Comparison of Objective Image Quality Metrics to Expert Radiologists’ Scoring of Diagnostic Quality of MR Images

Mason, Allister; Rioux, James A.; Clarke, Sharon E.; Costa, Andreu F.; Schmidt, Matthias; Keough, Valerie; Huynh, Thien; Beyea, Steven

doi:10.1109/tmi.2019.2930338

Cited by 100 publications

(73 citation statements)

References 28 publications

(28 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The authors were able to draw parallels between the quantitative and blind study results, which revealed that, in their data challenge, SSIM was a reasonable estimate for the radiologists’ ranking of the images. In contrast, Mason et al [ 71 ] found differences in their study between several image metrics and experts’ opinions on reconstructed MRI images.…”

Section: Discussionmentioning

confidence: 92%

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

et al. 2021

View full text Add to dashboard Cite

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.

show abstract

Section: Discussionmentioning

confidence: 92%

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

et al. 2021

View full text Add to dashboard Cite

show abstract

“…As an image processing technology, edge detection is a hotspot, and it needs to be optimized in response to the complexity of medical images and the increasing amount of information contained [ 11 , 12 ]. Compared with the classic edge detection algorithm, the improved Prewitt algorithm proposed by Sengupta et al [ 13 ] greatly shortens the image processing time.…”

Section: Introductionmentioning

confidence: 99%

Edge Detection Algorithm-Based Lung Ultrasound in Evaluation of Efficacy of High-Flow Oxygen Therapy on Critical Lung Injury

Lü¹,

Xie

Ding

2022

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

The study focused on the therapeutic effects of high-flow oxygen therapy on patients with critical lung injury using edge detection-based ultrasound images. Firstly, the traditional Canny edge detection algorithm was improved, and the optimal threshold was obtained by optimizing the median filter and combining Otsu algorithm and threshold iteration method. Then, the optimized algorithm was compared with the traditional Canny edge detection algorithm and applied to process the lung ultrasound images of 120 cases of critical lung injury, to compare the efficacy of high-flow oxygen therapy and the traditional oxygen therapy. It was found that the peak signal-to-noise ratio (PSNR) (20.34~31.3), edge intensity value (17.89~27.34), and edge detection effect of the improved Canny algorithm were better than the traditional Canny algorithm (15.2~28.61, 9.44~18.56). The failure rate of extubation (4.1%), reintubation rate (0.8%), comfort ( 2.38 ± 0.15 points), dry humidity score ( 1.07 ± 0.21 points), antibiotic use ( 7.41 ± 0.74 days), and hospital stay ( 8.66 ± 1.02 days) in the experimental group were significantly lower than the corresponding indexes in the control group (11.7%, 5%, 4.25 ± 0.26 minutes, 4.94 ± 0.78 minutes, 19.29 ± 1.7 days, and 27.49 ± 2.22 days), and the difference was statistically significant ( P < 0.05 ). In the experimental group, within 48 hours after extubation, the respiratory rate (RR), heart rate (HR), arterial partial pressure of carbon dioxide (PaCO2), and HCO3- were significantly lower than those of the control group; and the values of transcutaneous oxygen saturation (SpO2), mean arterial pressure (MAP), arterial partial pressure of oxygen (PaO2), and pH were significantly higher than the control group, and the difference was statistically significant ( P < 0.05 ). In conclusion, the algorithm in this study is superior to the traditional Canny algorithm, and the high-flow oxygen therapy can reduce the failure rate of extubation, strengthen patient comfort, improve the degree of gas humidification, stabilize the respiratory function and circulatory system, and shorten the time of antibiotic use and hospital stay.

show abstract

“…Visual inspection of the images agrees with the quantitative results. It is known that the VIF is highly correlated with the radiologist assessment of the image quality [49]. Although the VIF values of KIGAN are higher than that of RefineGAN, if we assume that an image with VIF>0.85 is good enough to be introduced into the clinical settings, there would be almost no differences between RefineGAN and KIGAN in terms of visual quality.…”

Section: Discussionmentioning

confidence: 99%

Which GAN? A comparative study of generative adversarial network-based fast MRI reconstruction

Zhu

Yang

2021

Phil. Trans. R. Soc. A.

View full text Add to dashboard Cite

Fast magnetic resonance imaging (MRI) is crucial for clinical applications that can alleviate motion artefacts and increase patient throughput. K -space undersampling is an obvious approach to accelerate MR acquisition. However, undersampling of k -space data can result in blurring and aliasing artefacts for the reconstructed images. Recently, several studies have been proposed to use deep learning-based data-driven models for MRI reconstruction and have obtained promising results. However, the comparison of these methods remains limited because the models have not been trained on the same datasets and the validation strategies may be different. The purpose of this work is to conduct a comparative study to investigate the generative adversarial network (GAN)-based models for MRI reconstruction. We reimplemented and benchmarked four widely used GAN-based architectures including DAGAN, ReconGAN, RefineGAN and KIGAN. These four frameworks were trained and tested on brain, knee and liver MRI images using twofold, fourfold and sixfold accelerations, respectively, with a random undersampling mask. Both quantitative evaluations and qualitative visualization have shown that the RefineGAN method has achieved superior performance in reconstruction with better accuracy and perceptual quality compared to other GAN-based methods. This article is part of the theme issue ‘Synergistic tomographic image reconstruction: part 1’.

show abstract

Comparison of Objective Image Quality Metrics to Expert Radiologists’ Scoring of Diagnostic Quality of MR Images

Cited by 100 publications

References 28 publications

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

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

Edge Detection Algorithm-Based Lung Ultrasound in Evaluation of Efficacy of High-Flow Oxygen Therapy on Critical Lung Injury

Which GAN? A comparative study of generative adversarial network-based fast MRI reconstruction

Contact Info

Product

Resources

About