Search papers, labs, and topics across Lattice.
This paper introduces the Hyperspectral Diffusion Equivariant Imaging (HyDiff-EI) framework, which addresses the hyperspectral image inpainting problem using a self-supervised approach that learns from a single corrupted hyperspectral image acquisition. By embedding equivariant consistency constraints within the diffusion process, the framework effectively leverages the geometric symmetries of hyperspectral data, enhancing noise robustness and generalizability. Extensive experiments on real-world datasets demonstrate that HyDiff-EI significantly outperforms existing self-supervised and diffusion-based methods in both noiseless and noisy scenarios, highlighting its practical applicability in remote sensing.
Coupling diffusion modeling with equivariant priors boosts inpainting quality, achieving remarkable results even with limited data.
A novel Hyperspectral diffusion Equivariant Imaging (HyDiff-EI) framework for solving the hyperspectral image (HSI) inpainting problem has been presented here. Unlike conventional diffusion-based methods that rely on large-scale pretraining, HyDiff-EI is a test-time optimization framework that learns directly from a single corrupted HSI acquisition. This makes it flexible for different sensor configurations and particularly well-suited for practical remote sensing scenarios where large annotated hyperspectral datasets are limited. To address the ill-posed nature of unsupervised inpainting, we embed equivariant consistency constraints within the diffusion process. By leveraging the inherent geometric symmetries and intrinsic characteristics of HSIs, HyDiff-EI bridges the gap between generative diffusion modeling and self-consistent physical priors. We empirically show that coupling diffusion modeling with equivariant priors substantially enhances noise robustness and generalizability. Extensive experiments on real-world datasets including Chikusei, Botswana, and EMIT demonstrate that HyDiff-EI offers remarkable inpainting quality over existing self-supervised and diffusion-based algorithms in both noiseless and noisy cases.
