FAHM: Frequency-Aware Hierarchical Mamba for Hyperspectral Image Classification

Convolutional neural networks and Transformers have garnered substantial attention in hyperspectral image (HSI) classification, and recently Mamba has made significant progress in modeling long sequences. However, existing Mamba-based approaches flatten 2-D images into 1-D sequences, inevitably disr...

Full description

Saved in:
Bibliographic Details
Main Authors: Peixian Zhuang, Xiaochen Zhang, Hao Wang, Tianxiang Zhang, Leiming Liu, Jiangyun Li
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10877784/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850204983665885184
author Peixian Zhuang
Xiaochen Zhang
Hao Wang
Tianxiang Zhang
Leiming Liu
Jiangyun Li
author_facet Peixian Zhuang
Xiaochen Zhang
Hao Wang
Tianxiang Zhang
Leiming Liu
Jiangyun Li
author_sort Peixian Zhuang
collection DOAJ
description Convolutional neural networks and Transformers have garnered substantial attention in hyperspectral image (HSI) classification, and recently Mamba has made significant progress in modeling long sequences. However, existing Mamba-based approaches flatten 2-D images into 1-D sequences, inevitably disrupting 2-D local dependencies, thereby disregarding the distinctive difference between high-frequency and low-frequency components in the frequency domain. For HSI tasks, these methods suffer from inherent receptive field limitations and insufficient exploitation of spectral information, resulting in unclear classification of regional boundaries and poor generalizability. To address these issues, we present a novel frequency-aware hierarchical Mamba (FAHM) for HSI classification. Specifically, FAHM is comprised of three main parts: a spatial-spectral interaction block (SSIB), a frequency group embedding module (FGEM), and an adaptive bidirectional Mamba block (ABMB). Building on the observation that the high-frequency component contains detailed features while the low-frequency information provides abundant high-level semantics, we design FGEM to enhance feature representation in the frequency domain. To prepare spectral–spatial contextual tokens for FGEM, we incorporate SSIB to extract shallow features, and ABMB is designed to capture spectral dependencies by modeling spectral variation bidirectionally. We construct our Mamba in a hierarchical manner, and refine multilevel features using a feature affinity module for improved accuracy. Experiments conducted on four HSI benchmark datasets show that the proposed FAHM outperforms eleven existing approaches in classification accuracy and model generalization.
format Article
id doaj-art-d302f0b5520b4083a879e627d2dc15c2
institution OA Journals
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-d302f0b5520b4083a879e627d2dc15c22025-08-20T02:11:11ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01186299631310.1109/JSTARS.2025.353979110877784FAHM: Frequency-Aware Hierarchical Mamba for Hyperspectral Image ClassificationPeixian Zhuang0https://orcid.org/0000-0002-7143-9569Xiaochen Zhang1https://orcid.org/0009-0002-5905-8541Hao Wang2https://orcid.org/0009-0002-8865-8589Tianxiang Zhang3https://orcid.org/0000-0002-0996-2586Leiming Liu4Jiangyun Li5https://orcid.org/0000-0003-2288-7901Key Laboratory of Knowledge Automation for Industrial Processes, Ministry of Education, University of Science and Technology Beijing, Beijing, ChinaSchool of Automation and Electrical Engineering, University of Science and Technology Beijing, Beijing, ChinaSchool of Automation and Electrical Engineering, University of Science and Technology Beijing, Beijing, ChinaKey Laboratory of Knowledge Automation for Industrial Processes, Ministry of Education, University of Science and Technology Beijing, Beijing, ChinaSchool of Automation and Electrical Engineering, University of Science and Technology Beijing, Beijing, ChinaKey Laboratory of Knowledge Automation for Industrial Processes, Ministry of Education, University of Science and Technology Beijing, Beijing, ChinaConvolutional neural networks and Transformers have garnered substantial attention in hyperspectral image (HSI) classification, and recently Mamba has made significant progress in modeling long sequences. However, existing Mamba-based approaches flatten 2-D images into 1-D sequences, inevitably disrupting 2-D local dependencies, thereby disregarding the distinctive difference between high-frequency and low-frequency components in the frequency domain. For HSI tasks, these methods suffer from inherent receptive field limitations and insufficient exploitation of spectral information, resulting in unclear classification of regional boundaries and poor generalizability. To address these issues, we present a novel frequency-aware hierarchical Mamba (FAHM) for HSI classification. Specifically, FAHM is comprised of three main parts: a spatial-spectral interaction block (SSIB), a frequency group embedding module (FGEM), and an adaptive bidirectional Mamba block (ABMB). Building on the observation that the high-frequency component contains detailed features while the low-frequency information provides abundant high-level semantics, we design FGEM to enhance feature representation in the frequency domain. To prepare spectral–spatial contextual tokens for FGEM, we incorporate SSIB to extract shallow features, and ABMB is designed to capture spectral dependencies by modeling spectral variation bidirectionally. We construct our Mamba in a hierarchical manner, and refine multilevel features using a feature affinity module for improved accuracy. Experiments conducted on four HSI benchmark datasets show that the proposed FAHM outperforms eleven existing approaches in classification accuracy and model generalization.https://ieeexplore.ieee.org/document/10877784/Classification accuracyfrequency domainhierarchical structurehyperspectral image (HSI) classificationstate space model (SSM)
spellingShingle Peixian Zhuang
Xiaochen Zhang
Hao Wang
Tianxiang Zhang
Leiming Liu
Jiangyun Li
FAHM: Frequency-Aware Hierarchical Mamba for Hyperspectral Image Classification
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Classification accuracy
frequency domain
hierarchical structure
hyperspectral image (HSI) classification
state space model (SSM)
title FAHM: Frequency-Aware Hierarchical Mamba for Hyperspectral Image Classification
title_full FAHM: Frequency-Aware Hierarchical Mamba for Hyperspectral Image Classification
title_fullStr FAHM: Frequency-Aware Hierarchical Mamba for Hyperspectral Image Classification
title_full_unstemmed FAHM: Frequency-Aware Hierarchical Mamba for Hyperspectral Image Classification
title_short FAHM: Frequency-Aware Hierarchical Mamba for Hyperspectral Image Classification
title_sort fahm frequency aware hierarchical mamba for hyperspectral image classification
topic Classification accuracy
frequency domain
hierarchical structure
hyperspectral image (HSI) classification
state space model (SSM)
url https://ieeexplore.ieee.org/document/10877784/
work_keys_str_mv AT peixianzhuang fahmfrequencyawarehierarchicalmambaforhyperspectralimageclassification
AT xiaochenzhang fahmfrequencyawarehierarchicalmambaforhyperspectralimageclassification
AT haowang fahmfrequencyawarehierarchicalmambaforhyperspectralimageclassification
AT tianxiangzhang fahmfrequencyawarehierarchicalmambaforhyperspectralimageclassification
AT leimingliu fahmfrequencyawarehierarchicalmambaforhyperspectralimageclassification
AT jiangyunli fahmfrequencyawarehierarchicalmambaforhyperspectralimageclassification