Progressive Flow-inspired Unfolding for Spectral Compressive Imaging

2025-09-15

AI Summary

In coded aperture snapshot spectral imaging (CASSI), reconstructing 3D hyperspectral images (HSIs) from single-frame 2D compressive measurements suffers from low reconstruction quality, uncontrollable unfolding trajectories, and non-progressive inter-stage optimization. To address these issues, this paper proposes a diffusion-inspired, trajectory-controllable progressive unfolding framework. It introduces a continuous optimization mechanism into deep unfolding networks for adjustable, smooth, and stage-wise refinement of the reconstruction path. The architecture integrates a spatial-spectral Transformer with a frequency-domain fusion module to jointly enforce explicit data fidelity constraints and implicit denoising priors. Extensive experiments on both simulated and real-world CASSI datasets demonstrate that the proposed method significantly outperforms state-of-the-art approaches in both reconstruction accuracy and inference efficiency.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Assisted, interactive, and conversational search
• Security and Privacy: Data transparency and provenance

Abstract

Coded aperture snapshot spectral imaging (CASSI) retrieves a 3D hyperspectral image (HSI) from a single 2D compressed measurement, which is a highly challenging reconstruction task. Recent deep unfolding networks (DUNs), empowered by explicit data-fidelity updates and implicit deep denoisers, have achieved the state of the art in CASSI reconstruction. However, existing unfolding approaches suffer from uncontrollable reconstruction trajectories, leading to abrupt quality jumps and non-gradual refinement across stages. Inspired by diffusion trajectories and flow matching, we propose a novel trajectory-controllable unfolding framework that enforces smooth, continuous optimization paths from noisy initial estimates to high-quality reconstructions. To achieve computational efficiency, we design an efficient spatial-spectral Transformer tailored for hyperspectral reconstruction, along with a frequency-domain fusion module to gurantee feature consistency. Experiments on simulation and real data demonstrate that our method achieves better reconstruction quality and efficiency than prior state-of-the-art approaches.

Problem

• Reconstructing 3D hyperspectral images from 2D compressed measurements
• Addressing uncontrollable reconstruction trajectories in unfolding networks
• Ensuring smooth optimization paths from noisy estimates to high-quality reconstructions

Innovation

• Trajectory-controllable unfolding framework for smooth optimization
• Efficient spatial-spectral Transformer for hyperspectral reconstruction
• Frequency-domain fusion module ensuring feature consistency