When Deep Learning Meets Multi-Task Learning in SAR ATR: Simultaneous Target Recognition and Segmentation
With the recent advances of deep learning, automatic target recognition (ATR) of synthetic aperture radar (SAR) has achieved superior performance. By not being limited to the target category, the SAR ATR system could benefit from the simultaneous extraction of multifarious target attributes. In this...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-11-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/12/23/3863 |
| _version_ | 1797546807505453056 |
|---|---|
| author | Chenwei Wang Jifang Pei Zhiyong Wang Yulin Huang Junjie Wu Haiguang Yang Jianyu Yang |
| author_facet | Chenwei Wang Jifang Pei Zhiyong Wang Yulin Huang Junjie Wu Haiguang Yang Jianyu Yang |
| author_sort | Chenwei Wang |
| collection | DOAJ |
| description | With the recent advances of deep learning, automatic target recognition (ATR) of synthetic aperture radar (SAR) has achieved superior performance. By not being limited to the target category, the SAR ATR system could benefit from the simultaneous extraction of multifarious target attributes. In this paper, we propose a new multi-task learning approach for SAR ATR, which could obtain the accurate category and precise shape of the targets simultaneously. By introducing deep learning theory into multi-task learning, we first propose a novel multi-task deep learning framework with two main structures: encoder and decoder. The encoder is constructed to extract sufficient image features in different scales for the decoder, while the decoder is a tasks-specific structure which employs these extracted features adaptively and optimally to meet the different feature demands of the recognition and segmentation. Therefore, the proposed framework has the ability to achieve superior recognition and segmentation performance. Based on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset, experimental results show the superiority of the proposed framework in terms of recognition and segmentation. |
| first_indexed | 2024-03-10T14:34:26Z |
| format | Article |
| id | doaj.art-d765d80a7e6a4b97898ff9feed2d2ed9 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T14:34:26Z |
| publishDate | 2020-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-d765d80a7e6a4b97898ff9feed2d2ed92023-11-20T22:18:37ZengMDPI AGRemote Sensing2072-42922020-11-011223386310.3390/rs12233863When Deep Learning Meets Multi-Task Learning in SAR ATR: Simultaneous Target Recognition and SegmentationChenwei Wang0Jifang Pei1Zhiyong Wang2Yulin Huang3Junjie Wu4Haiguang Yang5Jianyu Yang6The Department of Electrical Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaThe Department of Electrical Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaThe Department of Electrical Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaThe Department of Electrical Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaThe Department of Electrical Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaThe Department of Electrical Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaThe Department of Electrical Engineering, University of Electronic Science and Technology of China, Chengdu 611731, ChinaWith the recent advances of deep learning, automatic target recognition (ATR) of synthetic aperture radar (SAR) has achieved superior performance. By not being limited to the target category, the SAR ATR system could benefit from the simultaneous extraction of multifarious target attributes. In this paper, we propose a new multi-task learning approach for SAR ATR, which could obtain the accurate category and precise shape of the targets simultaneously. By introducing deep learning theory into multi-task learning, we first propose a novel multi-task deep learning framework with two main structures: encoder and decoder. The encoder is constructed to extract sufficient image features in different scales for the decoder, while the decoder is a tasks-specific structure which employs these extracted features adaptively and optimally to meet the different feature demands of the recognition and segmentation. Therefore, the proposed framework has the ability to achieve superior recognition and segmentation performance. Based on the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset, experimental results show the superiority of the proposed framework in terms of recognition and segmentation.https://www.mdpi.com/2072-4292/12/23/3863synthetic aperture radar (SAR)automatic target recognition (ATR)multi-task learningdeep learning |
| spellingShingle | Chenwei Wang Jifang Pei Zhiyong Wang Yulin Huang Junjie Wu Haiguang Yang Jianyu Yang When Deep Learning Meets Multi-Task Learning in SAR ATR: Simultaneous Target Recognition and Segmentation Remote Sensing synthetic aperture radar (SAR) automatic target recognition (ATR) multi-task learning deep learning |
| title | When Deep Learning Meets Multi-Task Learning in SAR ATR: Simultaneous Target Recognition and Segmentation |
| title_full | When Deep Learning Meets Multi-Task Learning in SAR ATR: Simultaneous Target Recognition and Segmentation |
| title_fullStr | When Deep Learning Meets Multi-Task Learning in SAR ATR: Simultaneous Target Recognition and Segmentation |
| title_full_unstemmed | When Deep Learning Meets Multi-Task Learning in SAR ATR: Simultaneous Target Recognition and Segmentation |
| title_short | When Deep Learning Meets Multi-Task Learning in SAR ATR: Simultaneous Target Recognition and Segmentation |
| title_sort | when deep learning meets multi task learning in sar atr simultaneous target recognition and segmentation |
| topic | synthetic aperture radar (SAR) automatic target recognition (ATR) multi-task learning deep learning |
| url | https://www.mdpi.com/2072-4292/12/23/3863 |
| work_keys_str_mv | AT chenweiwang whendeeplearningmeetsmultitasklearninginsaratrsimultaneoustargetrecognitionandsegmentation AT jifangpei whendeeplearningmeetsmultitasklearninginsaratrsimultaneoustargetrecognitionandsegmentation AT zhiyongwang whendeeplearningmeetsmultitasklearninginsaratrsimultaneoustargetrecognitionandsegmentation AT yulinhuang whendeeplearningmeetsmultitasklearninginsaratrsimultaneoustargetrecognitionandsegmentation AT junjiewu whendeeplearningmeetsmultitasklearninginsaratrsimultaneoustargetrecognitionandsegmentation AT haiguangyang whendeeplearningmeetsmultitasklearninginsaratrsimultaneoustargetrecognitionandsegmentation AT jianyuyang whendeeplearningmeetsmultitasklearninginsaratrsimultaneoustargetrecognitionandsegmentation |