Semantic Segmentation for Buildings of Large Intra-Class Variation in Remote Sensing Images with O-GAN
Remote sensing building extraction is of great importance to many applications, such as urban planning and economic status assessment. Deep learning with deep network structures and back-propagation optimization can automatically learn features of targets in high-resolution remote sensing images. Ho...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/13/3/475 |
| _version_ | 1827596944135946240 |
|---|---|
| author | Shuting Sun Lin Mu Lizhe Wang Peng Liu Xiaolei Liu Yuwei Zhang |
| author_facet | Shuting Sun Lin Mu Lizhe Wang Peng Liu Xiaolei Liu Yuwei Zhang |
| author_sort | Shuting Sun |
| collection | DOAJ |
| description | Remote sensing building extraction is of great importance to many applications, such as urban planning and economic status assessment. Deep learning with deep network structures and back-propagation optimization can automatically learn features of targets in high-resolution remote sensing images. However, it is also obvious that the generalizability of deep networks is almost entirely dependent on the quality and quantity of the labels. Therefore, building extraction performances will be greatly affected if there is a large intra-class variation among samples of one class target. To solve the problem, a subdivision method for reducing intra-class differences is proposed to enhance semantic segmentation. We proposed that backgrounds and targets be separately generated by two orthogonal generative adversarial networks (O-GAN). The two O-GANs are connected by adding the new loss function to their discriminators. To better extract building features, drawing on the idea of fine-grained image classification, feature vectors for a target are obtained through an intermediate convolution layer of O-GAN with selective convolutional descriptor aggregation (SCDA). Subsequently, feature vectors are clustered into new, different subdivisions to train semantic segmentation networks. In the prediction stages, the subdivisions will be merged into one class. Experiments were conducted with remote sensing images of the Tibet area, where there are both tall buildings and herdsmen’s tents. The results indicate that, compared with direct semantic segmentation, the proposed subdivision method can make an improvement on accuracy of about 4%. Besides, statistics and visualizing building features validated the rationality of features and subdivisions. |
| first_indexed | 2024-03-09T03:18:40Z |
| format | Article |
| id | doaj.art-30eae9f6f31a44baaf8cba6ec0a87594 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-09T03:18:40Z |
| publishDate | 2021-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-30eae9f6f31a44baaf8cba6ec0a875942023-12-03T15:13:11ZengMDPI AGRemote Sensing2072-42922021-01-0113347510.3390/rs13030475Semantic Segmentation for Buildings of Large Intra-Class Variation in Remote Sensing Images with O-GANShuting Sun0Lin Mu1Lizhe Wang2Peng Liu3Xiaolei Liu4Yuwei Zhang5Collage of Marine Science and Technology, China University of Geosciences (CUG), Wuhan 430074, ChinaCollege of Life Sciences and Oceanography, Shenzhen University, Shenzhen 518060, ChinaSchool of Computer Science, China University of Geosciences (CUG), Wuhan 430074, ChinaAerospace Information Research Institute, CAS, Beijing 100094, ChinaSchool of Information Science and Engineering, Yanshan University, Qinhuangdao 066012, ChinaAerospace Information Research Institute, CAS, Beijing 100094, ChinaRemote sensing building extraction is of great importance to many applications, such as urban planning and economic status assessment. Deep learning with deep network structures and back-propagation optimization can automatically learn features of targets in high-resolution remote sensing images. However, it is also obvious that the generalizability of deep networks is almost entirely dependent on the quality and quantity of the labels. Therefore, building extraction performances will be greatly affected if there is a large intra-class variation among samples of one class target. To solve the problem, a subdivision method for reducing intra-class differences is proposed to enhance semantic segmentation. We proposed that backgrounds and targets be separately generated by two orthogonal generative adversarial networks (O-GAN). The two O-GANs are connected by adding the new loss function to their discriminators. To better extract building features, drawing on the idea of fine-grained image classification, feature vectors for a target are obtained through an intermediate convolution layer of O-GAN with selective convolutional descriptor aggregation (SCDA). Subsequently, feature vectors are clustered into new, different subdivisions to train semantic segmentation networks. In the prediction stages, the subdivisions will be merged into one class. Experiments were conducted with remote sensing images of the Tibet area, where there are both tall buildings and herdsmen’s tents. The results indicate that, compared with direct semantic segmentation, the proposed subdivision method can make an improvement on accuracy of about 4%. Besides, statistics and visualizing building features validated the rationality of features and subdivisions.https://www.mdpi.com/2072-4292/13/3/475building extractionGF-2orthogonal generative adversarial networkssubdivision |
| spellingShingle | Shuting Sun Lin Mu Lizhe Wang Peng Liu Xiaolei Liu Yuwei Zhang Semantic Segmentation for Buildings of Large Intra-Class Variation in Remote Sensing Images with O-GAN Remote Sensing building extraction GF-2 orthogonal generative adversarial networks subdivision |
| title | Semantic Segmentation for Buildings of Large Intra-Class Variation in Remote Sensing Images with O-GAN |
| title_full | Semantic Segmentation for Buildings of Large Intra-Class Variation in Remote Sensing Images with O-GAN |
| title_fullStr | Semantic Segmentation for Buildings of Large Intra-Class Variation in Remote Sensing Images with O-GAN |
| title_full_unstemmed | Semantic Segmentation for Buildings of Large Intra-Class Variation in Remote Sensing Images with O-GAN |
| title_short | Semantic Segmentation for Buildings of Large Intra-Class Variation in Remote Sensing Images with O-GAN |
| title_sort | semantic segmentation for buildings of large intra class variation in remote sensing images with o gan |
| topic | building extraction GF-2 orthogonal generative adversarial networks subdivision |
| url | https://www.mdpi.com/2072-4292/13/3/475 |
| work_keys_str_mv | AT shutingsun semanticsegmentationforbuildingsoflargeintraclassvariationinremotesensingimageswithogan AT linmu semanticsegmentationforbuildingsoflargeintraclassvariationinremotesensingimageswithogan AT lizhewang semanticsegmentationforbuildingsoflargeintraclassvariationinremotesensingimageswithogan AT pengliu semanticsegmentationforbuildingsoflargeintraclassvariationinremotesensingimageswithogan AT xiaoleiliu semanticsegmentationforbuildingsoflargeintraclassvariationinremotesensingimageswithogan AT yuweizhang semanticsegmentationforbuildingsoflargeintraclassvariationinremotesensingimageswithogan |