Relationship Prior and Adaptive Knowledge Mimic Based Compressed Deep Network for Aerial Scene Classification
The aerial scene classification is one of the major tasks in the remote sensing community that automatically labels the corresponding semantic categories of aerial images. Recently, a lot of methods based on deep neural networks have been proposed, in which hierarchical internal feature are extracte...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8782458/ |
| _version_ | 1818911289150275584 |
|---|---|
| author | Chu He Bokun He Xiaohuo Yin Wenwei Wang Mingsheng Liao |
| author_facet | Chu He Bokun He Xiaohuo Yin Wenwei Wang Mingsheng Liao |
| author_sort | Chu He |
| collection | DOAJ |
| description | The aerial scene classification is one of the major tasks in the remote sensing community that automatically labels the corresponding semantic categories of aerial images. Recently, a lot of methods based on deep neural networks have been proposed, in which hierarchical internal feature are extracted for representations. However, these presented methods often have complex structures and require large volume of memory, and a large number of labeled aerial scene images are difficult to obtain, hindering their implementation in practical applications. In this paper, we present the between-class similarity priori and adaptive knowledge mimic (BPKM) method for aerial scene classification. First, the method extracts the efficient prior relationship information of the scene images from large-scale network. Then, a compressed network is generated through learning the output and the intermediate representations of the large-scale network, and the compressed network achieves better feature description ability; in addition, an improved cross-entropy method with an adaptive threshold is applied to reduce the training time consumption. A largescale data set (AID) and UC-Merced data set are considered for performance evaluation, and the experimental results indicate that the proposed method is about 24× parameters saving compared to popular networks, e.g., AlexNet, and has outstanding classification performance in classes with similar features. |
| first_indexed | 2024-12-19T22:56:20Z |
| format | Article |
| id | doaj.art-dcd6fdd759764a7c9adbf2eb2be29124 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-19T22:56:20Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-dcd6fdd759764a7c9adbf2eb2be291242022-12-21T20:02:38ZengIEEEIEEE Access2169-35362019-01-01713708013708910.1109/ACCESS.2019.29322298782458Relationship Prior and Adaptive Knowledge Mimic Based Compressed Deep Network for Aerial Scene ClassificationChu He0https://orcid.org/0000-0003-3662-5769Bokun He1Xiaohuo Yin2Wenwei Wang3Mingsheng Liao4Electronic and Information School, Wuhan University, Wuhan, ChinaElectronic and Information School, Wuhan University, Wuhan, ChinaElectronic and Information School, Wuhan University, Wuhan, ChinaElectronic and Information School, Wuhan University, Wuhan, ChinaCollaborative Innovation Center for Geospatial Technology, Wuhan, ChinaThe aerial scene classification is one of the major tasks in the remote sensing community that automatically labels the corresponding semantic categories of aerial images. Recently, a lot of methods based on deep neural networks have been proposed, in which hierarchical internal feature are extracted for representations. However, these presented methods often have complex structures and require large volume of memory, and a large number of labeled aerial scene images are difficult to obtain, hindering their implementation in practical applications. In this paper, we present the between-class similarity priori and adaptive knowledge mimic (BPKM) method for aerial scene classification. First, the method extracts the efficient prior relationship information of the scene images from large-scale network. Then, a compressed network is generated through learning the output and the intermediate representations of the large-scale network, and the compressed network achieves better feature description ability; in addition, an improved cross-entropy method with an adaptive threshold is applied to reduce the training time consumption. A largescale data set (AID) and UC-Merced data set are considered for performance evaluation, and the experimental results indicate that the proposed method is about 24× parameters saving compared to popular networks, e.g., AlexNet, and has outstanding classification performance in classes with similar features.https://ieeexplore.ieee.org/document/8782458/Aerial scene classificationdeep learningknowledge transfermodel compressionpriori knowledge |
| spellingShingle | Chu He Bokun He Xiaohuo Yin Wenwei Wang Mingsheng Liao Relationship Prior and Adaptive Knowledge Mimic Based Compressed Deep Network for Aerial Scene Classification IEEE Access Aerial scene classification deep learning knowledge transfer model compression priori knowledge |
| title | Relationship Prior and Adaptive Knowledge Mimic Based Compressed Deep Network for Aerial Scene Classification |
| title_full | Relationship Prior and Adaptive Knowledge Mimic Based Compressed Deep Network for Aerial Scene Classification |
| title_fullStr | Relationship Prior and Adaptive Knowledge Mimic Based Compressed Deep Network for Aerial Scene Classification |
| title_full_unstemmed | Relationship Prior and Adaptive Knowledge Mimic Based Compressed Deep Network for Aerial Scene Classification |
| title_short | Relationship Prior and Adaptive Knowledge Mimic Based Compressed Deep Network for Aerial Scene Classification |
| title_sort | relationship prior and adaptive knowledge mimic based compressed deep network for aerial scene classification |
| topic | Aerial scene classification deep learning knowledge transfer model compression priori knowledge |
| url | https://ieeexplore.ieee.org/document/8782458/ |
| work_keys_str_mv | AT chuhe relationshippriorandadaptiveknowledgemimicbasedcompresseddeepnetworkforaerialsceneclassification AT bokunhe relationshippriorandadaptiveknowledgemimicbasedcompresseddeepnetworkforaerialsceneclassification AT xiaohuoyin relationshippriorandadaptiveknowledgemimicbasedcompresseddeepnetworkforaerialsceneclassification AT wenweiwang relationshippriorandadaptiveknowledgemimicbasedcompresseddeepnetworkforaerialsceneclassification AT mingshengliao relationshippriorandadaptiveknowledgemimicbasedcompresseddeepnetworkforaerialsceneclassification |