RFA-Net: Reconstructed Feature Alignment Network for Domain Adaptation Object Detection in Remote Sensing Imagery
With the development of deep learning, great progress has been made in object detection of remote sensing (RS) imagery. However, the object detector is hard to generalize well from one labeled dataset (source domain) to another unlabeled dataset (target domain) due to the discrepancy of data distrib...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9829266/ |
| _version_ | 1818041683489587200 |
|---|---|
| author | Yangguang Zhu Xian Sun Wenhui Diao Hao Li Kun Fu |
| author_facet | Yangguang Zhu Xian Sun Wenhui Diao Hao Li Kun Fu |
| author_sort | Yangguang Zhu |
| collection | DOAJ |
| description | With the development of deep learning, great progress has been made in object detection of remote sensing (RS) imagery. However, the object detector is hard to generalize well from one labeled dataset (source domain) to another unlabeled dataset (target domain) due to the discrepancy of data distribution (domain shift). Currently, adversarial-based domain adaptation methods align the semantic features of source and target domain features to alleviate the domain shift. But they fail to avoid the alignment of noisy background features and neglect the instance-level features, which are inappropriate for detection models that focus on instance location and classification. To mitigate domain shift existing in object detection, we propose a reconstructed feature alignment network (RFA-Net) for unsupervised cross-domain object detection in RS imagery. The RFA-Net includes one sequential data augmentation module deployed on data level for providing solid gains on unlabeled data, one sparse feature reconstruction module deployed on feature level to intensify instance feature for feature alignment, and one pseudo-label generation module deployed on label level for the supervision of the unlabeled target domain. Extensive experiments illustrate that our proposed RFA-Net is effective to alleviate the domain shift problem in domain adaptation object detection of RS imagery. |
| first_indexed | 2024-12-10T08:34:19Z |
| format | Article |
| id | doaj.art-16896e4f2b0f4c9499e8d6a478dbaa60 |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-12-10T08:34:19Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-16896e4f2b0f4c9499e8d6a478dbaa602022-12-22T01:56:01ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352022-01-01155689570310.1109/JSTARS.2022.31906999829266RFA-Net: Reconstructed Feature Alignment Network for Domain Adaptation Object Detection in Remote Sensing ImageryYangguang Zhu0https://orcid.org/0000-0001-7598-4575Xian Sun1https://orcid.org/0000-0002-0038-9816Wenhui Diao2https://orcid.org/0000-0002-3931-3974Hao Li3Kun Fu4https://orcid.org/0000-0002-0450-6469Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaWith the development of deep learning, great progress has been made in object detection of remote sensing (RS) imagery. However, the object detector is hard to generalize well from one labeled dataset (source domain) to another unlabeled dataset (target domain) due to the discrepancy of data distribution (domain shift). Currently, adversarial-based domain adaptation methods align the semantic features of source and target domain features to alleviate the domain shift. But they fail to avoid the alignment of noisy background features and neglect the instance-level features, which are inappropriate for detection models that focus on instance location and classification. To mitigate domain shift existing in object detection, we propose a reconstructed feature alignment network (RFA-Net) for unsupervised cross-domain object detection in RS imagery. The RFA-Net includes one sequential data augmentation module deployed on data level for providing solid gains on unlabeled data, one sparse feature reconstruction module deployed on feature level to intensify instance feature for feature alignment, and one pseudo-label generation module deployed on label level for the supervision of the unlabeled target domain. Extensive experiments illustrate that our proposed RFA-Net is effective to alleviate the domain shift problem in domain adaptation object detection of RS imagery.https://ieeexplore.ieee.org/document/9829266/Data augmentationdomain adaptationfeature reconstructionobject detectionpseudo-label filtering |
| spellingShingle | Yangguang Zhu Xian Sun Wenhui Diao Hao Li Kun Fu RFA-Net: Reconstructed Feature Alignment Network for Domain Adaptation Object Detection in Remote Sensing Imagery IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Data augmentation domain adaptation feature reconstruction object detection pseudo-label filtering |
| title | RFA-Net: Reconstructed Feature Alignment Network for Domain Adaptation Object Detection in Remote Sensing Imagery |
| title_full | RFA-Net: Reconstructed Feature Alignment Network for Domain Adaptation Object Detection in Remote Sensing Imagery |
| title_fullStr | RFA-Net: Reconstructed Feature Alignment Network for Domain Adaptation Object Detection in Remote Sensing Imagery |
| title_full_unstemmed | RFA-Net: Reconstructed Feature Alignment Network for Domain Adaptation Object Detection in Remote Sensing Imagery |
| title_short | RFA-Net: Reconstructed Feature Alignment Network for Domain Adaptation Object Detection in Remote Sensing Imagery |
| title_sort | rfa net reconstructed feature alignment network for domain adaptation object detection in remote sensing imagery |
| topic | Data augmentation domain adaptation feature reconstruction object detection pseudo-label filtering |
| url | https://ieeexplore.ieee.org/document/9829266/ |
| work_keys_str_mv | AT yangguangzhu rfanetreconstructedfeaturealignmentnetworkfordomainadaptationobjectdetectioninremotesensingimagery AT xiansun rfanetreconstructedfeaturealignmentnetworkfordomainadaptationobjectdetectioninremotesensingimagery AT wenhuidiao rfanetreconstructedfeaturealignmentnetworkfordomainadaptationobjectdetectioninremotesensingimagery AT haoli rfanetreconstructedfeaturealignmentnetworkfordomainadaptationobjectdetectioninremotesensingimagery AT kunfu rfanetreconstructedfeaturealignmentnetworkfordomainadaptationobjectdetectioninremotesensingimagery |