learning for SAR ATR
Author(s): Michael Wilmanski; Chris Kreucher; Jim Lauer
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Recent breakthroughs in computational capabilities and optimization algorithms have enabled a new class of signal processing approaches based on deep neural networks (DNNs). These algorithms have been extremely successful in the classification of natural images, audio, and text data. In particular, a special type of DNNs, called convolutional neural networks (CNNs) have recently shown superior performance for object recognition in image processing applications. This paper discusses modern training approaches adopted from the image processing literature and shows how those approaches enable significantly improved performance for synthetic aperture radar (SAR) automatic target recognition (ATR). In particular, we show how a set of novel enhancements to the learning algorithm, based on new stochastic gradient descent approaches, generate significant classification improvement over previously published results on a standard dataset called MSTAR.

Paper Details

Date Published: 14 May 2016
PDF: 10 pages
Proc. SPIE 9843, Algorithms for Synthetic Aperture Radar Imagery XXIII, 98430N (14 May 2016); doi: 10.1117/12.2220290
Show Author Affiliations
Michael Wilmanski, Integrity Applications, Inc. (United States)
Chris Kreucher, Integrity Applications, Inc. (United States)
Jim Lauer, Integrity Applications, Inc. (United States)


Published in SPIE Proceedings Vol. 9843:
Algorithms for Synthetic Aperture Radar Imagery XXIII
Edmund Zelnio; Frederick D. Garber, Editor(s)

© SPIE. Terms of Use
Back to Top