Progressive $\mathcal{J}$-Invariant Self-supervised Learning for Low-Dose CT Denoising

TL;DR

This paper introduces a progressive $ ext{J}$-invariant self-supervised learning method for low-dose CT denoising, improving training efficiency and denoising quality without relying on paired data.

Contribution

It proposes a step-wise blind-spot mechanism and noise regularization strategy to enhance self-supervised LDCT denoising performance.

Findings

Outperforms existing self-supervised methods on Mayo LDCT dataset.

Achieves comparable or better results than some supervised denoising methods.

Demonstrates improved training efficiency and denoising quality.

Abstract

Self-supervised learning has been increasingly investigated for low-dose computed tomography (LDCT) image denoising, as it alleviates the dependence on paired normal-dose CT (NDCT) data, which are often difficult to collect. However, many existing self-supervised blind-spot denoising methods suffer from training inefficiencies and suboptimal performance due to restricted receptive fields. To mitigate this issue, we propose a novel Progressive $\mathcal{J}$-invariant Learning that maximizes the use of $\mathcal{J}$-invariant to enhance LDCT denoising performance. We introduce a step-wise blind-spot denoising mechanism that enforces conditional independence in a progressive manner, enabling more fine-grained learning for denoising. Furthermore, we explicitly inject a combination of controlled Gaussian and Poisson noise during training to regularize the denoising process and mitigate overfitting. Extensive experiments on the Mayo LDCT dataset demonstrate that the proposed method consistently outperforms existing self-supervised approaches and achieves performance comparable to, or better than, several representative supervised denoising methods.