본문 바로가기
KERD
논문KCI 등재

A Comparative Study of Unsupervised Deep Learning Methods for MRI Reconstruction

Zhuonan He|Cong Quan|Siyuan Wang|Yuanzheng Zhu|Minghui Zhang|Yanjie Zhu|Qiegen Liu
대한자기공명의과학회 (Korean Society Of Magnetic Resonance In Medicine)|2020.12|Investigative Magnetic Resonance Imaging|제24권 제4호|pp. 179-195

2

조회

0

다운로드

0

인용

0

참고문헌

키워드

Data distributionPrior informationMRI reconstructionUnsupervised deep learning
주제분류방사선과학

초록

Recently, unsupervised deep learning methods have shown great potential in image processing. Compared with a large-amount demand for paired training data of supervised methods with a specific task, unsupervised methods can learn a universal and explicit prior information on data distribution and integrate it into the reconstruction process. Therefore, it can be used in various image reconstruction environments without showing degraded performance. The importance of unsupervised learning in MRI reconstruction appears to be growing. Nevertheless, the establishment of prior formulation in unsupervised deep learning varies a lot depending on mathematical approximation and network architectures. In this work, we summarized basic concepts of unsupervised deep learning comprehensively and compared performances of several state-of-the-art unsupervised learning methods for MRI reconstruction.

Abstract

Recently, unsupervised deep learning methods have shown great potential in image processing. Compared with a large-amount demand for paired training data of supervised methods with a specific task, unsupervised methods can learn a universal and explicit prior information on data distribution and integrate it into the reconstruction process. Therefore, it can be used in various image reconstruction environments without showing degraded performance. The importance of unsupervised learning in MRI reconstruction appears to be growing. Nevertheless, the establishment of prior formulation in unsupervised deep learning varies a lot depending on mathematical approximation and network architectures. In this work, we summarized basic concepts of unsupervised deep learning comprehensively and compared performances of several state-of-the-art unsupervised learning methods for MRI reconstruction.

참고문헌 (72)

