Combining deep convolutional neural network and SVM to SAR image target recognition

Fei Gao, Teng Huang, Jun Wang, Jinping Sun, Erfu Yang, Amir Hussain

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

To address the challenging problem on target recognition from synthetic aperture radar (SAR) images, a novel method is proposed by combining Deep Convolutional Neural Network (DCNN) and Support Vector Machine (SVM). First, an improved DCNN is employed to learn the features of SAR images. Then, a SVM is utilized to map the leant features into the output labels. To enhance the feature extraction capability of DCNN, a class of separation information is also added to the cross-entropy cost function as a regularization term. As a result, this explicitly facilitates the intra-class compactness and separability in the process of feature learning. Numerical experiments are performed on the Moving and Stationary Target Acquisition and Recognition (MSTAR) database. The results demonstrate that the proposed method can achieve an average accuracy of 99.15% on ten types of targets.

LanguageEnglish
Title of host publication2017 IEEE International Conference on Internet of Things, IEEE Green Computing and Communications, IEEE Cyber, Physical and Social Computing, IEEE Smart Data, iThings-GreenCom-CPSCom-SmartData 2017
Place of PublicationPiscataway, NJ.
PublisherIEEE
Pages1082-1085
Number of pages4
Volume2018-January
ISBN (Electronic)9781538630655
DOIs
StatePublished - 30 Jan 2018
EventJoint 10th IEEE International Conference on Internet of Things, iThings 2017, 13th IEEE International Conference on Green Computing and Communications, GreenCom 2017, 10th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2017 and the 3rd IEEE International Conference on Smart Data, Smart Data 2017 - Exeter, United Kingdom
Duration: 21 Jun 201723 Jun 2017

Conference

ConferenceJoint 10th IEEE International Conference on Internet of Things, iThings 2017, 13th IEEE International Conference on Green Computing and Communications, GreenCom 2017, 10th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2017 and the 3rd IEEE International Conference on Smart Data, Smart Data 2017
CountryUnited Kingdom
CityExeter
Period21/06/1723/06/17

Fingerprint

Synthetic aperture radar
neural network
Support vector machines
Neural networks
entropy
Cost functions
Feature extraction
Labels
Entropy
experiment
costs
learning
Experiments

Keywords

  • automatic target recognition (ATR)
  • class separation information
  • deep convolutional neural network (DCNN)
  • support vector machine (SVM)
  • synthetic aperture radar (SAR)

Cite this

Gao, F., Huang, T., Wang, J., Sun, J., Yang, E., & Hussain, A. (2018). Combining deep convolutional neural network and SVM to SAR image target recognition. In 2017 IEEE International Conference on Internet of Things, IEEE Green Computing and Communications, IEEE Cyber, Physical and Social Computing, IEEE Smart Data, iThings-GreenCom-CPSCom-SmartData 2017 (Vol. 2018-January, pp. 1082-1085). Piscataway, NJ.: IEEE. DOI: 10.1109/iThings-GreenCom-CPSCom-SmartData.2017.165
Gao, Fei ; Huang, Teng ; Wang, Jun ; Sun, Jinping ; Yang, Erfu ; Hussain, Amir. / Combining deep convolutional neural network and SVM to SAR image target recognition. 2017 IEEE International Conference on Internet of Things, IEEE Green Computing and Communications, IEEE Cyber, Physical and Social Computing, IEEE Smart Data, iThings-GreenCom-CPSCom-SmartData 2017. Vol. 2018-January Piscataway, NJ. : IEEE, 2018. pp. 1082-1085
@inproceedings{f6978c76807446039c6b36ce469a93c3,
title = "Combining deep convolutional neural network and SVM to SAR image target recognition",
abstract = "To address the challenging problem on target recognition from synthetic aperture radar (SAR) images, a novel method is proposed by combining Deep Convolutional Neural Network (DCNN) and Support Vector Machine (SVM). First, an improved DCNN is employed to learn the features of SAR images. Then, a SVM is utilized to map the leant features into the output labels. To enhance the feature extraction capability of DCNN, a class of separation information is also added to the cross-entropy cost function as a regularization term. As a result, this explicitly facilitates the intra-class compactness and separability in the process of feature learning. Numerical experiments are performed on the Moving and Stationary Target Acquisition and Recognition (MSTAR) database. The results demonstrate that the proposed method can achieve an average accuracy of 99.15{\%} on ten types of targets.",
keywords = "automatic target recognition (ATR), class separation information, deep convolutional neural network (DCNN), support vector machine (SVM), synthetic aperture radar (SAR)",
author = "Fei Gao and Teng Huang and Jun Wang and Jinping Sun and Erfu Yang and Amir Hussain",
note = "{\circledC} 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.",
year = "2018",
month = "1",
day = "30",
doi = "10.1109/iThings-GreenCom-CPSCom-SmartData.2017.165",
language = "English",
volume = "2018-January",
pages = "1082--1085",
booktitle = "2017 IEEE International Conference on Internet of Things, IEEE Green Computing and Communications, IEEE Cyber, Physical and Social Computing, IEEE Smart Data, iThings-GreenCom-CPSCom-SmartData 2017",
publisher = "IEEE",

