JBFnet -- Low Dose CT Denoising by Trainable Joint Bilateral Filtering

TL;DR

JBFnet is a transparent, trainable neural network that implements iterative bilateral filtering for low dose CT denoising, achieving superior noise removal while maintaining interpretability and accountability.

Contribution

This paper introduces JBFnet, a neural network with a novel architecture that combines shallow convolutional filters with iterative bilateral filtering for improved low dose CT denoising.

Findings

JBFnet outperforms state-of-the-art methods like CPCE3D, GAN, and GFnet in noise reduction.

JBFnet maintains high structural fidelity and interpretability due to its low parameter count.

The approach demonstrates effective denoising with a small, understandable model.

Abstract

Deep neural networks have shown great success in low dose CT denoising. However, most of these deep neural networks have several hundred thousand trainable parameters. This, combined with the inherent non-linearity of the neural network, makes the deep neural network diffcult to understand with low accountability. In this study we introduce JBFnet, a neural network for low dose CT denoising. The architecture of JBFnet implements iterative bilateral filtering. The filter functions of the Joint Bilateral Filter (JBF) are learned via shallow convolutional networks. The guidance image is estimated by a deep neural network. JBFnet is split into four filtering blocks, each of which performs Joint Bilateral Filtering. Each JBF block consists of 112 trainable parameters, making the noise removal process comprehendable. The Noise Map (NM) is added after filtering to preserve high level features. We train JBFnet with the data from the body scans of 10 patients, and test it on the AAPM low dose CT Grand Challenge dataset. We compare JBFnet with state-of-the-art deep learning networks. JBFnet outperforms CPCE3D, GAN and deep GFnet on the test dataset in terms of noise removal while preserving structures. We conduct several ablation studies to test the performance of our network architecture and training method. Our current setup achieves the best performance, while still maintaining behavioural accountability.