MambaMoE: Mixture-of-Spectral-Spatial-Experts State Space Model for Hyperspectral Image Classification
This work addresses hyperspectral image classification, a domain-specific task, by introducing the first MoE-based approach in this area, which is incremental as it builds on existing Mamba-based models.
The paper tackles the problem of limited classification performance in hyperspectral image classification due to directional modeling heterogeneity across land-cover types, proposing MambaMoE, a spectral-spatial Mixture-of-Experts framework that achieves state-of-the-art performance in accuracy and computational efficiency on multiple benchmark datasets.
Mamba-based models have recently demonstrated significant potential in hyperspectral image (HSI) classification, primarily due to their ability to perform contextual modeling with linear computational complexity. However, existing Mamba-based approaches often overlook the directional modeling heterogeneity across different land-cover types, leading to limited classification performance. To address these limitations, we propose MambaMoE, a novel spectral-spatial Mixture-of-Experts (MoE) framework, which represents the first MoE-based approach in the HSI classification domain. Specifically, we design a Mixture of Mamba Expert Block (MoMEB) that performs adaptive spectral-spatial feature modeling via a sparse expert activation mechanism. Additionally, we introduce an uncertainty-guided corrective learning (UGCL) strategy that encourages the model to focus on complex regions prone to prediction ambiguity. This strategy dynamically samples supervision signals from regions with high predictive uncertainty, guiding the model to adaptively refine feature representations and thereby enhancing its focus on challenging areas. Extensive experiments conducted on multiple public HSI benchmark datasets show that MambaMoE achieves state-of-the-art performance in both classification accuracy and computational efficiency compared to existing advanced methods, particularly Mamba-based ones. The code will be available online at https://github.com/YichuXu/MambaMoE.