FCD-Mamba: A Frequency-Enhanced and Center-Pixel-Guided Dual-Branch Mamba Network for Hyperspectral Image Classification
Open Access
- 20 September 2026
- journal article
- research article
- Published by MDPI AG in Remote Sensing
- Vol. 18 (18) , 3234
Abstract
Hyperspectral image classification is an important task in remote sensing image interpretation. However, existing Mamba-based methods remain limited in exploiting shallow frequency information, preserving neighborhood continuity during spatial serialization, and protecting center-pixel semantics. To address these limitations, we propose a frequency-enhanced and center-pixel-guided dual-branch Mamba network (FCD-Mamba). First, the Frequency-Enhanced Multiscale Convolution Module (FEMSCM) applies serial Fourier filtering to enhance multiscale spatial features and calibrates spectral features using low-, middle-, and high-frequency DCT energy descriptors derived from the original spectrum of the center pixel. Second, the Spatial Center-Adaptive Mamba (SCA-Mamba) employs two orthogonal Hilbert curves and their reverse traversals to preserve spatial neighborhood continuity, while adaptively aggregating multi-path features using center and global representations. The Spectral Center-Preserving Mamba (SCP-Mamba) constructs center-preserved and global spectral sequences and models their complementary spectral dependencies through bidirectional scanning. Finally, the Center-Pixel-Guided Residual Fusion Module (CGRFM) exploits center-to-global discrepancies to calibrate dual-branch features and introduces auxiliary supervision to maintain the discriminability of each branch. Experiments on multiple public hyperspectral datasets demonstrate that FCD-Mamba achieves stable and competitive classification performance, validating the effectiveness of its constituent modules.Keywords
Funding Information
- Scientific Research Innovation Capability Support Project for Young Faculty (SRICSPYF-BS2025007)
- Central Government Guided Local Science and Technology Development Fund Projects of Qinghai Province (2026-GX-Z26)
- Basic Research Program of Jiangsu Province (BK20250043)
This publication has 51 references indexed in Scilit:
- S2PW-Mamba: Pinwheel and wavelet-based spatial-spectral mamba for hyperspectral image classificationExpert Systems with Applications, 2026
- MambaMoE: Mixture-of-spectral-spatial-experts state space model for hyperspectral image classificationInformation Fusion, 2025
- HMSNet: Hilbert curve enhanced Mamba for real-time semantic segmentationPattern Recognition, 2025
- Deep learning for hyperspectral image classification: A comprehensive review and future predictionsInformation Fusion, 2025
- CenterMamba: Enhancing semantic representation with center-scan Mamba network for hyperspectral image classificationExpert Systems with Applications, 2025
- CNN-Transformer and Channel-Spatial Attention based network for hyperspectral image classification with few samplesNeural Networks, 2025
- Tensor Transformer for hyperspectral image classificationPattern Recognition, 2025
- A synergistic CNN-transformer network with pooling attention fusion for hyperspectral image classificationDigital Signal Processing, 2025
- Mamba-in-Mamba: Centralized Mamba-Cross-Scan in Tokenized Mamba Model for Hyperspectral image classificationNeurocomputing, 2024
- SemanticFormer: Hyperspectral image classification via semantic transformerPattern Recognition Letters, 2024