Spectral-Spatial Convolutional Hybrid Transformer for Hyperspectral Image Classification

The combination of Convolutional neural networks (CNNs) and Transformers has achieved excellent performance in hyperspectral image classification due to the characteristics of local features extraction and long-range dependencies capture. However, how to integrate the spectral and spatial features e...

Full description

Saved in:
Bibliographic Details
Main Authors: Haixin Sun, Jingwen Xu, Fanlei Meng, Mengdi Cheng, Qiuguang Cao
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10945322/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849702072042127360
author Haixin Sun
Jingwen Xu
Fanlei Meng
Mengdi Cheng
Qiuguang Cao
author_facet Haixin Sun
Jingwen Xu
Fanlei Meng
Mengdi Cheng
Qiuguang Cao
author_sort Haixin Sun
collection DOAJ
description The combination of Convolutional neural networks (CNNs) and Transformers has achieved excellent performance in hyperspectral image classification due to the characteristics of local features extraction and long-range dependencies capture. However, how to integrate the spectral and spatial features effectively after dimensionality reduction while retaining key information remains to be optimized further. To address this issue, the Spectral-spatial convolutional hybrid Transformer (SSCHFormer) hyperspectral classification model is proposed in this article. First, the spectral pyramid 3D convolution and 2D convolution are combined to extract joint and detailed spectral-spatial features. Then, the hybrid convolutional transposed attention mechanism is introduced to enhance the correlation between the local and global features further and perceive the complex spatial and spectral information while reducing computational complexity. Furthermore, the gated feedforward network is designed to enhance the ability to distinguish subtle differences between different classes. Extensive experiments are implemented on four public datasets and demonstrate the superior classification performances of the proposed SSCHFormer to those of other advanced networks.
format Article
id doaj-art-b7cd0ca1f0f54ad6b21a0dc63a7b2ec6
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-b7cd0ca1f0f54ad6b21a0dc63a7b2ec62025-08-20T03:17:46ZengIEEEIEEE Access2169-35362025-01-0113591025911710.1109/ACCESS.2025.355559310945322Spectral-Spatial Convolutional Hybrid Transformer for Hyperspectral Image ClassificationHaixin Sun0https://orcid.org/0000-0002-6339-6589Jingwen Xu1https://orcid.org/0009-0006-3204-3568Fanlei Meng2Mengdi Cheng3Qiuguang Cao4College of Electronic and Information Engineering, Changchun University, Changchun, ChinaCollege of Electronic and Information Engineering, Changchun University, Changchun, ChinaCollege of Electronic and Information Engineering, Changchun University, Changchun, ChinaCollege of Electronic and Information Engineering, Changchun University, Changchun, ChinaCollege of Electronic and Information Engineering, Changchun University, Changchun, ChinaThe combination of Convolutional neural networks (CNNs) and Transformers has achieved excellent performance in hyperspectral image classification due to the characteristics of local features extraction and long-range dependencies capture. However, how to integrate the spectral and spatial features effectively after dimensionality reduction while retaining key information remains to be optimized further. To address this issue, the Spectral-spatial convolutional hybrid Transformer (SSCHFormer) hyperspectral classification model is proposed in this article. First, the spectral pyramid 3D convolution and 2D convolution are combined to extract joint and detailed spectral-spatial features. Then, the hybrid convolutional transposed attention mechanism is introduced to enhance the correlation between the local and global features further and perceive the complex spatial and spectral information while reducing computational complexity. Furthermore, the gated feedforward network is designed to enhance the ability to distinguish subtle differences between different classes. Extensive experiments are implemented on four public datasets and demonstrate the superior classification performances of the proposed SSCHFormer to those of other advanced networks.https://ieeexplore.ieee.org/document/10945322/Gated feed forward networkhybrid convolutional transposed attentionhyperspectral image classificationspectral-spatial convolution
spellingShingle Haixin Sun
Jingwen Xu
Fanlei Meng
Mengdi Cheng
Qiuguang Cao
Spectral-Spatial Convolutional Hybrid Transformer for Hyperspectral Image Classification
IEEE Access
Gated feed forward network
hybrid convolutional transposed attention
hyperspectral image classification
spectral-spatial convolution
title Spectral-Spatial Convolutional Hybrid Transformer for Hyperspectral Image Classification
title_full Spectral-Spatial Convolutional Hybrid Transformer for Hyperspectral Image Classification
title_fullStr Spectral-Spatial Convolutional Hybrid Transformer for Hyperspectral Image Classification
title_full_unstemmed Spectral-Spatial Convolutional Hybrid Transformer for Hyperspectral Image Classification
title_short Spectral-Spatial Convolutional Hybrid Transformer for Hyperspectral Image Classification
title_sort spectral spatial convolutional hybrid transformer for hyperspectral image classification
topic Gated feed forward network
hybrid convolutional transposed attention
hyperspectral image classification
spectral-spatial convolution
url https://ieeexplore.ieee.org/document/10945322/
work_keys_str_mv AT haixinsun spectralspatialconvolutionalhybridtransformerforhyperspectralimageclassification
AT jingwenxu spectralspatialconvolutionalhybridtransformerforhyperspectralimageclassification
AT fanleimeng spectralspatialconvolutionalhybridtransformerforhyperspectralimageclassification
AT mengdicheng spectralspatialconvolutionalhybridtransformerforhyperspectralimageclassification
AT qiuguangcao spectralspatialconvolutionalhybridtransformerforhyperspectralimageclassification