Deep Physics-Guided Unrolling Generalization for Compressed Sensing

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...

Full description

Saved in:
Bibliographic Details
Published inInternational journal of computer vision Vol. 131; no. 11; pp. 2864 - 2887
Main Authors Chen, Bin, Song, Jiechong, Xie, Jingfen, Zhang, Jian
Format Journal Article
LanguageEnglish
Published New York Springer US 01.11.2023
Springer Nature B.V
Subjects
Algorithms
Artificial Intelligence
Computer Imaging
Computer Science
Image Processing and Computer Vision
Image reconstruction
Inference
Iterative methods
Learning
Optimization models
Pattern Recognition
Pattern Recognition and Graphics
Physics
Recovery
Vision
Deep unrolling
Physics-guided network
Inverse imaging problem
Compressed sensing
Online AccessGet full text

Cover

More Information
Summary: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 Physics-guided unRolled recovery Learning (RL) 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: PRL-PGD and 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.
ISSN:0920-5691
1573-1405
DOI:10.1007/s11263-023-01814-w