Quantitative Comparison of Deep Learning-Based Image Reconstruction Methods for Low-Dose and Sparse-Angle CT Applications.

Johannes Leuschner,Maximilian Schmidt,Maureen Van Eijnatten,Amir Hadjifaradji,Poulami Somanya Ganguly,Dominik Bauer,Sophia Bethany Coban,Kees Joost Batenburg,Vladyslav Andriiashen,Peter Maass,Alexander Denker

doi:10.3390/jimaging7030044

Abstract

The reconstruction of computed tomography (CT) images is an active area of research. Following the rise of deep learning methods, many data-driven models have been proposed in recent years. In this work, we present the results of a data challenge that we organized, bringing together algorithm experts from different institutes to jointly work on quantitative evaluation of several data-driven methods on two large, public datasets during a ten day sprint. We focus on two applications of CT, namely, low-dose CT and sparse-angle CT. This enables us to fairly compare different methods using standardized settings. As a general result, we observe that the deep learning-based methods are able to improve the reconstruction quality metrics in both CT applications while the top performing methods show only minor differences in terms of peak signal-to-noise ratio (PSNR) and structural similarity (SSIM). We further discuss a number of other important criteria that should be taken into account when selecting a method, such as the availability of training data, the knowledge of the physical measurement model and the reconstruction speed.

Highlights

We focus on two reconstruction tasks with high relevance and impact—the first task is the reconstruction of low-dose medical computed tomography (CT) images, and the second is the reconstruction of sparse-angle CT images
Ten different reconstruction methods were evaluated on the challenge set of the low-dose parallel beam (LoDoPaB)-CT dataset
In order to assess the quality of the reconstructions, the peak signal-to-noise ratio (PSNR) and the structural similarity (SSIM) were calculated

Summary

Introduction

Computed tomography (CT) is a widely used (bio)medical imaging modality, with various applications in clinical settings, such as diagnostics [1], screening [2] and virtual treatment planning [3,4], as well as in industrial [5] and scientific [6,7,8] settings. One of the fundamental aspects of this modality is the reconstruction of images from multiple. Because each X-ray measurement exposes the sample or patient to harmful ionizing radiation, minimizing this exposure remains an active area of research [9]. The challenge is to either minimize the dose per measurement or the total number of measurements while maintaining sufficient image quality to perform subsequent diagnostic or analytic tasks

Objectives

Methods

Results

Discussion

Conclusion