Few-shot remote sensing scene classification based on multi subband deep feature fusion
Recently, convolutional neural networks (CNNs) have performed well in object classification and object recognition. However, due to the particularity of geographic data, the labeled samples are seriously insufficient, which limits the practical application of CNN methods in remote sensing (RS) image...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIMS Press
2023-06-01
|
| Series: | Mathematical Biosciences and Engineering |
| Subjects: | |
| Online Access: | https://www.aimspress.com/article/doi/10.3934/mbe.2023575?viewType=HTML |
| _version_ | 1797793971133480960 |
|---|---|
| author | Song Yang Huibin Wang Hongmin Gao Lili Zhang |
| author_facet | Song Yang Huibin Wang Hongmin Gao Lili Zhang |
| author_sort | Song Yang |
| collection | DOAJ |
| description | Recently, convolutional neural networks (CNNs) have performed well in object classification and object recognition. However, due to the particularity of geographic data, the labeled samples are seriously insufficient, which limits the practical application of CNN methods in remote sensing (RS) image processing. To address the problem of small sample RS image classification, a discrete wavelet-based multi-level deep feature fusion method is proposed. First, the deep features are extracted from the RS images using pre-trained deep CNNs and discrete wavelet transform (DWT) methods. Next, a modified discriminant correlation analysis (DCA) approach is proposed to distinguish easily confused categories effectively, which is based on the distance coefficient of between-class. The proposed approach can effectively integrate the deep feature information of various frequency bands. Thereby, the proposed method obtains the low-dimensional features with good discrimination, which is demonstrated through experiments on four benchmark datasets. Compared with several state-of-the-art methods, the proposed method achieves outstanding performance under limited training samples, especially one or two training samples per class. |
| first_indexed | 2024-03-13T02:55:58Z |
| format | Article |
| id | doaj.art-dd4487b5f5a24d80a574a2276c438074 |
| institution | Directory Open Access Journal |
| issn | 1551-0018 |
| language | English |
| last_indexed | 2024-03-13T02:55:58Z |
| publishDate | 2023-06-01 |
| publisher | AIMS Press |
| record_format | Article |
| series | Mathematical Biosciences and Engineering |
| spelling | doaj.art-dd4487b5f5a24d80a574a2276c4380742023-06-28T06:06:08ZengAIMS PressMathematical Biosciences and Engineering1551-00182023-06-01207128891290710.3934/mbe.2023575Few-shot remote sensing scene classification based on multi subband deep feature fusionSong Yang0Huibin Wang1Hongmin Gao2Lili Zhang 31. College of Computer and Information, Hohai University, Nanjing 211100, China 2. Faculty of Electronic Information Engineering, Huaiyin Institute of Technology, Huaian 223001, China1. College of Computer and Information, Hohai University, Nanjing 211100, China1. College of Computer and Information, Hohai University, Nanjing 211100, China1. College of Computer and Information, Hohai University, Nanjing 211100, ChinaRecently, convolutional neural networks (CNNs) have performed well in object classification and object recognition. However, due to the particularity of geographic data, the labeled samples are seriously insufficient, which limits the practical application of CNN methods in remote sensing (RS) image processing. To address the problem of small sample RS image classification, a discrete wavelet-based multi-level deep feature fusion method is proposed. First, the deep features are extracted from the RS images using pre-trained deep CNNs and discrete wavelet transform (DWT) methods. Next, a modified discriminant correlation analysis (DCA) approach is proposed to distinguish easily confused categories effectively, which is based on the distance coefficient of between-class. The proposed approach can effectively integrate the deep feature information of various frequency bands. Thereby, the proposed method obtains the low-dimensional features with good discrimination, which is demonstrated through experiments on four benchmark datasets. Compared with several state-of-the-art methods, the proposed method achieves outstanding performance under limited training samples, especially one or two training samples per class.https://www.aimspress.com/article/doi/10.3934/mbe.2023575?viewType=HTMLremote sensing scene classificationdeep feature fusiondiscriminant correlation analysisdiscrete wavelet transform |
| spellingShingle | Song Yang Huibin Wang Hongmin Gao Lili Zhang Few-shot remote sensing scene classification based on multi subband deep feature fusion Mathematical Biosciences and Engineering remote sensing scene classification deep feature fusion discriminant correlation analysis discrete wavelet transform |
| title | Few-shot remote sensing scene classification based on multi subband deep feature fusion |
| title_full | Few-shot remote sensing scene classification based on multi subband deep feature fusion |
| title_fullStr | Few-shot remote sensing scene classification based on multi subband deep feature fusion |
| title_full_unstemmed | Few-shot remote sensing scene classification based on multi subband deep feature fusion |
| title_short | Few-shot remote sensing scene classification based on multi subband deep feature fusion |
| title_sort | few shot remote sensing scene classification based on multi subband deep feature fusion |
| topic | remote sensing scene classification deep feature fusion discriminant correlation analysis discrete wavelet transform |
| url | https://www.aimspress.com/article/doi/10.3934/mbe.2023575?viewType=HTML |
| work_keys_str_mv | AT songyang fewshotremotesensingsceneclassificationbasedonmultisubbanddeepfeaturefusion AT huibinwang fewshotremotesensingsceneclassificationbasedonmultisubbanddeepfeaturefusion AT hongmingao fewshotremotesensingsceneclassificationbasedonmultisubbanddeepfeaturefusion AT lilizhang fewshotremotesensingsceneclassificationbasedonmultisubbanddeepfeaturefusion |