Hybrid Convolutional Network Combining Multiscale 3D Depthwise Separable Convolution and CBAM Residual Dilated Convolution for Hyperspectral Image Classification
In recent years, convolutional neural networks (CNNs) have been increasingly leveraged for the classification of hyperspectral imagery, displaying notable advancements. To address the issues of insufficient spectral and spatial information extraction and high computational complexity in hyperspectra...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-10-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/19/4796 |
| _version_ | 1827722253923516416 |
|---|---|
| author | Yicheng Hu Shufang Tian Jia Ge |
| author_facet | Yicheng Hu Shufang Tian Jia Ge |
| author_sort | Yicheng Hu |
| collection | DOAJ |
| description | In recent years, convolutional neural networks (CNNs) have been increasingly leveraged for the classification of hyperspectral imagery, displaying notable advancements. To address the issues of insufficient spectral and spatial information extraction and high computational complexity in hyperspectral image classification, we introduce the MDRDNet, an integrated neural network model. This novel architecture is comprised of two main components: a Multiscale 3D Depthwise Separable Convolutional Network and a CBAM-augmented Residual Dilated Convolutional Network. The first component employs depthwise separable convolutions in a 3D setting to efficiently capture spatial–spectral characteristics, thus substantially reducing the computational burden associated with 3D convolutions. Meanwhile, the second component enhances the network by integrating the Convolutional Block Attention Module (CBAM) with dilated convolutions via residual connections, effectively counteracting the issue of model degradation. We have empirically evaluated the MDRDNet’s performance by running comprehensive experiments on three publicly available datasets: Indian Pines, Pavia University, and Salinas. Our findings indicate that the overall accuracy of the MDRDNet on the three datasets reached 98.83%, 99.81%, and 99.99%, respectively, which is higher than the accuracy of existing models. Therefore, the MDRDNet proposed in this study can fully extract spatial–spectral joint information, providing a new idea for solving the problem of large model calculations in 3D convolutions. |
| first_indexed | 2024-03-10T21:35:27Z |
| format | Article |
| id | doaj.art-2e505d394d874190a82c735a26066e48 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T21:35:27Z |
| publishDate | 2023-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-2e505d394d874190a82c735a26066e482023-11-19T15:00:05ZengMDPI AGRemote Sensing2072-42922023-10-011519479610.3390/rs15194796Hybrid Convolutional Network Combining Multiscale 3D Depthwise Separable Convolution and CBAM Residual Dilated Convolution for Hyperspectral Image ClassificationYicheng Hu0Shufang Tian1Jia Ge2School of Earth Sciences and Resources, China University of Geosciences (Beijing), Beijing 100083, ChinaSchool of Earth Sciences and Resources, China University of Geosciences (Beijing), Beijing 100083, ChinaOil and Gas Resources Investigation Center of China Geological Survey, Beijing 100083, ChinaIn recent years, convolutional neural networks (CNNs) have been increasingly leveraged for the classification of hyperspectral imagery, displaying notable advancements. To address the issues of insufficient spectral and spatial information extraction and high computational complexity in hyperspectral image classification, we introduce the MDRDNet, an integrated neural network model. This novel architecture is comprised of two main components: a Multiscale 3D Depthwise Separable Convolutional Network and a CBAM-augmented Residual Dilated Convolutional Network. The first component employs depthwise separable convolutions in a 3D setting to efficiently capture spatial–spectral characteristics, thus substantially reducing the computational burden associated with 3D convolutions. Meanwhile, the second component enhances the network by integrating the Convolutional Block Attention Module (CBAM) with dilated convolutions via residual connections, effectively counteracting the issue of model degradation. We have empirically evaluated the MDRDNet’s performance by running comprehensive experiments on three publicly available datasets: Indian Pines, Pavia University, and Salinas. Our findings indicate that the overall accuracy of the MDRDNet on the three datasets reached 98.83%, 99.81%, and 99.99%, respectively, which is higher than the accuracy of existing models. Therefore, the MDRDNet proposed in this study can fully extract spatial–spectral joint information, providing a new idea for solving the problem of large model calculations in 3D convolutions.https://www.mdpi.com/2072-4292/15/19/4796convolutional neural networkshyperspectral image classificationdepthwise separableCBAMMDRDNet |
| spellingShingle | Yicheng Hu Shufang Tian Jia Ge Hybrid Convolutional Network Combining Multiscale 3D Depthwise Separable Convolution and CBAM Residual Dilated Convolution for Hyperspectral Image Classification Remote Sensing convolutional neural networks hyperspectral image classification depthwise separable CBAM MDRDNet |
| title | Hybrid Convolutional Network Combining Multiscale 3D Depthwise Separable Convolution and CBAM Residual Dilated Convolution for Hyperspectral Image Classification |
| title_full | Hybrid Convolutional Network Combining Multiscale 3D Depthwise Separable Convolution and CBAM Residual Dilated Convolution for Hyperspectral Image Classification |
| title_fullStr | Hybrid Convolutional Network Combining Multiscale 3D Depthwise Separable Convolution and CBAM Residual Dilated Convolution for Hyperspectral Image Classification |
| title_full_unstemmed | Hybrid Convolutional Network Combining Multiscale 3D Depthwise Separable Convolution and CBAM Residual Dilated Convolution for Hyperspectral Image Classification |
| title_short | Hybrid Convolutional Network Combining Multiscale 3D Depthwise Separable Convolution and CBAM Residual Dilated Convolution for Hyperspectral Image Classification |
| title_sort | hybrid convolutional network combining multiscale 3d depthwise separable convolution and cbam residual dilated convolution for hyperspectral image classification |
| topic | convolutional neural networks hyperspectral image classification depthwise separable CBAM MDRDNet |
| url | https://www.mdpi.com/2072-4292/15/19/4796 |
| work_keys_str_mv | AT yichenghu hybridconvolutionalnetworkcombiningmultiscale3ddepthwiseseparableconvolutionandcbamresidualdilatedconvolutionforhyperspectralimageclassification AT shufangtian hybridconvolutionalnetworkcombiningmultiscale3ddepthwiseseparableconvolutionandcbamresidualdilatedconvolutionforhyperspectralimageclassification AT jiage hybridconvolutionalnetworkcombiningmultiscale3ddepthwiseseparableconvolutionandcbamresidualdilatedconvolutionforhyperspectralimageclassification |