Validation and Generalizability of Self-Supervised Image Reconstruction Methods for Undersampled MRI
Center for Biomedical Imaging（生物医学成像中心）；Siemens Healthineers International AG（西门子医疗国际公司）；Lausanne University Hospital（洛桑大学医院）；University of Lausanne（洛桑大学）；sitem-insel；University of Bern（伯尔尼大学）
Thomas Yu, Tom Hilbert, Gian Franco Piredda, Arun Joseph, Gabriele Bonanno, Salim Zenkhri, Patrick Omoumi, Meritxell Bach Cuadra, Erick Jorge Canales-Rodríguez,… 展开作者
Thomas Yu, Tom Hilbert, Gian Franco Piredda, Arun Joseph, Gabriele Bonanno, Salim Zenkhri, Patrick Omoumi, Meritxell Bach Cuadra, Erick Jorge Canales-Rodríguez, Tobias Kober, Jean-Philippe Thiran
Deep learning methods have become the state of the art for undersampled MR reconstruction. Particularly for cases where it is infeasible or impossible for ground truth, fully sampled data to be acquired, self-supervised machine learning methods for reconstruction are becoming increasingly used. However potential issues in the validation of such methods, as well as their generalizability, remain underexplored. In this paper, we investigate important aspects of the validation of self-supervised algorithms for reconstruction of undersampled MR images: quantitative evaluation of prospective reconstructions, potential differences between prospective and retrospective reconstructions, suitability of commonly used quantitative metrics, and generalizability. Two self-supervised algorithms based on self-supervised denoising and the deep image prior were investigated. These methods are compared to a least squares fitting and a compressed sensing reconstruction using in-vivo and phantom data. Their generalizability was tested with prospectively under-sampled data from experimental conditions different to the training. We show that prospective reconstructions can exhibit significant distortion relative to retrospective reconstructions/ground truth. Furthermore, pixel-wise quantitative metrics may not capture differences in perceptual quality accurately, in contrast to a perceptual metric. In addition, all methods showed potential for generalization; however, generalizability is more affected by changes in anatomy/contrast than other changes. We further showed that no-reference image metrics correspond well with human rating of image quality for studying generalizability. Finally, we showed that a well-tuned compressed sensing reconstruction and learned denoising perform similarly on all data.
Fast Controllable Diffusion Models for Undersampled MRI Reconstruction
School of Electrical Engineering and Computer Science, University of Queensland（昆士兰大学电气工程与计算机科学学院）；School of Mathematics and Physics, University of Queensland（昆士兰大学数学与物理学院）
Sinogram upsampling using Primal-Dual UNet for undersampled CT and radial MRI reconstruction
Otto von Guericke University Magdeburg（马格德堡奥托·冯·格里克大学）；Human Technopole（人类技术园）；German Centre for Neurodegenerative Disease（德国神经退行性疾病中心）；Centre for Behavioural Brain Sciences（行为脑科学中心）
Undersampled MRI Reconstruction with Side Information-Guided Normalisation
The University of Queensland（昆士兰大学）；The Commonwealth Scientific and Industrial Research Organisation（联邦科学与工业研究组织）；Johns Hopkins University（约翰斯·霍普金斯大学）；University of Science and Technology of China（中国科学技术大学）
A plug-and-play synthetic data deep learning for undersampled magnetic resonance image reconstruction
Institute of Artificial Intelligence, Xiamen University（厦门大学人工智能学院）；Xiamen University（厦门大学）；Department of Electronic Science, Fujian Provincial Key Laboratory of Plasma and Magnetic Resonance, Xiamen University（厦门大学电子科学系）；Department of Microelectronics and Integrated Circuit, Xiamen University（厦门大学微电子与集成电路系）
Modeling human observer detection in undersampled magnetic resonance imaging (MRI) reconstruction with total variation and wavelet sparsity regularization
An Adaptive Intelligence Algorithm for Undersampled Knee MRI Reconstruction
Philips Research（飞利浦研究院）；Philips Healthcare（飞利浦医疗）；Leiden University Medical Center（莱顿大学医学中心）
自适应智能旨在通过额外使用领域知识来增强机器学习技术。在这项工作中，我们提出了应用自适应智能，以加快MR采集。从欠采样的k空间数据出发，采用一种受压缩感知理论启发的基于迭代学习的重建方案来重建图像。我们采用深度神经网络来改进和纠正给定训练数据的先验重建假设。该网络在Facebook AI Research和NYU Langone Health组织的2019年fastMRI挑战赛的膝关节MRI数据集上进行了训练和测试。所有提交的挑战最初根据与已知地面实况的相似性进行排名，之后对前4名提交进行放射学评价。我们的方法由fastMRI组织者在一个独立的挑战数据集上进行了评估。它在8倍加速多线圈、4倍多线圈和4倍单线圈跟踪中分别排名第1、第1和第3。这证明了该方法的优越性能和广泛适用性。
