AXS-Net：基于光谱基元分解与结构化噪声细化的高光谱图像去噪可解释深度展开
AXS-Net: Interpretable Deep Unfolding for Hyperspectral Image Denoising via Spectral Basis Unmixing and Structured Noise Refinement
- Xiangtan University（湘潭大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
AXS-Net 通过将高光谱去噪建模为光谱分解与结构化噪声分离的优化问题并展开为深度网络，实现了可解释去噪，在多个数据集上表现优异。
AI中文摘要：
高光谱图像（HSIs）常受到混合噪声的退化，包括与波段相关的高斯扰动以及结构化伪影，如条纹、死线和脉冲噪声。大多数深度去噪器直接回归干净图像，将信号与结构化噪声纠缠在一起。我们转而将高光谱图像去噪建模为 $\Y=\A\X+\Snoise+\Nnoise$，其中 $\A\X$ 是低秩光谱子空间（分解）重建，$\Snoise$ 是结构化稀疏噪声，$\Nnoise$ 是残差高斯噪声。由此产生的正则化优化问题被展开为 AXS-Net，一个 $K$ 阶段交替近端点框架。每个阶段结合了解析光谱基梯度步骤、用于丰度系数的 SSX-Block 近端算子，以及用于具有列一致性和稀疏先验的结构化残差的 SBlock 近端算子。这种优化对应关系揭示了可解释的端元、丰度图和结构化噪声估计。在 ICVL、CAVE 和 Harvard 数据集以及五种噪声配置上，所提出的 AXS-Net 实现了强大的域内精度和有竞争力的零样本迁移，在 ICVL 和 Harvard 上所有五种噪声情况下均获得一致的增益。恢复的结构化噪声紧密跟随合成参考，恢复的光谱基是平滑且按波段排序的，而非任意的潜在通道集。
英文摘要：
Hyperspectral images (HSIs) are often degraded by mixed noise, including band-dependent Gaussian perturbations and structured artifacts such as stripes, dead-lines, and impulse noise. Most deep denoisers regress the clean image directly, entangling signal and structured noise. We instead model HSI denoising as $\Y=\A\X+\Snoise+\Nnoise$, where $\A\X$ is a low-rank spectral-subspace (unmixing) reconstruction, $\Snoise$ is structured sparse noise and $\Nnoise$ is residual Gaussian noise. The resulting regularized optimization problem is unrolled into AXS-Net, a $K$-stage alternating proximal-point framework. Each stage combines an analytic spectral-basis gradient step, an SSX-Block proximal operator for abundance coefficients, and an SBlock proximal operator for the structured residual with column-consistent and sparse priors. This optimization correspondence exposes interpretable endmembers, abundance maps, and structured-noise estimates. Across ICVL, CAVE, and Harvard datasets and five noise configurations, the proposed AXS-Net achieves strong in-domain accuracy and competitive zero-shot transfer, with consistent gains across all five noise regimes on ICVL and Harvard. The recovered structured-noise closely follows the synthetic reference, and the recovered spectral basis is smooth and band-ordered rather than an arbitrary set of latent channels.
