Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Apr;33(4):2686-2698.
doi: 10.1007/s00330-022-09243-y. Epub 2022 Nov 15.

MR-self Noise2Noise: self-supervised deep learning-based image quality improvement of submillimeter resolution 3D MR images

Woojin Jung  1 , Hyun-Soo Lee  2 , Minkook Seo  3 , Yoonho Nam  4 , Yangsean Choi  3 , Na-Young Shin  3 , Kook-Jin Ahn  3 , Bum-Soo Kim  3 , Jinhee Jang  5
Affiliations

MR-self Noise2Noise: self-supervised deep learning-based image quality improvement of submillimeter resolution 3D MR images

Woojin Jung et al. Eur Radiol. 2023 Apr.

Abstract

Objectives: The study aimed to develop a deep neural network (DNN)-based noise reduction and image quality improvement by only using routine clinical scans and evaluate its performance in 3D high-resolution MRI.

Methods: This retrospective study included T1-weighted magnetization-prepared rapid gradient-echo (MP-RAGE) images from 185 clinical scans: 135 for DNN training, 11 for DNN validation, 20 for qualitative evaluation, and 19 for quantitative evaluation. Additionally, 18 vessel wall imaging (VWI) data were included to evaluate generalization. In each scan of the DNN training set, two noise-independent images were generated from the k-space data, resulting in an input-label pair. 2.5D U-net architecture was utilized for the DNN model. Qualitative evaluation between conventional MP-RAGE and DNN-based MP-RAGE was performed by two radiologists in image quality, fine structure delineation, and lesion conspicuity. Quantitative evaluation was performed with full sampled data as a reference by measuring quantitative error metrics and volumetry at 7 different simulated noise levels. DNN application on VWI was evaluated by two radiologists in image quality.

Results: Our DNN-based MP-RAGE outperformed conventional MP-RAGE in all image quality parameters (average scores = 3.7 vs. 4.9, p < 0.001). In the quantitative evaluation, DNN showed better error metrics (p < 0.001) and comparable (p > 0.09) or better (p < 0.02) volumetry results than conventional MP-RAGE. DNN application to VWI also revealed improved image quality (3.5 vs. 4.6, p < 0.001).

Conclusion: The proposed DNN model successfully denoises 3D MR image and improves its image quality by using routine clinical scans only.

Key points: • Our deep learning framework successfully improved conventional 3D high-resolution MRI in all image quality parameters, fine structure delineation, and lesion conspicuity. • Compared to conventional MRI, the proposed deep neural network-based MRI revealed better quantitative error metrics and comparable or better volumetry results. • Deep neural network application to 3D MRI whose pulse sequences and parameters were different from the training data showed improvement in image quality, revealing the potential to generalize on various clinical MRI.

Keywords: Deep learning; High-resolution; Magnetic resonance imaging; Quality improvement.

PubMed Disclaimer

References

    1. Obusez EC, Hui F, Hajj-ali RA et al (2014) High-resolution MRI vessel wall imaging: spatial and temporal patterns of reversible cerebral vasoconstriction syndrome and central nervous system vasculitis. AJNR Am J Neuroradiol 35:1527–1532. https://doi.org/10.3174/ajnr.a3909 - DOI - PubMed - PMC
    1. Pinker K, Noebauer-Huhmann IM, Stavrou I et al (2007) High-resolution contrast-enhanced, susceptibility-weighted MR imaging at 3T in patients with brain tumors: correlation with positron-emission tomography and histopathologic findings. AJNR Am J Neuroradiol 28:1280–1286. https://doi.org/10.3174/ajnr.a0540 - DOI - PubMed - PMC
    1. Noh Y, Sung YH, Lee J, Kim EY (2015) Nigrosome 1 detection at 3T MRI for the diagnosis of early-stage idiopathic parkinson disease: assessment of diagnostic accuracy and agreement on imaging asymmetry and clinical laterality. AJNR Am J Neuroradiol 36:2010–2016. https://doi.org/10.3174/ajnr.a4412 - DOI - PubMed - PMC
    1. Buades A, Coll B, Morel JM (2005) A review of image denoising algorithms, with a new one. Multiscale Model Simul 4:490–530. https://doi.org/10.1137/040616024 - DOI
    1. Zhang K, Zuo W, Chen Y et al (2017) Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising. IEEE Trans Image Process 26:3142–3155. https://doi.org/10.1109/tip.2017.2662206 - DOI - PubMed

MeSH terms

LinkOut - more resources