Few Shot Object Detection for SAR Images via Feature Enhancement and Dynamic Relationship Modeling
Current Synthetic Aperture Radar (SAR) image object detection methods require huge amounts of annotated data and can only detect the categories that appears in the training set. Due to the lack of training samples in the real applications, the performance decreases sharply on rare categories, which...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-07-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/15/3669 |
| _version_ | 1827601739693424640 |
|---|---|
| author | Shiqi Chen Jun Zhang Ronghui Zhan Rongqiang Zhu Wei Wang |
| author_facet | Shiqi Chen Jun Zhang Ronghui Zhan Rongqiang Zhu Wei Wang |
| author_sort | Shiqi Chen |
| collection | DOAJ |
| description | Current Synthetic Aperture Radar (SAR) image object detection methods require huge amounts of annotated data and can only detect the categories that appears in the training set. Due to the lack of training samples in the real applications, the performance decreases sharply on rare categories, which largely inhibits the detection model from reaching robustness. To tackle this problem, a novel few-shot SAR object detection framework is proposed, which is built upon the meta-learning architecture and aims at detecting objects of unseen classes given only a few annotated examples. Observing the quality of support features determines the performance of the few-shot object detection task, we propose an attention mechanism to highlight class-specific features while softening the irrelevant background information. Considering the variation between different support images, we also employ a support-guided module to enhance query features, thus generating high-qualified proposals more relevant to support images. To further exploit the relevance between support and query images, which is ignored in single class representation, a dynamic relationship learning paradigm is designed via constructing a graph convolutional network and imposing orthogonality constraint in hidden feature space, which both make features from the same category more closer and those from different classes more separable. Comprehensive experiments have been completed on the self-constructed SAR multi-class object detection dataset, which demonstrate the effectiveness of our few-shot object detection framework in learning more generalized features to both enhance the performance on novel classes and maintain the performance on base classes. |
| first_indexed | 2024-03-09T05:02:42Z |
| format | Article |
| id | doaj.art-1acb562e4ac84d389784d6641cc22b91 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T05:02:42Z |
| publishDate | 2022-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-1acb562e4ac84d389784d6641cc22b912023-12-03T12:58:15ZengMDPI AGRemote Sensing2072-42922022-07-011415366910.3390/rs14153669Few Shot Object Detection for SAR Images via Feature Enhancement and Dynamic Relationship ModelingShiqi Chen0Jun Zhang1Ronghui Zhan2Rongqiang Zhu3Wei Wang4National Key Laboratory of Science and Technology on Automatic Target Recognition, National University of Defense Technology, Changsha 410073, ChinaNational Key Laboratory of Science and Technology on Automatic Target Recognition, National University of Defense Technology, Changsha 410073, ChinaNational Key Laboratory of Science and Technology on Automatic Target Recognition, National University of Defense Technology, Changsha 410073, ChinaSouthwest Electronics and Telecommunication Technology Research Institute, Chengdu 610000, ChinaNational Key Laboratory of Science and Technology on Automatic Target Recognition, National University of Defense Technology, Changsha 410073, ChinaCurrent Synthetic Aperture Radar (SAR) image object detection methods require huge amounts of annotated data and can only detect the categories that appears in the training set. Due to the lack of training samples in the real applications, the performance decreases sharply on rare categories, which largely inhibits the detection model from reaching robustness. To tackle this problem, a novel few-shot SAR object detection framework is proposed, which is built upon the meta-learning architecture and aims at detecting objects of unseen classes given only a few annotated examples. Observing the quality of support features determines the performance of the few-shot object detection task, we propose an attention mechanism to highlight class-specific features while softening the irrelevant background information. Considering the variation between different support images, we also employ a support-guided module to enhance query features, thus generating high-qualified proposals more relevant to support images. To further exploit the relevance between support and query images, which is ignored in single class representation, a dynamic relationship learning paradigm is designed via constructing a graph convolutional network and imposing orthogonality constraint in hidden feature space, which both make features from the same category more closer and those from different classes more separable. Comprehensive experiments have been completed on the self-constructed SAR multi-class object detection dataset, which demonstrate the effectiveness of our few-shot object detection framework in learning more generalized features to both enhance the performance on novel classes and maintain the performance on base classes.https://www.mdpi.com/2072-4292/14/15/3669Synthetic Aperture Radar (SAR)few-shot object detectionattentionsupport-guided modulegraph convolutional network |
| spellingShingle | Shiqi Chen Jun Zhang Ronghui Zhan Rongqiang Zhu Wei Wang Few Shot Object Detection for SAR Images via Feature Enhancement and Dynamic Relationship Modeling Remote Sensing Synthetic Aperture Radar (SAR) few-shot object detection attention support-guided module graph convolutional network |
| title | Few Shot Object Detection for SAR Images via Feature Enhancement and Dynamic Relationship Modeling |
| title_full | Few Shot Object Detection for SAR Images via Feature Enhancement and Dynamic Relationship Modeling |
| title_fullStr | Few Shot Object Detection for SAR Images via Feature Enhancement and Dynamic Relationship Modeling |
| title_full_unstemmed | Few Shot Object Detection for SAR Images via Feature Enhancement and Dynamic Relationship Modeling |
| title_short | Few Shot Object Detection for SAR Images via Feature Enhancement and Dynamic Relationship Modeling |
| title_sort | few shot object detection for sar images via feature enhancement and dynamic relationship modeling |
| topic | Synthetic Aperture Radar (SAR) few-shot object detection attention support-guided module graph convolutional network |
| url | https://www.mdpi.com/2072-4292/14/15/3669 |
| work_keys_str_mv | AT shiqichen fewshotobjectdetectionforsarimagesviafeatureenhancementanddynamicrelationshipmodeling AT junzhang fewshotobjectdetectionforsarimagesviafeatureenhancementanddynamicrelationshipmodeling AT ronghuizhan fewshotobjectdetectionforsarimagesviafeatureenhancementanddynamicrelationshipmodeling AT rongqiangzhu fewshotobjectdetectionforsarimagesviafeatureenhancementanddynamicrelationshipmodeling AT weiwang fewshotobjectdetectionforsarimagesviafeatureenhancementanddynamicrelationshipmodeling |