  1. Heidemann RM, "A brief review of parallel magnetic resonance imaging", Eur Radiol, 13, pp.2323~2337 (2003)
  2. Donoho DL, "Compressed sensing", IEEE Trans Inf Theory, 52, pp.1289~1306 (2006)
  3. Lustig M, "Compressed sensing MRI", IEEE Signal Proc Mag, 25, pp.72~82 (2008)
  4. Liang D, "Accelerating SENSE using compressed sensing", Magn Reson Med, 62, pp.1574~1584 (2009)
  5. Huang J, "Efficient MR image reconstruction for compressed MR imaging", Med Image Anal, 15, pp.670~679 (2011)
  6. Lin FH, "Parallel imaging reconstruction using automatic regularization", Magn Reson Med, 51, pp.559~567 (2004)
  7. Dong W, "Compressive sensing via nonlocal low-rank regularization", IEEE Trans Image Process, 23, pp.3618~3632 (2014)
  8. Eksioglu EM, "Decoupled algorithm for MRI reconstruction using nonlocal block matching model : BM3D-MRI", J Math Imaging Vis, 56, pp.430~440 (2016)
  9. Ravishankar S, "MR image reconstruction from highly undersampled k-space data by dictionary learning", IEEE Trans Med Imaging, 30, pp.1028~1041 (2011)
  10. Huang Y, "Bayesian nonparametric dictionary learning for compressed sensing MRI", IEEE Trans Image Process, 23, pp.5007~5019 (2014)
  11. Wang S, "Accelerating magnetic resonance imaging via deep learning", Proc IEEE Int Symp Biomed Imaging, pp.514~517 (2016)
  12. Ronneberger O, "U-net: convolutional networks for biomedical image segmentation", MICCAI 2015: Medical Image Computing and Computer-Assisted Intervention, pp.234~241 (2015)
  13. Lee D, "Deep residual learning for compressed sensing MRI", Proc IEEE Int Symp Biomed Imaging, pp.15~18 (2017)
  14. Schlemper J, "A deep cascade of convolutional neural networks for MR image reconstruction", IMPI 2017: Information Processing in Medical Imaging, pp.647~658 (2017)
  15. Aggarwal HK, "MoDL : model-based deep learning architecture for inverse problems", IEEE Trans Med Imaging, 38, pp.394~405 (2019)
  16. Liu Q, "Highly undersampled magnetic resonance imaging reconstruction using autoencoding priors", Magn Reson Med, 83, pp.322~336 (2020)
  17. Zhang M, "High-dimensional embedding network derived prior for compressive sensing MRI reconstruction", Med Image Anal, 64, pp.101717~ (2020)
  18. Tezcan KC, "MR image reconstruction using deep density Priors", IEEE Trans Med Imaging, 38, pp.1633~1642 (2019)
  19. Mardani M, "Deep generative adversarial neural networks for compressive sensing MRI", IEEE Trans Med Imaging, 38, pp.167~179 (2019)
  20. Mardani M, "Recurrent generative adversarial networks for proximal learning and automated compressive image recovery", arXiv preprint arXiv 2017:1711.10046, ~ (2017)
  21. Luo G, "MRI reconstruction using deep Bayesian estimation", Magn Reson Med, 84, pp.2246~2261 (2020)
  22. Hussein S, "Lung and Pancreatic Tumor Characterization in the deep learning era : novel supervised and unsupervised learning approaches", IEEE Trans Med Imaging, 38, pp.1777~1787 (2019)
  23. Antun V, "On instabilities of deep learning in image reconstruction and the potential costs of AI", Proc Natl Acad Sci U S A &, 117, pp.30088~30095 (2020)
  24. Hastie T, "The elements of statistical learning: data mining, inference, and prediction" (2009), Springer Science & Business Media
  25. Bengio Y, "Unsupervised feature learning and deep learning : a review and new perspectives", CoRR, abs, pp.1206.5538 (2012), ~
  26. Erhan D, "Why does unsupervised pre-training help deep learning?", JMLR Workshop and Conference Proceedings, pp.201~208 (2010)
  27. Bengio Y, "Scaling learning algorithms towards AI", Large-scale Kernel Machines, 34, pp.1~41 (2007)
  28. Vincent P, "Stacked denoising autoencoders : learning useful representations in a deep network with a local denoising criterion", J Mach Learn Res, 11, pp.3371~3408 (2010)
  29. Vincent P, "Extracting and composing robust features with denoising autoencoders", Proceedings of International Conference on Machine Learning, pp.1096~1103 (2008)
  30. Rifai S, "Contractive auto-encoders: explicit invariance during feature extraction", Proceeding International Conference on Machine Learning (ICML), pp.833~840 (2011)
  31. Lee H, "Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations", Proceeding International Conference on Machine Learning (ICML), pp.609~616 (2009)
  32. Bengio Y, "Deep learning of representations: looking forward", International Conference on Statistical Language and Speech Processing, pp.1~37 (2013)
  33. Rolfe JT, "Discrete variational autoencoders", arXiv preprint arXiv:1609.02200, ~ (2016)
  34. Dilokthanakul N, "Deep unsupervised clustering with gaussian mixture variational autoencoders", arXiv preprint arXiv:1611.02648, ~ (2016)
  35. Kipf TN, "Variational graph auto-encoders", arXiv preprint arXiv:1611.07308, 1611~07308 (2016)
  36. Sutskever I, "Towards principled unsupervised learning", arXiv preprint arXiv:1511.06440, ~ (2015)
  37. Yi Z, "Dualgan: Unsupervised dual learning for image-to-image translation", Proceedings of the IEEE International Conference on Computer Vision (ICCV), pp.2849~2857 (2017)
