Subtask Attention Based Object Detection in Remote Sensing Images
Object detection in remote sensing images (RSIs) is one of the basic tasks in the field of remote sensing image automatic interpretation. In recent years, the deep object detection frameworks of natural scene images (NSIs) have been introduced into object detection on RSIs, and the detection perform...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-05-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/10/1925 |
| _version_ | 1797534082619408384 |
|---|---|
| author | Shengzhou Xiong Yihua Tan Yansheng Li Cai Wen Pei Yan |
| author_facet | Shengzhou Xiong Yihua Tan Yansheng Li Cai Wen Pei Yan |
| author_sort | Shengzhou Xiong |
| collection | DOAJ |
| description | Object detection in remote sensing images (RSIs) is one of the basic tasks in the field of remote sensing image automatic interpretation. In recent years, the deep object detection frameworks of natural scene images (NSIs) have been introduced into object detection on RSIs, and the detection performance has improved significantly because of the powerful feature representation. However, there are still many challenges concerning the particularities of remote sensing objects. One of the main challenges is the missed detection of small objects which have less than five percent of the pixels of the big objects. Generally, the existing algorithms choose to deal with this problem by multi-scale feature fusion based on a feature pyramid. However, the benefits of this strategy are limited, considering that the location of small objects in the feature map will disappear when the detection task is processed at the end of the network. In this study, we propose a subtask attention network (StAN), which handles the detection task directly on the shallow layer of the network. First, StAN contains one shared feature branch and two subtask attention branches of a semantic auxiliary subtask and a detection subtask based on the multi-task attention network (MTAN). Second, the detection branch uses only low-level features considering small objects. Third, the attention map guidance mechanism is put forward to optimize the network for keeping the identification ability. Fourth, the multi-dimensional sampling module (MdS), global multi-view channel weights (GMulW) and target-guided pixel attention (TPA) are designed for further improvement of the detection accuracy in complex scenes. The experimental results on the NWPU VHR-10 dataset and DOTA dataset demonstrated that the proposed algorithm achieved the SOTA performance, and the missed detection of small objects decreased. On the other hand, ablation experiments also proved the effects of MdS, GMulW and TPA. |
| first_indexed | 2024-03-10T11:24:36Z |
| format | Article |
| id | doaj.art-26eaff15cd064087b5917d6ef4c4fa08 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T11:24:36Z |
| publishDate | 2021-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-26eaff15cd064087b5917d6ef4c4fa082023-11-21T19:46:22ZengMDPI AGRemote Sensing2072-42922021-05-011310192510.3390/rs13101925Subtask Attention Based Object Detection in Remote Sensing ImagesShengzhou Xiong0Yihua Tan1Yansheng Li2Cai Wen3Pei Yan4National Key Laboratory of Science & Technology on Multi-Spectral Information Processing, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology, Wuhan 430074, ChinaNational Key Laboratory of Science & Technology on Multi-Spectral Information Processing, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology, Wuhan 430074, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, ChinaNational Key Laboratory of Science & Technology on Multi-Spectral Information Processing, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology, Wuhan 430074, ChinaNational Key Laboratory of Science & Technology on Multi-Spectral Information Processing, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology, Wuhan 430074, ChinaObject detection in remote sensing images (RSIs) is one of the basic tasks in the field of remote sensing image automatic interpretation. In recent years, the deep object detection frameworks of natural scene images (NSIs) have been introduced into object detection on RSIs, and the detection performance has improved significantly because of the powerful feature representation. However, there are still many challenges concerning the particularities of remote sensing objects. One of the main challenges is the missed detection of small objects which have less than five percent of the pixels of the big objects. Generally, the existing algorithms choose to deal with this problem by multi-scale feature fusion based on a feature pyramid. However, the benefits of this strategy are limited, considering that the location of small objects in the feature map will disappear when the detection task is processed at the end of the network. In this study, we propose a subtask attention network (StAN), which handles the detection task directly on the shallow layer of the network. First, StAN contains one shared feature branch and two subtask attention branches of a semantic auxiliary subtask and a detection subtask based on the multi-task attention network (MTAN). Second, the detection branch uses only low-level features considering small objects. Third, the attention map guidance mechanism is put forward to optimize the network for keeping the identification ability. Fourth, the multi-dimensional sampling module (MdS), global multi-view channel weights (GMulW) and target-guided pixel attention (TPA) are designed for further improvement of the detection accuracy in complex scenes. The experimental results on the NWPU VHR-10 dataset and DOTA dataset demonstrated that the proposed algorithm achieved the SOTA performance, and the missed detection of small objects decreased. On the other hand, ablation experiments also proved the effects of MdS, GMulW and TPA.https://www.mdpi.com/2072-4292/13/10/1925remote sensingobject detectiondeep learningsubtaskattention |
| spellingShingle | Shengzhou Xiong Yihua Tan Yansheng Li Cai Wen Pei Yan Subtask Attention Based Object Detection in Remote Sensing Images Remote Sensing remote sensing object detection deep learning subtask attention |
| title | Subtask Attention Based Object Detection in Remote Sensing Images |
| title_full | Subtask Attention Based Object Detection in Remote Sensing Images |
| title_fullStr | Subtask Attention Based Object Detection in Remote Sensing Images |
| title_full_unstemmed | Subtask Attention Based Object Detection in Remote Sensing Images |
| title_short | Subtask Attention Based Object Detection in Remote Sensing Images |
| title_sort | subtask attention based object detection in remote sensing images |
| topic | remote sensing object detection deep learning subtask attention |
| url | https://www.mdpi.com/2072-4292/13/10/1925 |
| work_keys_str_mv | AT shengzhouxiong subtaskattentionbasedobjectdetectioninremotesensingimages AT yihuatan subtaskattentionbasedobjectdetectioninremotesensingimages AT yanshengli subtaskattentionbasedobjectdetectioninremotesensingimages AT caiwen subtaskattentionbasedobjectdetectioninremotesensingimages AT peiyan subtaskattentionbasedobjectdetectioninremotesensingimages |