A Hybrid-Order Spectral-Spatial Feature Network for Hyperspectral Image Classification

doi:10.3390/rs14153555

CIOMP OpenIR

	A Hybrid-Order Spectral-Spatial Feature Network for Hyperspectral Image Classification
	D. X. Liu; G. L. Han; P. X. Liu; Y. R. Wang; H. Yang; D. B. Chen; Q. Q. Li and J. J. Wu
	2022
发表期刊	Remote Sensing
卷号	14 期号:15 页码:28
摘要	Convolutional neural networks are widely applied in hyperspectral image (HSI) classification and show excellent performance. However, there are two challenges: the first is that fine features are generally lost in the process of depth transfer; the second is that most existing studies usually restore to first-order features, whereas they rarely consider second-order representations. To tackle the above two problems, this article proposes a hybrid-order spectral-spatial feature network (HS(2)FNet) for hyperspectral image classification. This framework consists of a precedent feature extraction module (PFEM) and a feature rethinking module (FRM). The former is constructed to capture multiscale spectral-spatial features and focus on adaptively recalibrate channel-wise and spatial-wise feature responses to achieve first-order spectral-spatial feature distillation. The latter is devised to heighten the representative ability of HSI by capturing the importance of feature cross-dimension, while learning more discriminative representations by exploiting the second-order statistics of HSI, thereby improving the classification performance. Massive experiments demonstrate that the proposed network achieves plausible results compared with the state-of-the-art classification methods.
DOI	10.3390/rs14153555
URL	查看原文
收录类别	sci ; ei
语种	英语
引用统计
文献类型	期刊论文
条目标识符	http://ir.ciomp.ac.cn/handle/181722/66739
专题	中国科学院长春光学精密机械与物理研究所
推荐引用方式 GB/T 7714	D. X. Liu,G. L. Han,P. X. Liu,et al. A Hybrid-Order Spectral-Spatial Feature Network for Hyperspectral Image Classification[J]. Remote Sensing,2022,14(15):28.
APA	D. X. Liu.,G. L. Han.,P. X. Liu.,Y. R. Wang.,H. Yang.,...&Q. Q. Li and J. J. Wu.(2022).A Hybrid-Order Spectral-Spatial Feature Network for Hyperspectral Image Classification.Remote Sensing,14(15),28.
MLA	D. X. Liu,et al."A Hybrid-Order Spectral-Spatial Feature Network for Hyperspectral Image Classification".Remote Sensing 14.15(2022):28.

条目包含的文件		下载所有文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
A Hybrid-Order Spect（14051KB）	期刊论文	出版稿	开放获取	CC BY-NC-SA	浏览下载