CoChat
Sign Up
Cited by 17Advanced Image Processing TechniquesAdvanced MRI Techniques and ApplicationsMedical Image Segmentation Techniques

The DOI matches its Crossref record; 6 of the 6 references it cites were found in a registry or an index.

Deep attention super-resolution of brain magnetic resonance images acquired under clinical protocols

Open access
Li, Bryan M., Castorina, Leonardo V., Valdés Hernández, Maria del C., Clancy, Una, Wiseman, Stewart J., Sakka, Eleni, Storkey, Amos J., Jaime Garcia, Daniela, Cheng, Yajun, Doubal, Fergus, Thrippleton, Michael T., Stringer, Michael, Wardlaw, Joanna M.·2022·Frontiers in Computational Neuroscience
DOI 10.3389/fncom.2022.887633PMC PMC9458316

Published version of a preprint

Open this paper in CoChat — free account

Abstract

Vast quantities of Magnetic Resonance Images (MRI) are routinely acquired in clinical practice but, to speed up acquisition, these scans are typically of a quality that is sufficient for clinical diagnosis but sub-optimal for large-scale precision medicine, computational diagnostics, and large-scale neuroimaging collaborative research. Here, we present a critic-guided framework to upsample low-resolution (often 2D) MRI full scans to help overcome these limitations. We incorporate feature-importance and self-attention methods into our model to improve the interpretability of this study. We evaluate our framework on paired low- and high-resolution brain MRI structural full scans (i.e., T1-, T2-weighted, and FLAIR sequences are simultaneously input) obtained in clinical and research settings from scanners manufactured by Siemens, Phillips, and GE. We show that the upsampled MRIs are qualitatively faithful to the ground-truth high-quality scans ( PSNR = 35.39; MAE = 3.78 E −3; NMSE = 4.32 E −10; SSIM = 0.9852; mean normal-appearing gray/white matter ratio intensity differences ranging from 0.0363 to 0.0784 for FLAIR, from 0.0010 to 0.0138 for T1-weighted and from 0.0156 to 0.074 for T2-weighted sequences). The automatic raw segmentation of tissues and lesions using the super-resolved images has fewer false positives and higher accuracy than those obtained from interpolated images in protocols represented with more than three sets in the training sample, making our approach a strong candidate for practical application in clinical and collaborative research.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.3389/fncom.2022.887633

