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

39 Scopus citations


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.
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


OtherOCEANS 2017 - Anchorage
Country/TerritoryUnited States

All Science Journal Classification (ASJC) codes

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


Dive into the research topics of 'What's mine is yours: Pretrained CNNs for limited training sonar ATR'. Together they form a unique fingerprint.

Cite this