LiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification
Deep learning (DL) algorithms have been demonstrated to have great potential in hyperspectral image (HSI) classification. However, most DL methods mainly focus on improving classification performance, neglecting the computational cost. In order to broaden the application scenarios of DL-based HSI cl...
Guardado en:
Autores principales: | , , , |
---|---|
Formato: | article |
Lenguaje: | EN |
Publicado: |
IEEE
2021
|
Materias: | |
Acceso en línea: | https://doaj.org/article/71468806f9ef491ab9ebcf4065cd60e0 |
Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
id |
oai:doaj.org-article:71468806f9ef491ab9ebcf4065cd60e0 |
---|---|
record_format |
dspace |
spelling |
oai:doaj.org-article:71468806f9ef491ab9ebcf4065cd60e02021-12-01T00:00:17ZLiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification2151-153510.1109/JSTARS.2021.3124321https://doaj.org/article/71468806f9ef491ab9ebcf4065cd60e02021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9599575/https://doaj.org/toc/2151-1535Deep learning (DL) algorithms have been demonstrated to have great potential in hyperspectral image (HSI) classification. However, most DL methods mainly focus on improving classification performance, neglecting the computational cost. In order to broaden the application scenarios of DL-based HSI classification methods, it is necessary to develop lightweight and fast models to fit the deployment needs of computational-resource-limited platforms. Taking into account the efficiency and accuracy, this article designs a lightweight network architecture based on spectral and channel-wise attention modules, namely LiteSCANet, for HSI classification. It contains a residual double-branch structure, which makes the model effectively extract spectral-spatial features and achieve good performance with fast inference speed and low computational consumption (i.e., floating-point operations, number of parameters, and graphics processing unit memory usage). The experiment results on four benchmark datasets show that our proposed model achieves an excellent trade-off between efficiency and accuracy compared with the other six existing networks.Su QiaoXue-Mei DongJiangtao PengWeiwei SunIEEEarticleAttentionhyperspectral image (HSI) classificationiterative pruninglightweight networkresidual double-branchOcean engineeringTC1501-1800Geophysics. Cosmic physicsQC801-809ENIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol 14, Pp 11655-11668 (2021) |
institution |
DOAJ |
collection |
DOAJ |
language |
EN |
topic |
Attention hyperspectral image (HSI) classification iterative pruning lightweight network residual double-branch Ocean engineering TC1501-1800 Geophysics. Cosmic physics QC801-809 |
spellingShingle |
Attention hyperspectral image (HSI) classification iterative pruning lightweight network residual double-branch Ocean engineering TC1501-1800 Geophysics. Cosmic physics QC801-809 Su Qiao Xue-Mei Dong Jiangtao Peng Weiwei Sun LiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification |
description |
Deep learning (DL) algorithms have been demonstrated to have great potential in hyperspectral image (HSI) classification. However, most DL methods mainly focus on improving classification performance, neglecting the computational cost. In order to broaden the application scenarios of DL-based HSI classification methods, it is necessary to develop lightweight and fast models to fit the deployment needs of computational-resource-limited platforms. Taking into account the efficiency and accuracy, this article designs a lightweight network architecture based on spectral and channel-wise attention modules, namely LiteSCANet, for HSI classification. It contains a residual double-branch structure, which makes the model effectively extract spectral-spatial features and achieve good performance with fast inference speed and low computational consumption (i.e., floating-point operations, number of parameters, and graphics processing unit memory usage). The experiment results on four benchmark datasets show that our proposed model achieves an excellent trade-off between efficiency and accuracy compared with the other six existing networks. |
format |
article |
author |
Su Qiao Xue-Mei Dong Jiangtao Peng Weiwei Sun |
author_facet |
Su Qiao Xue-Mei Dong Jiangtao Peng Weiwei Sun |
author_sort |
Su Qiao |
title |
LiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification |
title_short |
LiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification |
title_full |
LiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification |
title_fullStr |
LiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification |
title_full_unstemmed |
LiteSCANet: An Efficient Lightweight Network Based on Spectral and Channel-Wise Attention for Hyperspectral Image Classification |
title_sort |
litescanet: an efficient lightweight network based on spectral and channel-wise attention for hyperspectral image classification |
publisher |
IEEE |
publishDate |
2021 |
url |
https://doaj.org/article/71468806f9ef491ab9ebcf4065cd60e0 |
work_keys_str_mv |
AT suqiao litescanetanefficientlightweightnetworkbasedonspectralandchannelwiseattentionforhyperspectralimageclassification AT xuemeidong litescanetanefficientlightweightnetworkbasedonspectralandchannelwiseattentionforhyperspectralimageclassification AT jiangtaopeng litescanetanefficientlightweightnetworkbasedonspectralandchannelwiseattentionforhyperspectralimageclassification AT weiweisun litescanetanefficientlightweightnetworkbasedonspectralandchannelwiseattentionforhyperspectralimageclassification |
_version_ |
1718406207641747456 |