Untrained Perceptual Loss for image denoising of line-like structures in MR images

TL;DR

This paper introduces an untrained perceptual loss (uPL) method for 3D MR image denoising, leveraging feature maps of untrained networks to improve denoising of line-like structures such as vessels and roots, outperforming traditional loss functions.

Contribution

The study adapts perceptual loss to 3D data using untrained networks and demonstrates its effectiveness for denoising line-like structures in MR images, with insights on network architecture impacts.

Findings

uPL outperforms L1 and SSIM-based losses in denoising quality

Network depth and pooling significantly affect performance

Small uPL networks achieve comparable or better results than large networks

Abstract

In the acquisition of Magnetic Resonance (MR) images shorter scan times lead to higher image noise. Therefore, automatic image denoising using deep learning methods is of high interest. MR images containing line-like structures such as roots or vessels yield special characteristics as they display connected structures and yield sparse information. For this kind of data, it is important to consider voxel neighborhoods when training a denoising network. In this paper, we translate the Perceptual Loss to 3D data by comparing feature maps of untrained networks in the loss function as done previously for 2D data. We tested the performance of untrained Perceptual Loss (uPL) on 3D image denoising of MR images displaying brain vessels (MR angiograms - MRA) and images of plant roots in soil. We investigate the impact of various uPL characteristics such as weight initialization, network depth, kernel size, and pooling operations on the results. We tested the performance of the uPL loss on four Rician noise levels using evaluation metrics such as the Structural Similarity Index Metric (SSIM). We observe, that our uPL outperforms conventional loss functions such as the L1 loss or a loss based on the Structural Similarity Index Metric (SSIM). The uPL network's initialization is not important, while network depth and pooling operations impact denoising performance. E.g. for both datasets a network with five convolutional layers led to the best performance while a network with more layers led to a performance drop. We also find that small uPL networks led to better or comparable results than using large networks such as VGG. We observe superior performance of our loss for both datasets, all noise levels, and three network architectures. In conclusion, for images containing line-like structures, uPL is an alternative to other loss functions for 3D image denoising.