An active deep learning approach for minimally supervised polsar image classification
Name:
Polarimetric SAR Image Semantic ...
Size:
8.277Mb
Format:
PDF
Description:
Accepted Manuscript
Authors
Xue, YongAffiliation
University of DerbyFudan University, Shanghai, China
X'ian Electronics and Engineering Institute, China
Issue Date
2019-08-01
Metadata
Show full item recordAbstract
Recently, deep neural networks have received intense interests in polarimetric synthetic aperture radar (PolSAR) image classification. However, its success is subject to the availability of large amounts of annotated data which require great efforts of experienced human annotators. Aiming at improving the classification performance with greatly reduced annotation cost, this paper presents an active deep learning approach for minimally supervised PolSAR image classification, which integrates active learning and fine-tuned convolutional neural network (CNN) into a principled framework. Starting from a CNN trained using a very limited number of labeled pixels, we iteratively and actively select the most informative candidates for annotation, and incrementally fine-tune the CNN by incorporating the newly annotated pixels. Moreover, to boost the performance and robustness of the proposed method, we employ Markov random field (MRF) to enforce class label smoothness, and data augmentation technique to enlarge the training set. We conducted extensive experiments on four real benchmark PolSAR images, and experiments demonstrated that our approach achieved state-of-the-art classification results with significantly reduced annotation cost.Citation
Bi, H., Xu, F., Wei, Z., Xue, Y. and Xu, Z., (2019). 'An active deep learning approach for minimally supervised PolSAR image classification'. IEEE Transactions on Geoscience and Remote Sensing, 57(11), pp.9378-9395. DOI: 10.1109/TGRS.2019.2926434Publisher
IEEEJournal
IEEE Transactions on Geoscience and Remote SensingDOI
10.1109/TGRS.2019.2926434Type
ArticleLanguage
enISSN
01962892EISSN
15580644ae974a485f413a2113503eed53cd6c53
10.1109/TGRS.2019.2926434
Scopus Count
Collections
The following license files are associated with this item:
- Creative Commons