Limited Parameter Denoising for Low-dose X-ray Computed Tomography Using Deep Reinforcement Learning

TL;DR

This paper introduces a novel low-dose CT denoising method using bilateral filtering combined with deep reinforcement learning to tune parameters, achieving high-quality results with limited data and fewer parameters.

Contribution

The study proposes a new interpretable CT denoising framework employing reinforcement learning for parameter tuning, outperforming larger deep CNNs with fewer parameters and no artifacts.

Findings

Increased PSNR from 28.53 to 28.93

Increased SSIM from 0.8952 to 0.9204

Outperforms state-of-the-art deep CNNs

Abstract

The use of deep learning has successfully solved several problems in the field of medical imaging. Deep learning has been applied to the CT denoising problem successfully. However, the use of deep learning requires large amounts of data to train deep convolutional networks (CNNs). Moreover, due to large parameter count, such deep CNNs may cause unexpected results. In this study, we introduce a novel CT denoising framework, which has interpretable behaviour, and provides useful results with limited data. We employ bilateral filtering in both the projection and volume domains to remove noise. To account for non-stationary noise, we tune the $\sigma$ parameters of the volume for every projection view, and for every volume pixel. The tuning is carried out by two deep CNNs. Due to impracticality of labelling, the two deep CNNs are trained via a Deep-Q reinforcement learning task. The reward for the task is generated by using a custom reward function represented by a neural network. Our experiments were carried out on abdominal scans for the Mayo Clinic TCIA dataset, and the AAPM Low Dose CT Grand Challenge. Our denoising framework has excellent denoising performance increasing the PSNR from 28.53 to 28.93, and increasing the SSIM from 0.8952 to 0.9204. We outperform several state-of-the-art deep CNNs, which have several orders of magnitude higher number of parameters (p-value (PSNR) = 0.000, p-value (SSIM) = 0.000). Our method does not introduce any blurring, which is introduced by MSE loss based methods, or any deep learning artifacts, which are introduced by WGAN based models. Our ablation studies show that parameter tuning and using our reward network results in the best possible results.