Cites

6 of 6 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 73

  1. Atkinson, 1997, Automatic correction of motion artifacts in magnetic resonance images using an entropy focus criterion, IEEE Trans. Med Imaging, 16, 903, 10.1109/42.650886 · DOI
  2. Bansal, 2018, “Recycle-gan: unsupervised video retargeting,”, 119
  3. Bergstra, 2012, Random search for hyper-parameter optimization, J. Mach. Learn. Res, 13, 281
  4. Bernal, 2020, “A framework for jointly assessing and reducing imaging artefacts automatically using texture analysis and total variation optimisation for improving perivascular spaces quantification in brain magnetic resonance imaging,”, 171
  5. Bernal, 2019, “Analysis of spatial spectral features of dynamic contrast-enhanced brain magnetic resonance images for studying small vessel disease,”, 282
  6. Bernal, 2021, A four-dimensional computational model of dynamic contrast-enhanced magnetic resonance imaging measurement of subtle blood-brain barrier leakage, Neuroimage, 230, 117786, 10.1016/j.neuroimage.2021.117786 · DOI
  7. Block, 2008, Suppression of mri truncation artifacts using total variation constrained data extrapolation, Int. J. Biomed. Imaging, 2008, 184123, 10.1155/2008/184123 · DOI
  8. Castorina, 2021, Metrics for quality control of results from super-resolution machine-learning algorithms - Data extracted from publications in the period 2017- May 2021, 2017-2021 [dataset]
  9. Chen, “Efficient and accurate mri super-resolution using a generative adversarial network and 3d multi-level densely connected network,”, 91
  10. Chen, “Brain mri super resolution using 3d deep densely connected neural networks,”, 739
  11. Chun, 2019, Mri super-resolution reconstruction for mri-guided adaptive radiotherapy using cascaded deep learning: In the presence of limited training data and unknown translation model, Med. Phys, 46, 4148, 10.1002/mp.13717 · DOI
  12. Clancy, 2021, Rationale and design of a longitudinal study of cerebral small vessel diseases, clinical and imaging outcomes in patients presenting with mild ischaemic stroke: mild stroke study 3, Eur. Stroke J, 6, 81, 10.1177/2396987320929617 · DOI
  13. Clayden, 2011, Tractor: magnetic resonance imaging and tractography with R, J. Stat. Softw, 44, 1, 10.18637/jss.v044.i08 · DOI
  14. Clevert, 2015, Fast and accurate deep network learning by exponential linear units (elus), arXiv preprint arXiv:1511.07289, 10.48550/arXiv.1511.07289 · DOI · Paper page
  15. Cohen, 2018, “Distribution matching losses can hallucinate features in medical image translation,”, 529
  16. Crowe, 2017, Expertise affects inter-observer agreement at peripheral locations within a brain tumor, Front. Psychol, 10.3389/fpsyg.2017.01628 · DOI
  17. Defazio, 2019, Offset sampling improves deep learning based accelerated mri reconstructions by exploiting symmetry, arXiv preprint arXiv:1912.01101, 10.48550/arXiv.1912.01101 · DOI
  18. Donahue, 2018, Adversarial audio synthesis, arXiv preprint arXiv:1802.04208, 10.48550/arXiv.1802.04208 · DOI
  19. Dong, 2015, Image super-resolution using deep convolutional networks, IEEE Trans. Pattern Anal. Mach. Intell, 38, 295, 10.1109/TPAMI.2015.2439281 · DOI · Paper page
  20. Dong, 2016, “Accelerating the super-resolution convolutional neural network,”, 391
  21. Ganzetti, 2016, Intensity inhomogeneity correction of structural mr images: a data-driven approach to define input algorithm parameters, Front. Neuroinform, 10.3389/fninf.2016.00010 · DOI
  22. Glasser, 2016, The human connectome projects neuroimaging approach, Nat. Neurosci, 19, 1175, 10.1038/nn.4361 · DOI · Paper page
  23. Goodfellow, 2014, Generative adversarial networks, arXiv preprint arXiv:1406.2661 · arXiv
  24. Greenspan, 2002, Mri inter-slice reconstruction using super-resolution, Magn. Reson. Imaging, 20, 437, 10.1016/S0730-725X(02)00511-8 · DOI
  25. He, 2015, “Delving deep into rectifiers: surpassing human-level performance on imagenet classification,”, 1026
  26. Hu, 2019, “Runet: a robust unet architecture for image super-resolution,”, 10.1109/CVPRW.2019.00073 · DOI
  27. Isola, 2017, “Image-to-image translation with conditional adversarial networks,”, 1125
  28. Job, 2017, A brain imaging repository of normal structural mri across the life course: brain images of normal subjects (brains), Neuroimage, 144, 299, 10.1016/j.neuroimage.2016.01.027 · DOI
  29. Johnson, 2016, “Perceptual losses for real-time style transfer and super-resolution,”, 694
  30. Kanazawa, 2018, “End-to-end recovery of human shape and pose,”, 7122
  31. Karras, 2017, Progressive growing of gans for improved quality, stability, and variation, arXiv preprint arXiv:1710.10196, 10.48550/arXiv.1710.10196 · DOI · Paper page
  32. Kim, 2016, “Accurate image super-resolution using very deep convolutional networks,”, 1646
  33. Kim, 2010, Single-image super-resolution using sparse regression and natural image prior, IEEE Trans. Pattern Anal. Mach. Intell, 32, 1127, 10.1109/TPAMI.2010.25 · DOI
  34. Ledig, 2017, “Photo-realistic single image super-resolution using a generative adversarial network,”, 4681
  35. Li, 2020, Calciumgan: a generative adversarial network model for synthesising realistic calcium imaging data of neuronal populations, arXiv preprint arXiv:2009.02707, 10.48550/arXiv.2009.02707 · DOI
