What's mine is yours: Pretrained CNNs for limited training sonar ATR

John McKay, Isaac Gerg, Vishal Monga, Raghu G. Raj

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

3 Citations (Scopus)

Abstract

Finding mines in Sonar imagery is a significant problem with a great deal of relevance for seafaring military and commercial endeavors. Unfortunately, the lack of enormous Sonar image data sets has prevented automatic target recognition (ATR) algorithms from some of the same advances seen in other computer vision fields. Namely, the boom in convolutional neural nets (CNNs) which have been able to achieve incredible results - even surpassing human actors - has not been an easily feasible route for many practitioners of Sonar ATR. We demonstrate the power of one avenue to incorporating CNNs into Sonar ATR: transfer learning. We first show how well a straightforward, flexible CNN feature-extraction strategy can be used to obtain impressive if not state-of-the-art results. Secondly, we propose a way to utilize the powerful transfer learning approach towards multiple instance target detection and identification within a provided synthetic aperture Sonar data set.

Original languageEnglish (US)
Title of host publicationOCEANS 2017 � Anchorage
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1-7
Number of pages7
ISBN (Electronic)9780692946909
StatePublished - Dec 19 2017
EventOCEANS 2017 - Anchorage - Anchorage, United States
Duration: Sep 18 2017Sep 21 2017

Publication series

NameOCEANS 2017 - Anchorage
Volume2017-January

Other

OtherOCEANS 2017 - Anchorage
CountryUnited States
CityAnchorage
Period9/18/179/21/17

Fingerprint

Automatic target recognition
neural nets
target recognition
sonar
Sonar
education
Neural networks
Synthetic aperture sonar
learning
sonar imagery
computer vision
boom
Target tracking
synthetic apertures
Computer vision
Feature extraction
pattern recognition
imagery
routes

All Science Journal Classification (ASJC) codes

  • Oceanography
  • Automotive Engineering
  • Water Science and Technology
  • Acoustics and Ultrasonics
  • Instrumentation
  • Ocean Engineering

Cite this

McKay, J., Gerg, I., Monga, V., & Raj, R. G. (2017). What's mine is yours: Pretrained CNNs for limited training sonar ATR. In OCEANS 2017 � Anchorage (pp. 1-7). (OCEANS 2017 - Anchorage; Vol. 2017-January). Institute of Electrical and Electronics Engineers Inc..
McKay, John ; Gerg, Isaac ; Monga, Vishal ; Raj, Raghu G. / What's mine is yours : Pretrained CNNs for limited training sonar ATR. OCEANS 2017 � Anchorage. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 1-7 (OCEANS 2017 - Anchorage).
@inproceedings{04467e4819d546b788da3fb700f377b6,
title = "What's mine is yours: Pretrained CNNs for limited training sonar ATR",
abstract = "Finding mines in Sonar imagery is a significant problem with a great deal of relevance for seafaring military and commercial endeavors. Unfortunately, the lack of enormous Sonar image data sets has prevented automatic target recognition (ATR) algorithms from some of the same advances seen in other computer vision fields. Namely, the boom in convolutional neural nets (CNNs) which have been able to achieve incredible results - even surpassing human actors - has not been an easily feasible route for many practitioners of Sonar ATR. We demonstrate the power of one avenue to incorporating CNNs into Sonar ATR: transfer learning. We first show how well a straightforward, flexible CNN feature-extraction strategy can be used to obtain impressive if not state-of-the-art results. Secondly, we propose a way to utilize the powerful transfer learning approach towards multiple instance target detection and identification within a provided synthetic aperture Sonar data set.",
author = "John McKay and Isaac Gerg and Vishal Monga and Raj, {Raghu G.}",
year = "2017",
month = "12",
day = "19",
language = "English (US)",
series = "OCEANS 2017 - Anchorage",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "1--7",
booktitle = "OCEANS 2017 � Anchorage",
address = "United States",

}

McKay, J, Gerg, I, Monga, V & Raj, RG 2017, What's mine is yours: Pretrained CNNs for limited training sonar ATR. in OCEANS 2017 � Anchorage. OCEANS 2017 - Anchorage, vol. 2017-January, Institute of Electrical and Electronics Engineers Inc., pp. 1-7, OCEANS 2017 - Anchorage, Anchorage, United States, 9/18/17.

What's mine is yours : Pretrained CNNs for limited training sonar ATR. / McKay, John; Gerg, Isaac; Monga, Vishal; Raj, Raghu G.

OCEANS 2017 � Anchorage. Institute of Electrical and Electronics Engineers Inc., 2017. p. 1-7 (OCEANS 2017 - Anchorage; Vol. 2017-January).

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

TY - GEN

T1 - What's mine is yours

T2 - Pretrained CNNs for limited training sonar ATR

AU - McKay, John

AU - Gerg, Isaac

AU - Monga, Vishal

AU - Raj, Raghu G.

PY - 2017/12/19

Y1 - 2017/12/19

N2 - Finding mines in Sonar imagery is a significant problem with a great deal of relevance for seafaring military and commercial endeavors. Unfortunately, the lack of enormous Sonar image data sets has prevented automatic target recognition (ATR) algorithms from some of the same advances seen in other computer vision fields. Namely, the boom in convolutional neural nets (CNNs) which have been able to achieve incredible results - even surpassing human actors - has not been an easily feasible route for many practitioners of Sonar ATR. We demonstrate the power of one avenue to incorporating CNNs into Sonar ATR: transfer learning. We first show how well a straightforward, flexible CNN feature-extraction strategy can be used to obtain impressive if not state-of-the-art results. Secondly, we propose a way to utilize the powerful transfer learning approach towards multiple instance target detection and identification within a provided synthetic aperture Sonar data set.

AB - Finding mines in Sonar imagery is a significant problem with a great deal of relevance for seafaring military and commercial endeavors. Unfortunately, the lack of enormous Sonar image data sets has prevented automatic target recognition (ATR) algorithms from some of the same advances seen in other computer vision fields. Namely, the boom in convolutional neural nets (CNNs) which have been able to achieve incredible results - even surpassing human actors - has not been an easily feasible route for many practitioners of Sonar ATR. We demonstrate the power of one avenue to incorporating CNNs into Sonar ATR: transfer learning. We first show how well a straightforward, flexible CNN feature-extraction strategy can be used to obtain impressive if not state-of-the-art results. Secondly, we propose a way to utilize the powerful transfer learning approach towards multiple instance target detection and identification within a provided synthetic aperture Sonar data set.

UR - http://www.scopus.com/inward/record.url?scp=85048211240&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048211240&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85048211240

T3 - OCEANS 2017 - Anchorage

SP - 1

EP - 7

BT - OCEANS 2017 � Anchorage

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

McKay J, Gerg I, Monga V, Raj RG. What's mine is yours: Pretrained CNNs for limited training sonar ATR. In OCEANS 2017 � Anchorage. Institute of Electrical and Electronics Engineers Inc. 2017. p. 1-7. (OCEANS 2017 - Anchorage).