}

Gao, F, Huang, T, Wang, J, Sun, J, Yang, E & Hussain, A 2018, Combining deep convolutional neural network and SVM to SAR image target recognition. in 2017 IEEE International Conference on Internet of Things, IEEE Green Computing and Communications, IEEE Cyber, Physical and Social Computing, IEEE Smart Data, iThings-GreenCom-CPSCom-SmartData 2017. vol. 2018-January, IEEE, Piscataway, NJ., pp. 1082-1085, Joint 10th IEEE International Conference on Internet of Things, iThings 2017, 13th IEEE International Conference on Green Computing and Communications, GreenCom 2017, 10th IEEE International Conference on Cyber, Physical and Social Computing, CPSCom 2017 and the 3rd IEEE International Conference on Smart Data, Smart Data 2017, Exeter, United Kingdom, 21/06/17. DOI: 10.1109/iThings-GreenCom-CPSCom-SmartData.2017.165

Combining deep convolutional neural network and SVM to SAR image target recognition. / Gao, Fei; Huang, Teng; Wang, Jun; Sun, Jinping; Yang, Erfu; Hussain, Amir.

2017 IEEE International Conference on Internet of Things, IEEE Green Computing and Communications, IEEE Cyber, Physical and Social Computing, IEEE Smart Data, iThings-GreenCom-CPSCom-SmartData 2017. Vol. 2018-January Piscataway, NJ. : IEEE, 2018. p. 1082-1085.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Combining deep convolutional neural network and SVM to SAR image target recognition

AU - Gao,Fei

AU - Huang,Teng

AU - Wang,Jun

AU - Sun,Jinping

AU - Yang,Erfu

AU - Hussain,Amir

N1 - © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

PY - 2018/1/30

Y1 - 2018/1/30

N2 - To address the challenging problem on target recognition from synthetic aperture radar (SAR) images, a novel method is proposed by combining Deep Convolutional Neural Network (DCNN) and Support Vector Machine (SVM). First, an improved DCNN is employed to learn the features of SAR images. Then, a SVM is utilized to map the leant features into the output labels. To enhance the feature extraction capability of DCNN, a class of separation information is also added to the cross-entropy cost function as a regularization term. As a result, this explicitly facilitates the intra-class compactness and separability in the process of feature learning. Numerical experiments are performed on the Moving and Stationary Target Acquisition and Recognition (MSTAR) database. The results demonstrate that the proposed method can achieve an average accuracy of 99.15% on ten types of targets.

AB - To address the challenging problem on target recognition from synthetic aperture radar (SAR) images, a novel method is proposed by combining Deep Convolutional Neural Network (DCNN) and Support Vector Machine (SVM). First, an improved DCNN is employed to learn the features of SAR images. Then, a SVM is utilized to map the leant features into the output labels. To enhance the feature extraction capability of DCNN, a class of separation information is also added to the cross-entropy cost function as a regularization term. As a result, this explicitly facilitates the intra-class compactness and separability in the process of feature learning. Numerical experiments are performed on the Moving and Stationary Target Acquisition and Recognition (MSTAR) database. The results demonstrate that the proposed method can achieve an average accuracy of 99.15% on ten types of targets.

KW - automatic target recognition (ATR)

KW - class separation information

KW - deep convolutional neural network (DCNN)

KW - support vector machine (SVM)

KW - synthetic aperture radar (SAR)

UR - http://www.scopus.com/inward/record.url?scp=85047438600&partnerID=8YFLogxK

U2 - 10.1109/iThings-GreenCom-CPSCom-SmartData.2017.165

DO - 10.1109/iThings-GreenCom-CPSCom-SmartData.2017.165

M3 - Conference contribution

VL - 2018-January

SP - 1082

EP - 1085

BT - 2017 IEEE International Conference on Internet of Things, IEEE Green Computing and Communications, IEEE Cyber, Physical and Social Computing, IEEE Smart Data, iThings-GreenCom-CPSCom-SmartData 2017

PB - IEEE

CY - Piscataway, NJ.

ER -

Gao F, Huang T, Wang J, Sun J, Yang E, Hussain A. Combining deep convolutional neural network and SVM to SAR image target recognition. In 2017 IEEE International Conference on Internet of Things, IEEE Green Computing and Communications, IEEE Cyber, Physical and Social Computing, IEEE Smart Data, iThings-GreenCom-CPSCom-SmartData 2017. Vol. 2018-January. Piscataway, NJ.: IEEE. 2018. p. 1082-1085. Available from, DOI: 10.1109/iThings-GreenCom-CPSCom-SmartData.2017.165