  38. Yuan Y, "Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops", Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pp.701~710 (2018)
  39. Wang C, "Unsupervised learning for cross-domain medical image synthesis using deformation invariant cycle consistency networks", International Workshop on Simulation and Synthesis in Medical Imaging, pp.52~60 (2018)
  40. Lugmayr A, "Unsupervised learning for real-world super-resolution", IEEE, 2019, 3408~3416
  41. Song Y, "Generative modeling by estimating gradients of the data distribution", 33rd Conference on Neural Information Processing Systems, pp.11918~11930 (2019)
  42. Kingma DP, "Glow: Generative flow with invertible 1x1 convolution", Advances in Neural Information Processing Systems 31 (NeurIPS 2018), pp.10215~10224 (2018)
  43. Oord A, "Pixel recurrent neural networks", arXiv preprint arXiv:1601.06759, ~ (2016)
  44. Quan TM, "Compressed sensing MRI reconstruction using a generative adversarial network with a cyclic loss", IEEE Trans Med Imaging, 37, pp.1488~1497 (2018)
  45. Radford A, "Unsupervised representation learning with deep convolutional generative adversarial networks", arXiv preprint arXiv: 1511.06434, ~ (2015)
  46. Lustig M, "Compressed sensing MRI", IEEE Signal Proc Mag, 25, pp.72~82 (2008)
  47. Lin FH, "Parallel MRI reconstruction using variance partitioning regularization", Magn Reson Med, 58, pp.735~744 (2007)
  48. Akcakaya M, "Compressed sensing with wavelet domain dependencies for coronary MRI : a retrospective study", IEEE Trans Med Imaging, 30, pp.1090~1099 (2011)
  49. Liu Q, "Synthesis-analysis deconvolutional network for compressed sensing", 2017 IEEE International Conference on Image Processing (ICIP), pp.1940~1944 (2017)
  50. Liu Q, "Highly undersampled magnetic resonance image reconstruction using two-level Bregman method with dictionary updating", IEEE Trans Med Imaging, 32, pp.1290~1301 (2013)
  51. Qu X, "Magnetic resonance image reconstruction from undersampled measurements using a patch-based nonlocal operator", Med Image Anal, 18, pp.843~856 (2014)
  52. Xiong J, "A two-stage convolutional sparse prior model for image restoration", J Vis Commun Image R, 48, pp.268~280 (2017)
  53. He J, "Accelerated High-Dimensional MR Imaging With Sparse Sampling Using Low-Rank Tensors", IEEE Trans Med Imaging, 35, pp.2119~2129 (2016)
  54. Hammernik K, "Learning a variational model for compressed sensing MRI reconstruction", Proceedings of the International Society of Magnetic Resonance in Medicine (ISMRM), pp.1088~ (2016)
  55. Mardani M, "Deep generative adversarial networks for compressed sensing automates MRI", arXiv preprint arXiv:1706.00051, ~ (2017)
  56. Ye JC, "Deep convolutional framelets : a general deep learning framework for inverse problems", SIAM J Imaging Sci, 11, pp.991~1048 (2018)
  57. Alain G, "What regularized auto-encoders learn from the data-generating distribution", J Mach Learn Res, 15, pp.3743~3773 (2014)
  58. Nguyen A, "Plug & play generative networks: conditional iterative generation of images in latent space", IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.4467~4477 (2017)
  59. Bigdeli SA, "Image restoration using autoencoding priors", arXiv preprint arXiv:1703.09964, ~ (2017)
  60. Bigdeli SA, "Deep meanshift priors for image restoration", Advances in Neural Information Processing Systems 30 (NIPS 2017), pp.763~772 (2017)
  61. Kingma DP, "Auto-encoding variational bayes", arXiv preprint arXiv:1312.6114, ~ (2013)
  62. Rezende DJ, "Stochastic backpropagation and approximate inference in deep generative models", arXiv preprint arXiv:1401.4082, ~ (2014)
  63. Bishop CM, "Pattern recognition and machine learning", pp.738~ (2006), Springer
  64. Salimans T, "Pixelcnn++:improving the PixelCNN with discretized logistic mixture likelihood and other modifications", arXiv preprint arXiv:1701.05517, ~ (2017)
  65. Larochelle H, "The neural autoregressive distribution estimator", Proceeding International Conference on Artificial Intelligence and Statistics, JMLR Workshop and Conference Proceedings, pp.29~37 (2011)
  66. Grover A, "Flow-GAN: combining maximum likelihood and adversarial learning in generative models", arXiv preprint arXiv:1705.08868, ~ (2017)
  67. Dinh L, "Nice: non-linear independent components estimation", arXiv preprint arXiv:1410.8516, ~ (2014)
  68. Dinh L, "Density estimation using real NVP", arXiv preprint arXiv:1605.08803, ~ (2016)
  69. Ma X, "MaCow: masked convolutional generative flow", Advances in Neural Information Processing Systems 32 (NeurIPS 2019), pp.5893~5902 (2019)
  70. Hoogeboom E, "Emerging convolutions for generative normalizing flows", arXiv preprint arXiv:1901.11137, ~ (2019)
  71. Uecker M, "Image reconstruction by regularized nonlinear inversion--joint estimation of coil sensitivities and image content", Magn Reson Med, 60, pp.674~682 (2008)
  72. Goodfellow I, "Generative adversarial nets", Advances in neural information processing systems 27 (NIPS 2014), pp.2672~2680 (2014)

