U-DuDoNet: Unpaired dual-domain network for CT metal artifact reduction

TL;DR

U-DuDoNet is an innovative unpaired dual-domain deep learning framework that effectively reduces metal artifacts in CT images by modeling artifact generation and utilizing a sinogram prior, outperforming existing methods on clinical data.

Contribution

The paper introduces U-DuDoNet, a novel unpaired dual-domain network that models artifact generation directly and incorporates a sinogram prior for improved clinical CT metal artifact reduction.

Findings

Outperforms state-of-the-art unpaired methods on clinical data.

Effectively models artifact generation in sinogram and image domains.

Utilizes cyclic constraints and a sinogram prior for enhanced artifact removal.

Abstract

Recently, both supervised and unsupervised deep learning methods have been widely applied on the CT metal artifact reduction (MAR) task. Supervised methods such as Dual Domain Network (Du-DoNet) work well on simulation data; however, their performance on clinical data is limited due to domain gap. Unsupervised methods are more generalized, but do not eliminate artifacts completely through the sole processing on the image domain. To combine the advantages of both MAR methods, we propose an unpaired dual-domain network (U-DuDoNet) trained using unpaired data. Unlike the artifact disentanglement network (ADN) that utilizes multiple encoders and decoders for disentangling content from artifact, our U-DuDoNet directly models the artifact generation process through additions in both sinogram and image domains, which is theoretically justified by an additive property associated with metal artifact. Our design includes a self-learned sinogram prior net, which provides guidance for restoring the information in the sinogram domain, and cyclic constraints for artifact reduction and addition on unpaired data. Extensive experiments on simulation data and clinical images demonstrate that our novel framework outperforms the state-of-the-art unpaired approaches.