Phy-CoSF: Physics-Guided Continuous Spectral Fields Reconstruction and Super-Resolution for Snapshot Compressive Imaging

2026-05-13

AI Summary

This work addresses the limitations of existing snapshot compressive spectral imaging methods, which are restricted to reconstructing hyperspectral images at fixed discrete wavelengths and struggle to model continuous spectra or achieve spectral super-resolution. To overcome this, the authors propose a two-stage deep unfolding architecture that integrates implicit neural representation with cross-domain feature learning. In the first stage, a deep unfolding network recovers hyperspectral images at discrete wavelengths from snapshot measurements. The second stage introduces a novel Continuous Spectral Field (CoSF) module as a dynamic prior, enabling high-fidelity spectral rendering at arbitrary target wavelengths. The framework supports arbitrary spectral resolution and employs a three-branch cross-domain mixer to fuse spatial, frequency, and channel features. Experiments demonstrate that the proposed method significantly outperforms state-of-the-art approaches in reconstruction accuracy and spectral detail preservation, achieving, for the first time, continuous spectral super-resolution from snapshot imaging.

Abstract

Recent advances have demonstrated that coded aperture snapshot spectral imaging (CASSI) systems show great potential for capturing 3D hyperspectral images (HSIs) from a single 2D measurement. Despite the inherent spectral continuity of scenes captured by CASSI, most existing reconstruction methods are restricted to fixed, discrete spectral outputs, thereby precluding continuous spectral reconstruction or spectral super-resolution. To address this challenge, we propose Phy-CoSF, which synergizes deep unfolding networks with implicit neural representations, establishing a new paradigm for continuous spectral reconstruction and super-resolution in CASSI. Specifically, we propose a two-phase architecture that bridges discrete-wavelength training with continuous spectral rendering, enabling the synthesis of high-fidelity HSIs at arbitrary target wavelengths. At the core of our framework lies the continuous spectral fields (CoSF) module, embedded within each unfolding stage as a dynamic prior, which comprises a triple-branch cross-domain feature mixer for comprehensive spatial-frequency-channel feature fusion, alongside a spectral synthesis head that generates spectral intensities by querying continuous wavelength coordinates. Extensive experimental results demonstrate that Phy-CoSF not only achieves continuous modeling at arbitrary spectral resolutions but also outperforms many state-of-the-art methods in both reconstruction fidelity and spectral detail preservation. Our code and more results are available at: https://github.com/PaiDii/Phy-CoSF.git.

Problem

• snapshot compressive imaging
• hyperspectral image reconstruction
• spectral super-resolution
• continuous spectral representation
• coded aperture

Innovation

• continuous spectral reconstruction
• spectral super-resolution
• implicit neural representation
• deep unfolding network
• snapshot compressive imaging