SEMI-SUPERVISED SEMANTIC SEGMENTATION NETWORK VIA LEARNING CONSISTENCY FOR REMOTE SENSING LAND-COVER CLASSIFICATION
Current popular deep neural networks for semantic segmentation are almost supervised and highly rely on a large amount of labeled data. However, obtaining a large amount of pixel-level labeled data is time-consuming and laborious. In remote sensing area, this problem is more urgent. To alleviate thi...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2020-08-01
|
| Series: | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| Online Access: | https://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/V-2-2020/609/2020/isprs-annals-V-2-2020-609-2020.pdf |
| _version_ | 1828352796079947776 |
|---|---|
| author | B. Zhang Y. Zhang Y. Li Y. Wan F. Wen |
| author_facet | B. Zhang Y. Zhang Y. Li Y. Wan F. Wen |
| author_sort | B. Zhang |
| collection | DOAJ |
| description | Current popular deep neural networks for semantic segmentation are almost supervised and highly rely on a large amount of labeled data. However, obtaining a large amount of pixel-level labeled data is time-consuming and laborious. In remote sensing area, this problem is more urgent. To alleviate this problem, we propose a novel semantic segmentation neural network (S4Net) based on semi-supervised learning by using unlabeled data. Our model can learn from unlabeled data by consistency regularization, which enforces the consistency of output under different random transforms and perturbations, such as random affine transform. Thus, the network is trained by the weighted sum of a supervised loss from labeled data and a consistency regularization loss from unlabeled data. The experiments we conducted on DeepGlobe land cover classification challenge dataset verified that our network can make use of unlabeled data to obtain precise results of semantic segmentation and achieve competitive performance when compared to other methods. |
| first_indexed | 2024-04-14T01:58:09Z |
| format | Article |
| id | doaj.art-8d4a101afe3346f99e011fa39c132385 |
| institution | Directory Open Access Journal |
| issn | 2194-9042 2194-9050 |
| language | English |
| last_indexed | 2024-04-14T01:58:09Z |
| publishDate | 2020-08-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| spelling | doaj.art-8d4a101afe3346f99e011fa39c1323852022-12-22T02:18:57ZengCopernicus PublicationsISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences2194-90422194-90502020-08-01V-2-202060961510.5194/isprs-annals-V-2-2020-609-2020SEMI-SUPERVISED SEMANTIC SEGMENTATION NETWORK VIA LEARNING CONSISTENCY FOR REMOTE SENSING LAND-COVER CLASSIFICATIONB. Zhang0Y. Zhang1Y. Li2Y. Wan3F. Wen4School of Remote Sensing and Information Engineering, Wuhan University, Wuhan, Hubei, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, Hubei, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, Hubei, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, Hubei, ChinaSchool of Remote Sensing and Information Engineering, Wuhan University, Wuhan, Hubei, ChinaCurrent popular deep neural networks for semantic segmentation are almost supervised and highly rely on a large amount of labeled data. However, obtaining a large amount of pixel-level labeled data is time-consuming and laborious. In remote sensing area, this problem is more urgent. To alleviate this problem, we propose a novel semantic segmentation neural network (S4Net) based on semi-supervised learning by using unlabeled data. Our model can learn from unlabeled data by consistency regularization, which enforces the consistency of output under different random transforms and perturbations, such as random affine transform. Thus, the network is trained by the weighted sum of a supervised loss from labeled data and a consistency regularization loss from unlabeled data. The experiments we conducted on DeepGlobe land cover classification challenge dataset verified that our network can make use of unlabeled data to obtain precise results of semantic segmentation and achieve competitive performance when compared to other methods.https://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/V-2-2020/609/2020/isprs-annals-V-2-2020-609-2020.pdf |
| spellingShingle | B. Zhang Y. Zhang Y. Li Y. Wan F. Wen SEMI-SUPERVISED SEMANTIC SEGMENTATION NETWORK VIA LEARNING CONSISTENCY FOR REMOTE SENSING LAND-COVER CLASSIFICATION ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| title | SEMI-SUPERVISED SEMANTIC SEGMENTATION NETWORK VIA LEARNING CONSISTENCY FOR REMOTE SENSING LAND-COVER CLASSIFICATION |
| title_full | SEMI-SUPERVISED SEMANTIC SEGMENTATION NETWORK VIA LEARNING CONSISTENCY FOR REMOTE SENSING LAND-COVER CLASSIFICATION |
| title_fullStr | SEMI-SUPERVISED SEMANTIC SEGMENTATION NETWORK VIA LEARNING CONSISTENCY FOR REMOTE SENSING LAND-COVER CLASSIFICATION |
| title_full_unstemmed | SEMI-SUPERVISED SEMANTIC SEGMENTATION NETWORK VIA LEARNING CONSISTENCY FOR REMOTE SENSING LAND-COVER CLASSIFICATION |
| title_short | SEMI-SUPERVISED SEMANTIC SEGMENTATION NETWORK VIA LEARNING CONSISTENCY FOR REMOTE SENSING LAND-COVER CLASSIFICATION |
| title_sort | semi supervised semantic segmentation network via learning consistency for remote sensing land cover classification |
| url | https://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/V-2-2020/609/2020/isprs-annals-V-2-2020-609-2020.pdf |
| work_keys_str_mv | AT bzhang semisupervisedsemanticsegmentationnetworkvialearningconsistencyforremotesensinglandcoverclassification AT yzhang semisupervisedsemanticsegmentationnetworkvialearningconsistencyforremotesensinglandcoverclassification AT yli semisupervisedsemanticsegmentationnetworkvialearningconsistencyforremotesensinglandcoverclassification AT ywan semisupervisedsemanticsegmentationnetworkvialearningconsistencyforremotesensinglandcoverclassification AT fwen semisupervisedsemanticsegmentationnetworkvialearningconsistencyforremotesensinglandcoverclassification |