  36. Li, Neuronal learning analysis using cycle-consistent adversarial networks, arXiv preprint arXiv:2111.13073, 10.48550/arXiv.2111.13073 · DOI
  37. Li, A review of the deep learning methods for medical images super resolution problems, IRBM, 42, 120, 10.1016/j.irbm.2020.08.004 · DOI
  38. “Vision transformers enable fast and robust accelerated mri,” LinK. HeckelR. Proceedings of Machine Learning Research 1-222022
  39. Liu, 2019, “An edge enhanced srgan for mri super resolution in slice-selection direction,”, 12, 10.1007/978-3-030-33226-6_2 · DOI
  40. Lu, 2019, Single image super resolution based on a modified u-net with mixed gradient loss, arXiv preprint arXiv:1911.09428, 10.48550/arXiv.1911.09428 · DOI
  41. Lyu, 2018, Super-resolution mri through deep learning, 10.48550/arXiv.1810.06776 · DOI
  42. Masutani, 2020, Deep learning single-frame and multiframe super-resolution for cardiac mri, Radiology, 295, 552, 10.1148/radiol.2020192173 · DOI
  43. Micikevicius, 2017, Mixed precision training, arXiv preprint arXiv:1710.03740, 10.48550/arXiv.1710.03740 · DOI · Paper page
  44. Modat, 2010, Fast free-form deformation using graphics processing units, Comput. Methods Programs Biomed, 98, 278, 10.1016/j.cmpb.2009.09.002 · DOI
  45. Oktay, 2018, Attention u-net: learning where to look for the pancreas, arXiv preprint arXiv:1804.03999, 10.48550/arXiv.1804.03999 · DOI · Paper page
  46. Paszke, 2019, Pytorch: an imperative style, high-performance deep learning library, arXiv preprint arXiv:1912.01703, 10.48550/arXiv.1912.01703 · DOI · Paper page
  47. Pathak, 2016, “Context encoders: feature learning by inpainting,”, 2536
  48. Pérez-García, 2021, Torchio: a python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning, Comput. Methods Programs Biomed, 2018, 106236, 10.1016/j.cmpb.2021.106236 · DOI
  49. Pham, 2017, “Brain mri super-resolution using deep 3d convolutional networks,”, 197, 10.1109/ISBI.2017.7950500 · DOI
  50. Preston, 2006, Magnetic Resonance Imaging of the Brain and Spine by Scott W. Atlas, 5th Edn
  51. Radford, 2015, Unsupervised representation learning with deep convolutional generative adversarial networks, arXiv preprint arXiv:1511.06434, 10.48550/arXiv.1511.06434 · DOI · Paper page
  52. Selvaraju, 2017, “Grad-cam: visual explanations from deep networks via gradient-based localization,”, 618
  53. Shi, 2018, Super-resolution reconstruction of mr image with a novel residual learning network algorithm, Phys. Med. Biol, 10.1088/1361-6560/aab9e9 · DOI
  54. Smith, 2002, Fast robust automated brain extraction, Hum. Brain Mapp, 17, 143, 10.1002/hbm.10062 · DOI · Paper page
  55. Timofte, 2013, “Anchored neighborhood regression for fast example-based super-resolution,”, 1920
  56. Ulyanov, 2016, Instance normalization: the missing ingredient for fast stylization, arXiv preprint arXiv:1607.08022, 10.48550/arXiv.1607.08022 · DOI
  57. Valdes, 2001, Improvement of remotely sensed low spatial resolution images by back-propagated neural networks using data fusion techniques, Int. J. Remote Sens, 22, 629, 10.1080/01431160050505892 · DOI
  58. Van der Walt, 2014, scikit-image: image processing in python, PeerJ, 10.7717/peerj.453 · DOI
  59. Van Essen, 2012, The human connectome project: a data acquisition perspective, Neuroimage, 62, 2222, 10.1016/j.neuroimage.2012.02.018 · DOI
  60. Wang, 2018, “Esrgan: enhanced super-resolution generative adversarial networks,”
  61. Wardlaw, 2011, Brain aging, cognition in youth and old age and vascular disease in the lothian birth cohort 1936: rationale, design and methodology of the imaging protocol, Int. J. Stroke, 6, 547, 10.1111/j.1747-4949.2011.00683.x · DOI
  62. Wardlaw, 2010, A large web-based observer reliability study of early ischaemic signs on computed tomography. The acute cerebral ct evaluation of stroke study (access), PLoS ONE, 10.1371/journal.pone.0015757 · DOI
  63. Woolrich, 2009, Bayesian analysis of neuroimaging data in fsl, Neuroimage, 45, S173, 10.1016/j.neuroimage.2008.10.055 · DOI · Paper page
  64. Wu, “Learning a probabilistic latent space of object shapes via 3d generative-adversarial modeling,”, 82
  65. Wu, Google's neural machine translation system: bridging the gap between human and machine translation
  66. Yang, 2010, Image super-resolution via sparse representation, IEEE Trans. Image Process, 19, 2861, 10.1109/TIP.2010.2050625 · DOI
  67. You, 2019, Ct super-resolution gan constrained by the identical, residual, and cycle learning ensemble (gan-circle), IEEE Trans. Med. Imaging, 39, 188, 10.1109/TMI.2019.2922960 · DOI
  68. Zbontar, 2018, fastmri: an open dataset and benchmarks for accelerated mri, arXiv preprint arXiv:1811.08839, 10.48550/arXiv.1811.08839 · DOI
  69. Zhan, 2015, Automatic method for white matter lesion segmentation based on t1-fluid-attenuated inversion recovery images, IET Comput. Vis, 9, 447, 10.1049/iet-cvi.2014.0121 · DOI
  70. Zhang, 2001, Segmentation of brain MR images through a hidden markov random field model and the expectation-maximization algorithm, IEEE Trans. Med. Imaging, 20, 45, 10.1109/42.906424 · DOI
  71. Zhang, 2018, “Image super-resolution using very deep residual channel attention networks,”, 286
  72. Zhang, 2019, “Reducing uncertainty in undersampled mri reconstruction with active acquisition,”, 2049
  73. Zhu, 2017, “Unpaired image-to-image translation using cycle-consistent adversarial networks,”, 2223

