Spectral-spatial self-attention networks for hyperspectral image classification.

Zhang, Xuming; Sun, Genyun; Jia, Xiuping; Wu, Lixin; Zhang, Aizhu; Ren, Jinchang; Fu, Hang; Yao, Yanjuan

Abstract

This study presents a spectral-spatial self-attention network (SSSAN) for classification of hyperspectral images (HSIs), which can adaptively integrate local features with long-range dependencies related to the pixel to be classified. Specifically, it has two subnetworks. The spatial subnetwork introduces the proposed spatial self-attention module to exploit rich patch-based contextual information related to the center pixel. The spectral subnetwork introduces the proposed spectral self-attention module to exploit the long-range spectral correlation over local spectral features. The extracted spectral and spatial features are then adaptively fused for HSI classification. Experiments conducted on four HSI datasets demonstrate that the proposed network outperforms several state-of-the-art methods.

Citation

ZHANG, X., SUN, G., JIA, X., WU, L., ZHANG, A., REN, J., FU, H. and YAO, Y. 2022. Spectral-spatial self-attention networks for hyperspectral image classification. IEEE transactions on geoscience and remote sensing [online], 60, article 5512115. Available from: https://doi.org/10.1109/TGRS.2021.3102143

• Journal Article Type: Article
• Acceptance Date: Jul 4, 2021
• Online Publication Date: Aug 6, 2021
• Publication Date: Jan 17, 2022
• Deposit Date: Aug 13, 2021
• Publicly Available Date: Aug 13, 2021
• Journal: IEEE Transactions on Geoscience and Remote Sensing
• Print ISSN: 0196-2892
• Electronic ISSN: 1558-0644
• Publisher: Institute of Electrical and Electronics Engineers (IEEE)
• Peer Reviewed: Peer Reviewed
• Volume: 60
• Article Number: 5512115
• DOI: https://doi.org/10.1109/TGRS.2021.3102143
• Keywords: Adaptation models; Convolution; Convolutional neural network (CNN); Correlation; Deep learning; Feature extraction; Hyperspectral image (HSI) classification; Hyperspectral imaging; Image segmentation; Spatial self-attention module; Spectral self-attention
• Public URL: https://rgu-repository.worktribe.com/output/1406215