Sparse Mixture-of-Experts for Non-Uniform Noise Reduction in MRI Images

TL;DR

This paper introduces a sparse mixture-of-experts neural network framework for MRI image denoising, effectively handling non-uniform noise and outperforming existing methods on various datasets.

Contribution

It presents a novel sparse mixture-of-experts approach with specialized CNN experts for region-specific noise reduction in MRI images.

Findings

Outperforms state-of-the-art denoising techniques

Effective on synthetic and real-world datasets

Generalizes well to unseen data

Abstract

Magnetic Resonance Imaging (MRI) is an essential diagnostic tool in clinical settings but its utility is often hindered by noise artifacts introduced during the imaging process. Effective denoising is critical for enhancing image quality while preserving anatomical structures. However traditional denoising methods which typically assume uniform noise distributions struggle to handle the non-uniform noise commonly present in MRI images. In this paper we introduce a novel approach leveraging a sparse mixture-of-experts framework for MRI image denoising. Each expert is a specialized denoising convolutional neural network fine-tuned to target specific noise characteristics associated with different image regions. Our method demonstrates superior performance over state-of-the-art denoising techniques on both synthetic and real-world MRI datasets. Furthermore we show that it generalizes effectively to unseen datasets highlighting its robustness and adaptability.