局部精确精修：一种用于增强基础模型对抗光谱漂移泛化能力的双门控混合专家框架
Local Precise Refinement: A Dual-Gated Mixture-of-Experts for Enhancing Foundation Model Generalization against Spectral Shifts
把 MoE 的双门控路由用在基础模型特征的空间自适应精修上，并引入由 RGB 波段估计的深度特征作为结构先验，通过交叉注意力融合——这套「模态独立路由 + 结构引导融合」的微调范式可迁移到其他存在域偏移的密集预测任务。
原文依据 a dual-gated MoE architecture that independently routes visual and depth features to top-k selected experts for specialized refinement
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论光谱遥感中的域泛化语义分割（DGSS）受到不同采集条件下光谱漂移的严重挑战，这会导致模型在未见域上部署时性能显著下降。尽管微调基础模型是一个有前景的方向，现有方法采用全局、同质的调整。这种「一刀切」的调优难以应对地表覆盖的空间异质性，从而造成语义混淆。我们认为，鲁棒 DGSS 的关键不在于单一的全局自适应，而在于对基础模型的特征进行细粒度、空间自适应的精修。为此，我们提出 SpectralMoE，一个面向 DGSS 的新型微调框架。该框架通过采用混合专家（MoE）架构对基础模型的特征执行局部精确精修来落实这一原则，并引入由光谱遥感影像中选定 RGB 波段估计得到的深度特征来引导微调过程。具体而言，SpectralMoE 采用双门控 MoE 架构，将视觉特征与深度特征分别路由至 top-k 选中的专家进行专门精修，从而实现模态特定的调整。随后的交叉注意力机制将精修后的结构线索审慎地融合进视觉流，缓解由光谱变化引起的语义歧义。大量实验表明，SpectralMoE 在高光谱、多光谱和 RGB 遥感影像的多个 DGSS 基准上取得了新的最优性能。
英文原文摘要 1585 字符
Domain Generalization Semantic Segmentation (DGSS) in spectral remote sensing is severely challenged by spectral shifts across diverse acquisition conditions, which cause significant performance degradation for models deployed in unseen domains. While fine-tuning foundation models is a promising direction, existing methods employ global, homogeneous adjustments. This "one-size-fits-all" tuning struggles with the spatial heterogeneity of land cover, causing semantic confusion. We argue that the key to robust DGSS lies not in a single global adaptation, but in performing fine-grained, spatially-adaptive refinement of a foundation model's features. To achieve this, we propose SpectralMoE, a novel fine-tuning framework for DGSS. It operationalizes this principle by utilizing a Mixture-of-Experts (MoE) architecture to perform local precise refinement on the foundation model's features, incorporating depth features estimated from selected RGB bands of the spectral remote sensing imagery to guide the fine-tuning process. Specifically, SpectralMoE employs a dual-gated MoE architecture that independently routes visual and depth features to top-k selected experts for specialized refinement, enabling modality-specific adjustments. A subsequent cross-attention mechanism then judiciously fuses the refined structural cues into the visual stream, mitigating semantic ambiguities caused by spectral variations. Extensive experiments show that SpectralMoE sets a new state-of-the-art on multiple DGSS benchmarks across hyperspectral, multispectral, and RGB remote sensing imagery.