A Multidimensional Spectral Transformer with Channel-Wise Correlation for Hyperspectral Image Classification
Convolutional neural networks (CNNs) have been developed as an effective strategy for hyperspectral image (HSI) classification. However, the lack of feature extraction by CNN networks is due to the network failing to effectively extract global features and poor capability in distinguishing between d...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/9/5482 |
| _version_ | 1797603015691075584 |
|---|---|
| author | Kai Zhang Zheng Tan Jianying Sun Baoyu Zhu Yuanbo Yang Qunbo Lv |
| author_facet | Kai Zhang Zheng Tan Jianying Sun Baoyu Zhu Yuanbo Yang Qunbo Lv |
| author_sort | Kai Zhang |
| collection | DOAJ |
| description | Convolutional neural networks (CNNs) have been developed as an effective strategy for hyperspectral image (HSI) classification. However, the lack of feature extraction by CNN networks is due to the network failing to effectively extract global features and poor capability in distinguishing between different feature categories that are similar. In order to solve these problems, this paper proposes a novel approach to hyperspectral image classification using a multidimensional spectral transformer with channel-wise correlation. The proposed method consists of two key components: an input mask and a channel correlation block. The input mask is used to extract relevant spectral information from hyperspectral images and discard irrelevant information, reducing the dimensionality of the input data and improving classification accuracy. The channel correlation block captures the correlations between different spectral channels and is integrated into the transformer network to improve the model’s discrimination power. The experimental results demonstrate that the proposed method achieves great performance with several benchmark hyperspectral image datasets. The input mask and channel correlation block effectively improve classification accuracy and reduce computational complexity. |
| first_indexed | 2024-03-11T04:24:33Z |
| format | Article |
| id | doaj.art-d0c9f2a3e1a44350bfe9eaf650b48d01 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-11T04:24:33Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-d0c9f2a3e1a44350bfe9eaf650b48d012023-11-17T22:34:38ZengMDPI AGApplied Sciences2076-34172023-04-01139548210.3390/app13095482A Multidimensional Spectral Transformer with Channel-Wise Correlation for Hyperspectral Image ClassificationKai Zhang0Zheng Tan1Jianying Sun2Baoyu Zhu3Yuanbo Yang4Qunbo Lv5Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, ChinaConvolutional neural networks (CNNs) have been developed as an effective strategy for hyperspectral image (HSI) classification. However, the lack of feature extraction by CNN networks is due to the network failing to effectively extract global features and poor capability in distinguishing between different feature categories that are similar. In order to solve these problems, this paper proposes a novel approach to hyperspectral image classification using a multidimensional spectral transformer with channel-wise correlation. The proposed method consists of two key components: an input mask and a channel correlation block. The input mask is used to extract relevant spectral information from hyperspectral images and discard irrelevant information, reducing the dimensionality of the input data and improving classification accuracy. The channel correlation block captures the correlations between different spectral channels and is integrated into the transformer network to improve the model’s discrimination power. The experimental results demonstrate that the proposed method achieves great performance with several benchmark hyperspectral image datasets. The input mask and channel correlation block effectively improve classification accuracy and reduce computational complexity.https://www.mdpi.com/2076-3417/13/9/5482hyperspectral image classificationtransformerchannel-wise correlation |
| spellingShingle | Kai Zhang Zheng Tan Jianying Sun Baoyu Zhu Yuanbo Yang Qunbo Lv A Multidimensional Spectral Transformer with Channel-Wise Correlation for Hyperspectral Image Classification Applied Sciences hyperspectral image classification transformer channel-wise correlation |
| title | A Multidimensional Spectral Transformer with Channel-Wise Correlation for Hyperspectral Image Classification |
| title_full | A Multidimensional Spectral Transformer with Channel-Wise Correlation for Hyperspectral Image Classification |
| title_fullStr | A Multidimensional Spectral Transformer with Channel-Wise Correlation for Hyperspectral Image Classification |
| title_full_unstemmed | A Multidimensional Spectral Transformer with Channel-Wise Correlation for Hyperspectral Image Classification |
| title_short | A Multidimensional Spectral Transformer with Channel-Wise Correlation for Hyperspectral Image Classification |
| title_sort | multidimensional spectral transformer with channel wise correlation for hyperspectral image classification |
| topic | hyperspectral image classification transformer channel-wise correlation |
| url | https://www.mdpi.com/2076-3417/13/9/5482 |
| work_keys_str_mv | AT kaizhang amultidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT zhengtan amultidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT jianyingsun amultidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT baoyuzhu amultidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT yuanboyang amultidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT qunbolv amultidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT kaizhang multidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT zhengtan multidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT jianyingsun multidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT baoyuzhu multidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT yuanboyang multidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification AT qunbolv multidimensionalspectraltransformerwithchannelwisecorrelationforhyperspectralimageclassification |