ResearchSpace

Transfer learning for multi-frequency synthetic aperture radar applications

Show simple item record

dc.contributor.author Schwegmann, Colin P
dc.contributor.author Kleynhans, Waldo
dc.contributor.author Salmon, BP
dc.contributor.author Mdakane, Lizwe W
dc.contributor.author Meyer, Rory GV
dc.contributor.author Janoth, J
dc.contributor.author Lumsdon, P
dc.date.accessioned 2019-04-10T10:54:11Z
dc.date.available 2019-04-10T10:54:11Z
dc.date.issued 2018-07
dc.identifier.citation Schwegmann, Colin, P., Kleynhans, W., Salmon, B.P. Mdakane, L.W. Meyer, R.G.V. Janoth, J. & Lumsdon, P .2018. Transfer learning for multi-frequency synthetic aperture radar applications. In: 2018 IGARSS: International Geoscience and Remote Sensing Symposium, 22-27 July 2018, Valencia, Spain en_US
dc.identifier.isbn 978-1-5386-7150-4
dc.identifier.isbn 978-1-5386-7149-8
dc.identifier.uri https://ieeexplore.ieee.org/document/8518401
dc.identifier.uri DOI: 10.1109/IGARSS.2018.8518401
dc.identifier.uri https://www.igarss2018.org/Papers/viewpapers.asp?papernum=3926
dc.identifier.uri http://hdl.handle.net/10204/10939
dc.description Presented in: 2018 IGARSS: International Geoscience and Remote Sensing Symposium, 22-27 July 2018, Valencia, Spain. Due to copyright restrictions, the attached PDF file only contains the abstract of the full-text item. For access to the full-text item, please consult the publisher's website. While waiting for the post-print or published PDF document from the publisher en_US
dc.description.abstract To correctly train a system to detect ships in Sythetic Aperture Radar imagery requires the creation of a validated dataset. This dataset is used to train a Machine Learning system to identify ships and non-ships. Recent advances in Deep Learning have focused on Transfer Learning which uses characteristics from one dataset to classify another without retraining the system. In this work two Deep Learning architectures were trained on lower resolution C-band SAR ship data and then tested against an unseen set of higher resolution Xband SAR ship data. Transfer Learning allowed the system to correctly identify 81% of the ships and 93% of the unseen data. This represents a possible reduction in the amount of effort required to curate new SAR datasets in the future by pre-identifying likely candidates as ships or non-ships. en_US
dc.language.iso en en_US
dc.publisher IEEE en_US
dc.relation.ispartofseries Workflow;21853
dc.subject Machine learning en_US
dc.subject Marine technology en_US
dc.subject Syntetic aperture radar en_US
dc.title Transfer learning for multi-frequency synthetic aperture radar applications en_US
dc.type Conference Presentation en_US
dc.identifier.apacitation Schwegmann, C. P., Kleynhans, W., Salmon, B., Mdakane, L. W., Meyer, R. G., Janoth, J., & Lumsdon, P. (2018). Transfer learning for multi-frequency synthetic aperture radar applications. IEEE. http://hdl.handle.net/10204/10939 en_ZA
dc.identifier.chicagocitation Schwegmann, Colin P, Waldo Kleynhans, BP Salmon, Lizwe W Mdakane, Rory GV Meyer, J Janoth, and P Lumsdon. "Transfer learning for multi-frequency synthetic aperture radar applications." (2018): http://hdl.handle.net/10204/10939 en_ZA
dc.identifier.vancouvercitation Schwegmann CP, Kleynhans W, Salmon B, Mdakane LW, Meyer RG, Janoth J, et al, Transfer learning for multi-frequency synthetic aperture radar applications; IEEE; 2018. http://hdl.handle.net/10204/10939 . en_ZA
dc.identifier.ris TY - Conference Presentation AU - Schwegmann, Colin P AU - Kleynhans, Waldo AU - Salmon, BP AU - Mdakane, Lizwe W AU - Meyer, Rory GV AU - Janoth, J AU - Lumsdon, P AB - To correctly train a system to detect ships in Sythetic Aperture Radar imagery requires the creation of a validated dataset. This dataset is used to train a Machine Learning system to identify ships and non-ships. Recent advances in Deep Learning have focused on Transfer Learning which uses characteristics from one dataset to classify another without retraining the system. In this work two Deep Learning architectures were trained on lower resolution C-band SAR ship data and then tested against an unseen set of higher resolution Xband SAR ship data. Transfer Learning allowed the system to correctly identify 81% of the ships and 93% of the unseen data. This represents a possible reduction in the amount of effort required to curate new SAR datasets in the future by pre-identifying likely candidates as ships or non-ships. DA - 2018-07 DB - ResearchSpace DP - CSIR KW - Machine learning KW - Marine technology KW - Syntetic aperture radar LK - https://researchspace.csir.co.za PY - 2018 SM - 978-1-5386-7150-4 SM - 978-1-5386-7149-8 T1 - Transfer learning for multi-frequency synthetic aperture radar applications TI - Transfer learning for multi-frequency synthetic aperture radar applications UR - http://hdl.handle.net/10204/10939 ER - en_ZA


Files in this item

This item appears in the following Collection(s)

Show simple item record