FDGNet: Frequency Disentanglement and Data Geometry for Domain Generalization in Cross-Scene Hyperspectral Image Classification
IEEE Transactions on Neural Networks and Learning Systems · 2025-06 · 36 (6) · 10297-10310
DOI: 10.1109/tnnls.2024.3445136 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Neural Networks and Learning Systems, 2025-06. Verfasst von Boao Qin und 5 weiteren Autor:innen. Die Arbeit wurde bisher 102-mal zitiert (Crossref) und nennt 55 Literaturangaben. Gefördert von: Open Fund of State Key Laboratory of Integrated Services Networks, China Postdoctoral Science Foundation, National Natural Science Foundation of China.
Abstract (Originalsprache)
Cross-scene hyperspectral image classification (HSIC) poses a significant challenge in recognizing hyperspectral images (HSIs) from different domains. The current mainstream approaches based on domain adaptation (DA) methods need to access target data when aligning distributions between domains, limiting the applicability of the model. In contrast, recent domain generalization (DG) methods aim to directly generalize to unseen domains, eliminating the requirements for target data during training. Nonetheless, most DG-based methods overly focus on randomizing sample styles, leading to semantically compromised samples. In addition, broadening the source distribution without ensuring reasonable support may result in undesired extended distributions. To address these issues, we propose a novel DG network with frequency disentanglement and data geometry (FDGNet) for cross-scene HSIC. Specifically, we first develop a spectral-spatial encoder based on frequency disentanglement (FDSS encoder), which facilitates synthesized domains to preserve their semantic consistency while simulating interdomain gaps with the source domain. Second, to avoid the generation of unrealistic samples, we incorporate data geometry into adversarial training. This helps diversify new domains while keeping the data geometry of extended domains in an explainable support. To improve the learning of domain-invariant representation, we propose an intermediate domain sampling strategy based on the class-wise perceptual manifold. This strategy synthesizes reliable intermediate domains by sampling from class-wise manifold flows estimated over the source and extended domains. Extensive experiments and analysis on three public HSI datasets yield the superiority of our proposed FDGNet. The codes will be available from the website: https://github.com/Qba-heu/FDGNet.
Förderung
- Open Fund of State Key Laboratory of Integrated Services Networks
- China Postdoctoral Science Foundation
- National Natural Science Foundation of China
- Fundamental Research Funds for the Central Universities
- Open Fund of State Key Laboratory of Remote Sensing Science
Zitieren
APA
Qin, B., Feng, S., Zhao, C., Xi, B., Li, W., & Tao, R. (2025). FDGNet: Frequency Disentanglement and Data Geometry for Domain Generalization in Cross-Scene Hyperspectral Image Classification. IEEE Transactions on Neural Networks and Learning Systems, 36(6), 10297-10310. https://doi.org/10.1109/tnnls.2024.3445136
Vancouver
Qin B, Feng S, Zhao C, Xi B, Li W, Tao R. FDGNet: Frequency Disentanglement and Data Geometry for Domain Generalization in Cross-Scene Hyperspectral Image Classification. IEEE Transactions on Neural Networks and Learning Systems. 2025;36(6):10297-10310. doi:10.1109/tnnls.2024.3445136
BibTeX
@article{Qin2025,
author = {Qin, Boao and Feng, Shou and Zhao, Chunhui and Xi, Bobo and Li, Wei and Tao, Ran},
title = {FDGNet: Frequency Disentanglement and Data Geometry for Domain Generalization in Cross-Scene Hyperspectral Image Classification},
journal = {IEEE Transactions on Neural Networks and Learning Systems},
year = {2025},
volume = {36},
number = {6},
pages = {10297--10310},
doi = {10.1109/tnnls.2024.3445136}
}
EUVIMED