CSA-MSO3DCNN: Multiscale Octave 3D CNN with Channel and Spatial Attention for Hyperspectral Image Classification
3D convolutional neural networks (CNNs) have been demonstrated to be a powerful tool in hyperspectral images (HSIs) classification. However, using the conventional 3D CNNs to extract the spectral−spatial feature for HSIs results in too many parameters as HSIs have plenty of spatial redunda...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/12/1/188 |
| _version_ | 1819002531914711040 |
|---|---|
| author | Qin Xu Yong Xiao Dongyue Wang Bin Luo |
| author_facet | Qin Xu Yong Xiao Dongyue Wang Bin Luo |
| author_sort | Qin Xu |
| collection | DOAJ |
| description | 3D convolutional neural networks (CNNs) have been demonstrated to be a powerful tool in hyperspectral images (HSIs) classification. However, using the conventional 3D CNNs to extract the spectral−spatial feature for HSIs results in too many parameters as HSIs have plenty of spatial redundancy. To address this issue, in this paper, we first design multiscale convolution to extract the contextual feature of different scales for HSIs and then propose to employ the octave 3D CNN which factorizes the mixed feature maps by their frequency to replace the normal 3D CNN in order to reduce the spatial redundancy and enlarge the receptive field. To further explore the discriminative features, a channel attention module and a spatial attention module are adopted to optimize the feature maps and improve the classification performance. The experiments on four hyperspectral image data sets demonstrate that the proposed method outperforms other state-of-the-art deep learning methods. |
| first_indexed | 2024-12-20T23:06:36Z |
| format | Article |
| id | doaj.art-48aed36251a4484d8f80f2206669963d |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-12-20T23:06:36Z |
| publishDate | 2020-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-48aed36251a4484d8f80f2206669963d2022-12-21T19:23:51ZengMDPI AGRemote Sensing2072-42922020-01-0112118810.3390/rs12010188rs12010188CSA-MSO3DCNN: Multiscale Octave 3D CNN with Channel and Spatial Attention for Hyperspectral Image ClassificationQin Xu0Yong Xiao1Dongyue Wang2Bin Luo3Key Laboratory of Intelligent Computing and Signal Processing of Ministry of Education, School of Computer Science and Technology, Anhui University, Hefei 230601, ChinaKey Laboratory of Intelligent Computing and Signal Processing of Ministry of Education, School of Computer Science and Technology, Anhui University, Hefei 230601, ChinaKey Laboratory of Intelligent Computing and Signal Processing of Ministry of Education, School of Computer Science and Technology, Anhui University, Hefei 230601, ChinaKey Laboratory of Intelligent Computing and Signal Processing of Ministry of Education, School of Computer Science and Technology, Anhui University, Hefei 230601, China3D convolutional neural networks (CNNs) have been demonstrated to be a powerful tool in hyperspectral images (HSIs) classification. However, using the conventional 3D CNNs to extract the spectral−spatial feature for HSIs results in too many parameters as HSIs have plenty of spatial redundancy. To address this issue, in this paper, we first design multiscale convolution to extract the contextual feature of different scales for HSIs and then propose to employ the octave 3D CNN which factorizes the mixed feature maps by their frequency to replace the normal 3D CNN in order to reduce the spatial redundancy and enlarge the receptive field. To further explore the discriminative features, a channel attention module and a spatial attention module are adopted to optimize the feature maps and improve the classification performance. The experiments on four hyperspectral image data sets demonstrate that the proposed method outperforms other state-of-the-art deep learning methods.https://www.mdpi.com/2072-4292/12/1/188hypersctral image classificationoctave convolutionfeature extractionchannel and spatial attention |
| spellingShingle | Qin Xu Yong Xiao Dongyue Wang Bin Luo CSA-MSO3DCNN: Multiscale Octave 3D CNN with Channel and Spatial Attention for Hyperspectral Image Classification Remote Sensing hypersctral image classification octave convolution feature extraction channel and spatial attention |
| title | CSA-MSO3DCNN: Multiscale Octave 3D CNN with Channel and Spatial Attention for Hyperspectral Image Classification |
| title_full | CSA-MSO3DCNN: Multiscale Octave 3D CNN with Channel and Spatial Attention for Hyperspectral Image Classification |
| title_fullStr | CSA-MSO3DCNN: Multiscale Octave 3D CNN with Channel and Spatial Attention for Hyperspectral Image Classification |
| title_full_unstemmed | CSA-MSO3DCNN: Multiscale Octave 3D CNN with Channel and Spatial Attention for Hyperspectral Image Classification |
| title_short | CSA-MSO3DCNN: Multiscale Octave 3D CNN with Channel and Spatial Attention for Hyperspectral Image Classification |
| title_sort | csa mso3dcnn multiscale octave 3d cnn with channel and spatial attention for hyperspectral image classification |
| topic | hypersctral image classification octave convolution feature extraction channel and spatial attention |
| url | https://www.mdpi.com/2072-4292/12/1/188 |
| work_keys_str_mv | AT qinxu csamso3dcnnmultiscaleoctave3dcnnwithchannelandspatialattentionforhyperspectralimageclassification AT yongxiao csamso3dcnnmultiscaleoctave3dcnnwithchannelandspatialattentionforhyperspectralimageclassification AT dongyuewang csamso3dcnnmultiscaleoctave3dcnnwithchannelandspatialattentionforhyperspectralimageclassification AT binluo csamso3dcnnmultiscaleoctave3dcnnwithchannelandspatialattentionforhyperspectralimageclassification |