Self-supervised Noise2noise Method Utilizing Corrupted Images with a Modular Network for LDCT Denoising

TL;DR

This paper introduces a self-supervised LDCT denoising method that uses corrupted images and a modular U-Net, eliminating the need for paired clean images and outperforming existing techniques.

Contribution

It presents a novel self-supervised noise2noise approach with a modular U-Net for LDCT denoising using only LDCT data, avoiding reliance on clean images.

Findings

Effective denoising on Mayo LDCT dataset

Outperforms state-of-the-art deep learning methods

Uses only LDCT data without paired clean images

Abstract

Deep learning is a very promising technique for low-dose computed tomography (LDCT) image denoising. However, traditional deep learning methods require paired noisy and clean datasets, which are often difficult to obtain. This paper proposes a new method for performing LDCT image denoising with only LDCT data, which means that normal-dose CT (NDCT) is not needed. We adopt a combination including the self-supervised noise2noise model and the noisy-as-clean strategy. First, we add a second yet similar type of noise to LDCT images multiple times. Note that we use LDCT images based on the noisy-as-clean strategy for corruption instead of NDCT images. Then, the noise2noise model is executed with only the secondary corrupted images for training. We select a modular U-Net structure from several candidates with shared parameters to perform the task, which increases the receptive field without increasing the parameter size. The experimental results obtained on the Mayo LDCT dataset show the effectiveness of the proposed method compared with that of state-of-the-art deep learning methods. The developed code is available at https://github.com/XYuan01/Self-supervised-Noise2Noise-for-LDCT.