D3C2-Net: Dual-Domain Deep Convolutional Coding Network for Compressive Sensing

Weiqi Li, Bin Chen, Jian Zhang

Keywords: image reconstruction, compressive sensing (CS), convolutional coding, dual-domain optimization, deep unfolding networks

TL;DR: We propose a novel D3C2-Net for compressive sensing based on our new proposed generalized dual-domain optimization framework, achieving higher performance than other state-of-the-arts.

Abstract: Mapping optimization algorithms into neural networks, deep unfolding networks (DUNs) have achieved impressive success in compressive sensing (CS). From the perspective of optimization, DUNs inherit a well-defined and interpretable structure from iterative steps. However, from the viewpoint of neural network design, most existing DUNs are inherently established based on traditional image-domain unfolding, which takes single-channel images as inputs and outputs between adjacent stages, resulting in insufficient information transmission capability and the inevitable loss of the image details. In this paper, to break the above bottleneck, we propose a generalized dual-domain optimization framework, which is general for inverse imaging problems and integrates the merits of both (1) image-domain and (2) convolutional-coding-domain priors to constrain the feasible region of the solution space. By unfolding the proposed optimization framework into deep neural networks, we further design a novel Dual-Domain Deep Convolutional Coding Network ($\mathrm{D^3C^2}$-Net) for CS imaging with the ability of transmitting high-capacity feature through all the unfolded stages. Experiments on multiple natural and MR image datasets demonstrate that our $\mathrm{D^3C^2}$-Net achieves higher performance and better accuracy-complexity trade-offs than other state-of-the-art.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Applications (eg, speech processing, computer vision, NLP)