Recurrent Generative Adversarial Networks for Proximal Learning and Automated Compressive Image Recovery

TL;DR

This paper introduces a recurrent generative adversarial network architecture that unrolls proximal gradient iterations for efficient and plausible image recovery from undersampled measurements, demonstrated on MRI and face superresolution tasks.

Contribution

It proposes a cascaded network unrolling proximal gradient methods with generative ResNets for improved image reconstruction and superresolution, emphasizing efficiency and physical plausibility.

Findings

Recurrent ResNet with one residual block outperforms deeper ResNets in MRI reconstruction by 2dB SNR.

The proposed method achieves 4dB SNR improvement over conventional compressed-sensing MRI.

The architecture enables 100x faster inference in MRI reconstruction.

Abstract

Recovering images from undersampled linear measurements typically leads to an ill-posed linear inverse problem, that asks for proper statistical priors. Building effective priors is however challenged by the low train and test overhead dictated by real-time tasks; and the need for retrieving visually "plausible" and physically "feasible" images with minimal hallucination. To cope with these challenges, we design a cascaded network architecture that unrolls the proximal gradient iterations by permeating benefits from generative residual networks (ResNet) to modeling the proximal operator. A mixture of pixel-wise and perceptual costs is then deployed to train proximals. The overall architecture resembles back-and-forth projection onto the intersection of feasible and plausible images. Extensive computational experiments are examined for a global task of reconstructing MR images of pediatric patients, and a more local task of superresolving CelebA faces, that are insightful to design efficient architectures. Our observations indicate that for MRI reconstruction, a recurrent ResNet with a single residual block effectively learns the proximal. This simple architecture appears to significantly outperform the alternative deep ResNet architecture by 2dB SNR, and the conventional compressed-sensing MRI by 4dB SNR with 100x faster inference. For image superresolution, our preliminary results indicate that modeling the denoising proximal demands deep ResNets.