A Low-dose CT Reconstruction Network Based on TV-regularized OSEM Algorithm

TL;DR

This paper introduces a novel low-dose CT reconstruction method that integrates TV regularization into the EM algorithm, employing view-by-view processing and neural network unrolling for high-quality, efficient image reconstruction.

Contribution

It proposes a new TV-regularized EM algorithm with view-by-view processing and an end-to-end neural network based on unrolled OSEM-CP for improved LDCT reconstruction.

Findings

Outperforms traditional methods in image quality.

Achieves high-quality reconstructions with just one iteration.

Demonstrates superior performance on various datasets.

Abstract

Low-dose computed tomography (LDCT) offers significant advantages in reducing the potential harm to human bodies. However, reducing the X-ray dose in CT scanning often leads to severe noise and artifacts in the reconstructed images, which might adversely affect diagnosis. By utilizing the expectation maximization (EM) algorithm, statistical priors could be combined with artificial priors to improve LDCT reconstruction quality. However, conventional EM-based regularization methods adopt an alternating solving strategy, i.e. full reconstruction followed by image-regularization, resulting in over-smoothing and slow convergence. In this paper, we propose to integrate TV regularization into the ``M''-step of the EM algorithm, thus achieving effective and efficient regularization. Besides, by employing the Chambolle-Pock (CP) algorithm and the ordered subset (OS) strategy, we propose the OSEM-CP algorithm for LDCT reconstruction, in which both reconstruction and regularization are conducted view-by-view. Furthermore, by unrolling OSEM-CP, we propose an end-to-end reconstruction neural network (NN), named OSEM-CPNN, with remarkable performance and efficiency that achieves high-quality reconstructions in just one full-view iteration. Experiments on different models and datasets demonstrate our methods' outstanding performance compared to traditional and state-of-the-art deep-learning methods.