High-Dimensional Noise to Low-Dimensional Manifolds: A Manifold-Space Diffusion Framework for Degraded Hyperspectral Image Classification

2026-04-29

AI Summary

This work addresses the degradation of classification performance in high-dimensional hyperspectral remote sensing imagery under complex degradation conditions, where the intrinsic low-dimensional manifold structure is often disrupted. To mitigate this issue, the authors propose a novel approach that first embeds degraded data into a low-dimensional manifold preserving class semantics and then, for the first time, introduces a diffusion generative model within this manifold space. By iteratively denoising latent features, the method effectively disentangles degradation-induced interference from discriminative structural information. Enhanced by discriminative spectral-spatial reconstruction and manifold constraints, the framework substantially improves representation stability. Extensive experiments on multiple hyperspectral benchmark datasets under composite degradation scenarios demonstrate that the proposed method significantly outperforms current state-of-the-art approaches in classification accuracy.

Abstract

Recently, Hyperspectral Image (HSI) classification has attracted increasing attention in remote sensing. However, HSI data are inherently high-dimensional but low-rank, with discriminative information concentrated on a low-dimensional latent manifold. In real-world remote sensing scenarios, the superposition of multiple degradation factors disrupts this intrinsic manifold structure, driving samples away from their original low-dimensional distribution and introducing substantial redundant and non-discriminative variations. To better handle this challenge, this paper proposes a manifold-space diffusion framework (MSDiff) for robust hyperspectral classification under complex degradation conditions. Specifically, the proposed method first maps high-dimensional, degradation-affected HSI data into a compact low-dimensional manifold through a discriminative spectral-spatial reconstruction task, preserving class semantics and reducing redundant variations. A diffusion-based generative model is then applied to regularize the spectral-spatial distribution within the manifold, enabling progressive refinement and stabilization of latent features against residual degradations. The key advantage of the proposed framework lies in performing diffusion-based distribution modeling directly on the low-dimensional manifold, effectively decoupling degradation-induced disturbances from intrinsic discriminative structures and enhancing representation stability under complex degradations. Experimental results on multiple hyperspectral benchmarks demonstrate consistent performance improvements over state-of-the-art methods under diverse composite degradation settings. The code will be available at https://github.com/yangboxiang1207/MSDiff

Problem

• Hyperspectral Image Classification
• Degradation
• Low-Dimensional Manifold
• High-Dimensional Noise
• Manifold Structure

Innovation

• manifold-space diffusion
• hyperspectral image classification
• low-dimensional manifold
• spectral-spatial reconstruction
• degradation robustness