Deep Physics-Guided Unrolling Generalization for Compressed Sensing

TL;DR

This paper introduces a novel deep physics-guided unrolling framework for compressed sensing that enhances efficiency and accuracy by operating in a high-dimensional feature space, enabling real-time image reconstruction.

Contribution

It generalizes traditional iterative recovery models to feature domains and develops a multiscale unrolling architecture for improved performance and speed.

Findings

PRL networks outperform state-of-the-art methods in accuracy.

PRL achieves real-time inference speeds.

Framework shows potential for broader inverse imaging applications.

Abstract

By absorbing the merits of both the model- and data-driven methods, deep physics-engaged learning scheme achieves high-accuracy and interpretable image reconstruction. It has attracted growing attention and become the mainstream for inverse imaging tasks. Focusing on the image compressed sensing (CS) problem, we find the intrinsic defect of this emerging paradigm, widely implemented by deep algorithm-unrolled networks, in which more plain iterations involving real physics will bring enormous computation cost and long inference time, hindering their practical application. A novel deep $\textbf{P}$hysics-guided un$\textbf{R}$olled recovery $\textbf{L}$earning ($\textbf{PRL}$) framework is proposed by generalizing the traditional iterative recovery model from image domain (ID) to the high-dimensional feature domain (FD). A compact multiscale unrolling architecture is then developed to enhance the network capacity and keep real-time inference speeds. Taking two different perspectives of optimization and range-nullspace decomposition, instead of building an algorithm-specific unrolled network, we provide two implementations: $\textbf{PRL-PGD}$ and $\textbf{PRL-RND}$. Experiments exhibit the significant performance and efficiency leading of PRL networks over other state-of-the-art methods with a large potential for further improvement and real application to other inverse imaging problems or optimization models.