관련 논문AI추천

Deep Learning in MR Image Processing고진규, 남윤호, 류강현, 윤재연, 이두희, 이진구대한자기공명의과학회 · Investigative Magnetic Resonance Imaging2019 · 제23권 제2호DEMO: Deep MR Parametric Mapping with Unsupervised Multi-Tasking FrameworkDong Liang, Jing Cheng, Yanjie Zhu, Yuanyuan Liu대한자기공명의과학회 · Investigative Magnetic Resonance Imaging2021 · 제25권 제4호A Self-Supervised Learning Framework for Under-Sampling Pattern Design Using Graph Convolution NetworkHuijun Chen, Yuze Li대한자기공명의과학회 · Investigative Magnetic Resonance Imaging2020 · 제24권 제4호Deep Learning in MR Motion Correction: a Brief Review and a New Motion Simulation Tool (view2Dmotion)김동현, Kyu-Jin Jung, Seul Lee, Soozy Jung대한자기공명의과학회 · Investigative Magnetic Resonance Imaging2020 · 제24권 제4호Cascaded Residual Dense Networks for Dynamic MR Imaging with Edge-Enhanced Loss ConstraintDong Liang, Yanjie Zhu, Ziwen Ke대한자기공명의과학회 · Investigative Magnetic Resonance Imaging2020 · 제24권 제4호Blended-Transfer Learning for Compressed-Sensing Cardiac CINE MRIChang-Beom Ahn, Seong Jae Park대한자기공명의과학회 · Investigative Magnetic Resonance Imaging2021 · 제25권 제1호

이 간행물의 다른 논문 (6건)

Deep Learning in MR Motion Correction: a Brief Review and a New Motion Simulation Tool (view2Dmotion)Kyu-Jin Jung, 김동현, Seul Lee, Soozy Jungpp. 196-206Exploring Generalization Capacity of Artificial Neural Network for Myelin Water ImagingDongmyung Shin, 이종호, 오세홍, 이지은, 김응엽, Joon Yul Choipp. 207-213Cascaded Residual Dense Networks for Dynamic MR Imaging with Edge-Enhanced Loss ConstraintDong Liang, Yanjie Zhu, Ziwen Kepp. 214-222Unsupervised Deformable Image Registration Using Polyphase UNet for 3D Brain MRI VolumesAntoinette D. Martin, 김보아, 예종철pp. 223-231A Self-Supervised Learning Framework for Under-Sampling Pattern Design Using Graph Convolution NetworkYuze Li, Huijun Chenpp. 232-240
원문 보기