File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Active Ensemble Deep Learning for Polarimetric Synthetic Aperture Radar Image Classification

TitleActive Ensemble Deep Learning for Polarimetric Synthetic Aperture Radar Image Classification
Authors
KeywordsActive learning
convolutional neural network (CNN)
deep learning
ensemble learning
image classification
Issue Date2021
PublisherInstitute of Electrical and Electronics Engineers. The Journal's web site is located at https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=8859
Citation
IEEE Geoscience and Remote Sensing Letters, 2021, v. 18 n. 9, p. 1580-1584 How to Cite?
AbstractAlthough deep learning has achieved great success in the image-classification tasks, its performance is subject to the quantity and quality of the training samples. For the classification of the polarimetric synthetic aperture radar (PolSAR) images, it is nearly impossible to annotate the images from visual interpretation. Therefore, it is urgent for remote-sensing scientists to develop new techniques for PolSAR image classification under the condition of very few training samples. In this letter, we take the advantage of active learning and propose active ensemble deep learning (AEDL) for PolSAR image classification. We first show that only 35% of the predicted labels of the deep-learning model’s snapshots near its convergence were exactly the same. The disagreement between the snapshots is nonnegligible. From the perspective of multiview learning, the snapshots together serve as a good committee to evaluate the importance of the unlabeled instances. Using the snapshot committee to give out the informativeness of the unlabeled data, the proposed AEDL achieved better performance on two real PolSAR images than the standard active learning strategies. It achieved the same classification accuracy with only 86% and 55% of the training samples compared to the breaking tie active learning and random selection for the Flevoland data set.
Persistent Identifierhttp://hdl.handle.net/10722/304571
ISSN
2023 Impact Factor: 4.0
2023 SCImago Journal Rankings: 1.248
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorLiu, SJ-
dc.contributor.authorLuo, H-
dc.contributor.authorShi, Q-
dc.date.accessioned2021-09-23T09:01:58Z-
dc.date.available2021-09-23T09:01:58Z-
dc.date.issued2021-
dc.identifier.citationIEEE Geoscience and Remote Sensing Letters, 2021, v. 18 n. 9, p. 1580-1584-
dc.identifier.issn1545-598X-
dc.identifier.urihttp://hdl.handle.net/10722/304571-
dc.description.abstractAlthough deep learning has achieved great success in the image-classification tasks, its performance is subject to the quantity and quality of the training samples. For the classification of the polarimetric synthetic aperture radar (PolSAR) images, it is nearly impossible to annotate the images from visual interpretation. Therefore, it is urgent for remote-sensing scientists to develop new techniques for PolSAR image classification under the condition of very few training samples. In this letter, we take the advantage of active learning and propose active ensemble deep learning (AEDL) for PolSAR image classification. We first show that only 35% of the predicted labels of the deep-learning model’s snapshots near its convergence were exactly the same. The disagreement between the snapshots is nonnegligible. From the perspective of multiview learning, the snapshots together serve as a good committee to evaluate the importance of the unlabeled instances. Using the snapshot committee to give out the informativeness of the unlabeled data, the proposed AEDL achieved better performance on two real PolSAR images than the standard active learning strategies. It achieved the same classification accuracy with only 86% and 55% of the training samples compared to the breaking tie active learning and random selection for the Flevoland data set.-
dc.languageeng-
dc.publisherInstitute of Electrical and Electronics Engineers. The Journal's web site is located at https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=8859-
dc.relation.ispartofIEEE Geoscience and Remote Sensing Letters-
dc.rightsIEEE Geoscience and Remote Sensing Letters. Copyright © Institute of Electrical and Electronics Engineers.-
dc.rights©20xx 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.-
dc.subjectActive learning-
dc.subjectconvolutional neural network (CNN)-
dc.subjectdeep learning-
dc.subjectensemble learning-
dc.subjectimage classification-
dc.titleActive Ensemble Deep Learning for Polarimetric Synthetic Aperture Radar Image Classification-
dc.typeArticle-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1109/LGRS.2020.3005076-
dc.identifier.hkuros325236-
dc.identifier.volume18-
dc.identifier.issue9-
dc.identifier.spage1580-
dc.identifier.epage1584-
dc.identifier.isiWOS:000690441200022-
dc.publisher.placeUnited States-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats