IMPROVING SEMANTIC SEGMENTATION OF HIGH-RESOLUTION REMOTE SENSING IMAGES USING WASSERSTEIN GENERATIVE ADVERSARIAL NETWORK
Semantic segmentation of remote sensing images with high spatial resolution has many applications in a wide range of problems in this field. In recent years, the use of advanced techniques based on fully convolutional neural networks have achieved high and impressive accuracies. However, the labels...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2023-01-01
|
| Series: | The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| Online Access: | https://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLVIII-4-W2-2022/45/2023/isprs-archives-XLVIII-4-W2-2022-45-2023.pdf |
| _version_ | 1797954224207691776 |
|---|---|
| author | H. R. Hosseinpour F. Samadzadegan F. Dadrass Javan F. Dadrass Javan S. Motayyeb |
| author_facet | H. R. Hosseinpour F. Samadzadegan F. Dadrass Javan F. Dadrass Javan S. Motayyeb |
| author_sort | H. R. Hosseinpour |
| collection | DOAJ |
| description | Semantic segmentation of remote sensing images with high spatial resolution has many applications in a wide range of problems in this field. In recent years, the use of advanced techniques based on fully convolutional neural networks have achieved high and impressive accuracies. However, the labels of different classes are estimated independently in this method. In general, the segmentation effect is too coarse to take the relationship between pixels into account. On the other hand, due to the use of convolution filters and limitations of calculations, the field of view information of these filters will be limited in deep layers. In this study, a method based on generative adversarial network (GAN) is proposed to strengthen spatial vicinity in the output segmentation map. The segmentation model receive assistance from the GAN model in the form of a higher order potential loss. Furthermore, for better stability and performance in model training the Wasserstein GAN is used for optimization of the model. We successfully show an increase in semantic segmentation accuracy using the challenging ISPRS Vaihingen benchmark dataset. |
| first_indexed | 2024-04-10T23:14:17Z |
| format | Article |
| id | doaj.art-442f49d32204449bba7366b9c445dee4 |
| institution | Directory Open Access Journal |
| issn | 1682-1750 2194-9034 |
| language | English |
| last_indexed | 2024-04-10T23:14:17Z |
| publishDate | 2023-01-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| spelling | doaj.art-442f49d32204449bba7366b9c445dee42023-01-12T21:55:52ZengCopernicus PublicationsThe International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences1682-17502194-90342023-01-01XLVIII-4-W2-2022455110.5194/isprs-archives-XLVIII-4-W2-2022-45-2023IMPROVING SEMANTIC SEGMENTATION OF HIGH-RESOLUTION REMOTE SENSING IMAGES USING WASSERSTEIN GENERATIVE ADVERSARIAL NETWORKH. R. Hosseinpour0F. Samadzadegan1F. Dadrass Javan2F. Dadrass Javan3S. Motayyeb4School of Surveying and Geospatial Information Engineering, College of Engineering, University of Tehran, Tehran, IranSchool of Surveying and Geospatial Information Engineering, College of Engineering, University of Tehran, Tehran, IranSchool of Surveying and Geospatial Information Engineering, College of Engineering, University of Tehran, Tehran, IranFaculty of Geo-Information Science and Earth Observation (ITC), University of Twente, 7522 NB Enschede, the NetherlandsSchool of Surveying and Geospatial Information Engineering, College of Engineering, University of Tehran, Tehran, IranSemantic segmentation of remote sensing images with high spatial resolution has many applications in a wide range of problems in this field. In recent years, the use of advanced techniques based on fully convolutional neural networks have achieved high and impressive accuracies. However, the labels of different classes are estimated independently in this method. In general, the segmentation effect is too coarse to take the relationship between pixels into account. On the other hand, due to the use of convolution filters and limitations of calculations, the field of view information of these filters will be limited in deep layers. In this study, a method based on generative adversarial network (GAN) is proposed to strengthen spatial vicinity in the output segmentation map. The segmentation model receive assistance from the GAN model in the form of a higher order potential loss. Furthermore, for better stability and performance in model training the Wasserstein GAN is used for optimization of the model. We successfully show an increase in semantic segmentation accuracy using the challenging ISPRS Vaihingen benchmark dataset.https://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLVIII-4-W2-2022/45/2023/isprs-archives-XLVIII-4-W2-2022-45-2023.pdf |
| spellingShingle | H. R. Hosseinpour F. Samadzadegan F. Dadrass Javan F. Dadrass Javan S. Motayyeb IMPROVING SEMANTIC SEGMENTATION OF HIGH-RESOLUTION REMOTE SENSING IMAGES USING WASSERSTEIN GENERATIVE ADVERSARIAL NETWORK The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| title | IMPROVING SEMANTIC SEGMENTATION OF HIGH-RESOLUTION REMOTE SENSING IMAGES USING WASSERSTEIN GENERATIVE ADVERSARIAL NETWORK |
| title_full | IMPROVING SEMANTIC SEGMENTATION OF HIGH-RESOLUTION REMOTE SENSING IMAGES USING WASSERSTEIN GENERATIVE ADVERSARIAL NETWORK |
| title_fullStr | IMPROVING SEMANTIC SEGMENTATION OF HIGH-RESOLUTION REMOTE SENSING IMAGES USING WASSERSTEIN GENERATIVE ADVERSARIAL NETWORK |
| title_full_unstemmed | IMPROVING SEMANTIC SEGMENTATION OF HIGH-RESOLUTION REMOTE SENSING IMAGES USING WASSERSTEIN GENERATIVE ADVERSARIAL NETWORK |
| title_short | IMPROVING SEMANTIC SEGMENTATION OF HIGH-RESOLUTION REMOTE SENSING IMAGES USING WASSERSTEIN GENERATIVE ADVERSARIAL NETWORK |
| title_sort | improving semantic segmentation of high resolution remote sensing images using wasserstein generative adversarial network |
| url | https://www.int-arch-photogramm-remote-sens-spatial-inf-sci.net/XLVIII-4-W2-2022/45/2023/isprs-archives-XLVIII-4-W2-2022-45-2023.pdf |
| work_keys_str_mv | AT hrhosseinpour improvingsemanticsegmentationofhighresolutionremotesensingimagesusingwassersteingenerativeadversarialnetwork AT fsamadzadegan improvingsemanticsegmentationofhighresolutionremotesensingimagesusingwassersteingenerativeadversarialnetwork AT fdadrassjavan improvingsemanticsegmentationofhighresolutionremotesensingimagesusingwassersteingenerativeadversarialnetwork AT fdadrassjavan improvingsemanticsegmentationofhighresolutionremotesensingimagesusingwassersteingenerativeadversarialnetwork AT smotayyeb improvingsemanticsegmentationofhighresolutionremotesensingimagesusingwassersteingenerativeadversarialnetwork |