elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

Learning to Generate SAR Images with Adversarial Autoencoder

Song, Qian and Xu, Feng and Zhu, Xiao Xiang and Jin, Ya-Qiu (2022) Learning to Generate SAR Images with Adversarial Autoencoder. IEEE Transactions on Geoscience and Remote Sensing, 60, p. 5210015. IEEE - Institute of Electrical and Electronics Engineers. doi: 10.1109/TGRS.2021.3086817. ISSN 0196-2892.

[img] PDF - Postprint version (accepted manuscript)
6MB

Official URL: https://ieeexplore.ieee.org/document/9461232

Abstract

Deep learning-based synthetic aperture radar (SAR) target recognition often suffers from sparsely distributed training samples and rapid angular variations due to scattering scintillation. Thus, data-driven SAR target recognition is considered a typical few-shot learning (FSL) task. This paper first reviews the key issues of FSL and provides a definition of the FSL task. A novel adversarial autoencoder (AAE) is then proposed as a SAR representation and generation network. It consists of a generator network that decodes target knowledge to SAR images and an adversarial discriminator network that not only learns to discriminate “fake” generated images from real ones but also encodes the input SAR image back to a target knowledge. The discriminator employs progressively expanding convolution layers and a corresponding layer-by-layer training strategy. It uses two cyclic loss functions to enforce consistency between the inputs and outputs. Moreover, rotated cropping is introduced as a mechanism to address the challenge of representing the target orientation. The MSTAR 7-target dataset is used to evaluate the AAE’s performance, and the results demonstrate its ability to generate SAR images with aspect angular diversity. Using only 90 training samples with at least 25 degrees of orientation interval, the trained AAE is able to generate the remaining 1,748 samples of other orientation angles with an unprecedented level of fidelity. Thus, it can be used for data augmentation in SAR target recognition FSL tasks. Our experimental results show that the AAE could boost the test accuracy by 5.77%.

Item URL in elib:https://elib.dlr.de/142830/
Document Type:Article
Title:Learning to Generate SAR Images with Adversarial Autoencoder
Authors:
AuthorsInstitution or Email of AuthorsAuthor's ORCID iD
Song, QianQian.Song (at) dlr.deUNSPECIFIED
Xu, Fengfengxu (at) fudan.edu.cnUNSPECIFIED
Zhu, Xiao XiangXiaoxiang.Zhu (at) dlr.deUNSPECIFIED
Jin, Ya-Qiuyqjin (at) fudan.edu.cnUNSPECIFIED
Date:January 2022
Journal or Publication Title:IEEE Transactions on Geoscience and Remote Sensing
Refereed publication:Yes
Open Access:Yes
Gold Open Access:No
In SCOPUS:Yes
In ISI Web of Science:Yes
Volume:60
DOI :10.1109/TGRS.2021.3086817
Page Range:p. 5210015
Publisher:IEEE - Institute of Electrical and Electronics Engineers
ISSN:0196-2892
Status:Published
Keywords:Synthetic Aperture Radar (SAR), Image Representation, Adversarial Autoencoder, Few-shot Learning (FSL), Deep Learning
HGF - Research field:Aeronautics, Space and Transport
HGF - Program:Space
HGF - Program Themes:Earth Observation
DLR - Research area:Raumfahrt
DLR - Program:R EO - Earth Observation
DLR - Research theme (Project):R - Artificial Intelligence
Location: Oberpfaffenhofen
Institutes and Institutions:Remote Sensing Technology Institute > EO Data Science
Deposited By: Song, Qian
Deposited On:24 Nov 2021 13:38
Last Modified:14 Jan 2022 15:29

Repository Staff Only: item control page

Browse
Search
Help & Contact
Information
electronic library is running on EPrints 3.3.12
Copyright © 2008-2017 German Aerospace Center (DLR). All rights reserved.