MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR
Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR), most algorithms of which have employed and relied on sufficient training samples to receive a strong discriminative classification model, has remained a challenging task in recent years, among which the challenge of SAR data acquisit...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-04-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/12/9/1385 |
| _version_ | 1797569395611926528 |
|---|---|
| author | Yikui Zhai Wenbo Deng Tian Lan Bing Sun Zilu Ying Junying Gan Chaoyun Mai Jingwen Li Ruggero Donida Labati Vincenzo Piuri Fabio Scotti |
| author_facet | Yikui Zhai Wenbo Deng Tian Lan Bing Sun Zilu Ying Junying Gan Chaoyun Mai Jingwen Li Ruggero Donida Labati Vincenzo Piuri Fabio Scotti |
| author_sort | Yikui Zhai |
| collection | DOAJ |
| description | Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR), most algorithms of which have employed and relied on sufficient training samples to receive a strong discriminative classification model, has remained a challenging task in recent years, among which the challenge of SAR data acquisition and further insight into the intuitive features of SAR images are the main concerns. In this paper, a deep transferred multi-level feature fusion attention network with dual optimized loss, called a multi-level feature attention Synthetic Aperture Radar network (MFFA-SARNET), is proposed to settle the problem of small samples in SAR ATR tasks. Firstly, a multi-level feature attention (MFFA) network is established to learn more discriminative features from SAR images with a fusion method, followed by alleviating the impact of background features on images with the following attention module that focuses more on the target features. Secondly, a novel dual optimized loss is incorporated to further optimize the classification network, which enhances the robust and discriminative learning power of features. Thirdly, transfer learning is utilized to validate the variances and small-sample classification tasks. Extensive experiments conducted on a public database with three different configurations consistently demonstrate the effectiveness of our proposed network, and the significant improvements yielded to surpass those of the state-of-the-art methods under small-sample conditions. |
| first_indexed | 2024-03-10T20:11:07Z |
| format | Article |
| id | doaj.art-19bdce6842ed4d64b581eaee6fb0e989 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T20:11:07Z |
| publishDate | 2020-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-19bdce6842ed4d64b581eaee6fb0e9892023-11-19T22:53:49ZengMDPI AGRemote Sensing2072-42922020-04-01129138510.3390/rs12091385MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATRYikui Zhai0Wenbo Deng1Tian Lan2Bing Sun3Zilu Ying4Junying Gan5Chaoyun Mai6Jingwen Li7Ruggero Donida Labati8Vincenzo Piuri9Fabio Scotti10Department of Intelligent Manufacturing, Wuyi University, Jiangmen 529020, ChinaDepartment of Intelligent Manufacturing, Wuyi University, Jiangmen 529020, ChinaDepartment of Intelligent Manufacturing, Wuyi University, Jiangmen 529020, ChinaSchool of Electronics and Information Engineering, Beihang University, Beijing 100191, ChinaDepartment of Intelligent Manufacturing, Wuyi University, Jiangmen 529020, ChinaDepartment of Intelligent Manufacturing, Wuyi University, Jiangmen 529020, ChinaDepartment of Intelligent Manufacturing, Wuyi University, Jiangmen 529020, ChinaSchool of Electronics and Information Engineering, Beihang University, Beijing 100191, ChinaDepartimento di Information, Universita, Degli Studi di Milano, via Celoria 18, 20133 Milano (MI), ItalyDepartimento di Information, Universita, Degli Studi di Milano, via Celoria 18, 20133 Milano (MI), ItalyDepartimento di Information, Universita, Degli Studi di Milano, via Celoria 18, 20133 Milano (MI), ItalySynthetic Aperture Radar (SAR) Automatic Target Recognition (ATR), most algorithms of which have employed and relied on sufficient training samples to receive a strong discriminative classification model, has remained a challenging task in recent years, among which the challenge of SAR data acquisition and further insight into the intuitive features of SAR images are the main concerns. In this paper, a deep transferred multi-level feature fusion attention network with dual optimized loss, called a multi-level feature attention Synthetic Aperture Radar network (MFFA-SARNET), is proposed to settle the problem of small samples in SAR ATR tasks. Firstly, a multi-level feature attention (MFFA) network is established to learn more discriminative features from SAR images with a fusion method, followed by alleviating the impact of background features on images with the following attention module that focuses more on the target features. Secondly, a novel dual optimized loss is incorporated to further optimize the classification network, which enhances the robust and discriminative learning power of features. Thirdly, transfer learning is utilized to validate the variances and small-sample classification tasks. Extensive experiments conducted on a public database with three different configurations consistently demonstrate the effectiveness of our proposed network, and the significant improvements yielded to surpass those of the state-of-the-art methods under small-sample conditions.https://www.mdpi.com/2072-4292/12/9/1385SAR ATRattention networkfeature fusiondual optimized losstransfer learningsmall samples |
| spellingShingle | Yikui Zhai Wenbo Deng Tian Lan Bing Sun Zilu Ying Junying Gan Chaoyun Mai Jingwen Li Ruggero Donida Labati Vincenzo Piuri Fabio Scotti MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR Remote Sensing SAR ATR attention network feature fusion dual optimized loss transfer learning small samples |
| title | MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR |
| title_full | MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR |
| title_fullStr | MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR |
| title_full_unstemmed | MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR |
| title_short | MFFA-SARNET: Deep Transferred Multi-Level Feature Fusion Attention Network with Dual Optimized Loss for Small-Sample SAR ATR |
| title_sort | mffa sarnet deep transferred multi level feature fusion attention network with dual optimized loss for small sample sar atr |
| topic | SAR ATR attention network feature fusion dual optimized loss transfer learning small samples |
| url | https://www.mdpi.com/2072-4292/12/9/1385 |
| work_keys_str_mv | AT yikuizhai mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT wenbodeng mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT tianlan mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT bingsun mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT ziluying mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT junyinggan mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT chaoyunmai mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT jingwenli mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT ruggerodonidalabati mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT vincenzopiuri mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr AT fabioscotti mffasarnetdeeptransferredmultilevelfeaturefusionattentionnetworkwithdualoptimizedlossforsmallsamplesaratr |