Metadata

Authors
Li, B. M., Castorina, L. V., Valdés Hernández, M. D. C., Clancy, U., Wiseman, S. J., Sakka, E., Storkey, A. J., Jaime Garcia, D., Cheng, Y., Doubal, F., Thrippleton, M. T., Stringer, M., & Wardlaw, J. M.
Affiliations
  • University of Edinburgh (GB)
  • National Health Service (GB)
ORCID
Venue
Frontiers in Computational Neuroscience
Publisher
Frontiers Media SA
Published
2022-08-25
Volume
16
Pages
887633-887633
ISSN
1662-5188 (ISSN-L)
DOI
10.3389/fncom.2022.887633
PMC
PMC9458316
OpenAlex
W4293088946
Cited by
17
References
6
Citations by year
2026: 5 · 2025: 3 · 2024: 3 · 2023: 6
FWCI
0.98
Type
journal article
Language
en
Open access
Licence
Locations
  • Frontiers in Computational Neuroscience (publishedVersion, cc-by)
  • medRxiv (acceptedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Edinburgh Research Explorer (University of Edinburgh) (submittedVersion, cc-by)
Funders
  • Wellcome Trust
  • UK Research and Innovation
  • China Scholarship Council
  • Mrs Gladys Row Fogo Charitable Trust
  • Stroke Association
  • Fondation Leducq
  • Garfield Weston Foundation
  • Dunhill Medical Trust
  • Medical Research Council
  • Edinburgh and Lothians Health Foundation
Topics
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image Processing Techniques
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Advanced MRI Techniques and Applications
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Medical Image Segmentation Techniques
Keywords
super-resolution, brain MRI, low-resolution MRI, structural MRI, FLAIR, tissue segmentation, lesion segmentation, interpretability, feature importance

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review