MEPNet: A Model-Driven Equivariant Proximal Network for Joint Sparse-View Reconstruction and Metal Artifact Reduction in CT Images

TL;DR

MEPNet is a model-driven equivariant proximal network designed for joint sparse-view CT reconstruction and metal artifact reduction, effectively embedding physical constraints and prior knowledge to improve image quality with fewer parameters.

Contribution

The paper introduces a novel dual-domain, optimization-inspired neural network that incorporates rotation equivariance to better utilize inherent imaging priors in CT reconstruction.

Findings

Outperforms conventional methods in reconstruction quality

Uses fewer network parameters due to rotation equivariance

Effectively reduces metal artifacts in CT images

Abstract

Sparse-view computed tomography (CT) has been adopted as an important technique for speeding up data acquisition and decreasing radiation dose. However, due to the lack of sufficient projection data, the reconstructed CT images often present severe artifacts, which will be further amplified when patients carry metallic implants. For this joint sparse-view reconstruction and metal artifact reduction task, most of the existing methods are generally confronted with two main limitations: 1) They are almost built based on common network modules without fully embedding the physical imaging geometry constraint of this specific task into the dual-domain learning; 2) Some important prior knowledge is not deeply explored and sufficiently utilized. Against these issues, we specifically construct a dual-domain reconstruction model and propose a model-driven equivariant proximal network, called MEPNet. The main characteristics of MEPNet are: 1) It is optimization-inspired and has a clear working mechanism; 2) The involved proximal operator is modeled via a rotation equivariant convolutional neural network, which finely represents the inherent rotational prior underlying the CT scanning that the same organ can be imaged at different angles. Extensive experiments conducted on several datasets comprehensively substantiate that compared with the conventional convolution-based proximal network, such a rotation equivariance mechanism enables our proposed method to achieve better reconstruction performance with fewer network parameters. We will release the code at \url{https://github.com/hongwang01/MEPNet}.