Multi-Cycle-Consistent Adversarial Networks for Edge Denoising of Computed Tomography Images

TL;DR

This paper introduces MCCAN, a multi-cycle-consistent adversarial network that improves edge denoising in low-dose CT images by modeling intermediate domains and enforcing cycle-consistency, leading to higher quality results.

Contribution

The paper proposes a novel MCCAN framework that incorporates both local and global cycle-consistency for enhanced CT image denoising, outperforming previous methods.

Findings

MCCAN outperforms CCADN in denoising quality.

Both local and global cycle-consistency are crucial for success.

MCCAN requires slightly less computational resources.

Abstract

As one of the most commonly ordered imaging tests, computed tomography (CT) scan comes with inevitable radiation exposure that increases the cancer risk to patients. However, CT image quality is directly related to radiation dose, thus it is desirable to obtain high-quality CT images with as little dose as possible. CT image denoising tries to obtain high dose like high-quality CT images (domain X) from low dose low-quality CTimages (domain Y), which can be treated as an image-to-image translation task where the goal is to learn the transform between a source domain X (noisy images) and a target domain Y (clean images). In this paper, we propose a multi-cycle-consistent adversarial network (MCCAN) that builds intermediate domains and enforces both local and global cycle-consistency for edge denoising of CT images. The global cycle-consistency couples all generators together to model the whole denoising process, while the local cycle-consistency imposes effective supervision on the process between adjacent domains. Experiments show that both local and global cycle-consistency are important for the success of MCCAN, which outperformsCCADN in terms of denoising quality with slightly less computation resource consumption.