Deep learning-based reconstruction of highly accelerated 3D MRI
GE Research（通用电气研究院）；GE Healthcare（通用电气医疗）；Mayo Clinic（妙佑医疗国际）；Walter Reed National Military Medical Center（沃尔特里德国家军事医疗中心）；Uniformed Services University（美国三军卫生科学大学）
LDPM: Towards undersampled MRI reconstruction with MR-VAE and Latent Diffusion Prior
Shenzhen Technology University（深圳科技大学）；Shenzhen University（深圳大学）；Huawei（华为）；Nanjing University of Science and Technology（南京理工大学）；Qilu University of Technology（齐鲁工业大学）
Seeking Common Ground While Reserving Differences: Multiple Anatomy Collaborative Framework for Undersampled MRI Reconstruction
Tsinghua Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）；Tencent AI Lab（腾讯人工智能实验室）；The Chinese University of Hong Kong（香港中文大学）；University of Texas at Arlington（德克萨斯大学阿灵顿分校）
MRPD: Undersampled MRI reconstruction by prompting a large latent diffusion model
School of Biomedical Engineering, Division of Life Sciences and Medicine, University of Science and Technology of China (USTC)（中国科学技术大学生命科学与医学部生物医学工程学院）；Suzhou Institute for Advance Research, USTC（中国科学技术大学苏州高等研究院）；Key Laboratory of Precision and Intelligent Chemistry, USTC（中国科学技术大学精准与智能化学重点实验室）；Key Laboratory of Intelligent Information Processing of Chinese Academy of Sciences (CAS), Institute of Computing Technology, CAS（中国科学院计算技术研究所中国科学院智能信息处理重点实验室）
Highly Undersampled MRI Reconstruction via a Single Posterior Sampling of Diffusion Models
Central South University（中南大学）；University of Sydney（悉尼大学）；Soochow University（苏州大学）；Nanjing University（南京大学）；University of Queensland（昆士兰大学）；University of Calgary（卡尔加里大学）；University of Newcastle（纽卡斯尔大学）
Real-time cardiac cine MRI -- A comparison of a diffusion probabilistic model with alternative state-of-the-art image reconstruction techniques for undersampled spiral acquisitions
Inference Stage Denoising for Undersampled MRI Reconstruction
School of Engineering, The University of Edinburgh（爱丁堡大学工程学院）；Department of Electrical and Electronic Engineering and I-X, Imperial College London（帝国理工学院电气与电子工程系及I-X）
Quality-aware Cine Cardiac MRI Reconstruction and Analysis from Undersampled k-space Data
School of Biomedical Engineering & Imaging Sciences, King’s College London, UK（伦敦国王学院生物医学工程与成像科学学院）；Technical University of Munich, Germany（慕尼黑工业大学）；Biomedical Image Analysis Group, Imperial College London, UK（帝国理工学院生物医学图像分析组）；Department of Adult and Paediatric Cardiology, Guy’s and St Thomas’ NHS Foundation Trust, London, UK（盖伊和圣托马斯国民健康服务信托基金会成人及儿科心脏病科）；Coimbra Institute for Biomedical Imaging and Translational Research, University of Coimbra, Portugal（科英布拉大学科英布拉生物医学成像与转化研究学院）；Helmholtz Center Munich, Germany（慕尼黑亥姆霍兹中心）
DuDoUniNeXt: Dual-domain unified hybrid model for single and multi-contrast undersampled MRI reconstruction
School of Biomedical Engineering, Division of Life Sciences and Medicine, University of Science and Technology of China（中国科学技术大学生命科学与医学部生物医学工程学院）；Center for Medical Imaging, Robotics, Analytic Computing & Learning (MIRACLE), Suzhou Institute for Advance Research, University of Science and Technology of China（中国科学技术大学苏州高等研究院医学影像、机器人、分析计算与学习中心）；School of Electrical Engineering and Computer Science, The University of Queensland（昆士兰大学电气工程与计算机科学学院）
Multi-Coil MRI Reconstruction Challenge -- Assessing Brain MRI Reconstruction Models and their Generalizability to Varying Coil Configurations
University of Calgary（卡尔加里大学）；Hotchkiss Brain Institute, University of Calgary（卡尔加里大学霍奇基斯脑研究所）；Foothills Medical Center（山麓医学中心）；Radboud University Medical Center（拉德堡德大学医学中心）；Netherlands Cancer Institute（荷兰癌症研究所）；University of Amsterdam（阿姆斯特丹大学）；Amsterdam UMC（阿姆斯特丹大学医学中心）；University of Campinas（坎皮纳斯大学）
Self-Supervised Learning for MRI Reconstruction with a Parallel Network Training Framework
Paul C. Lauterbur Research Center for Biomedical Imaging, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences（中国科学院深圳先进技术研究院保罗·C·劳特伯生物医学成像研究中心）；University of Chinese Academy of Sciences（中国科学院大学）；Department of Electronic Information Engineering, Nanchang University（南昌大学电子信息工程系）；Peng Cheng Laboratory（鹏城实验室）；Pazhou Lab（琶洲实验室）
Model-based Dynamic 3D MRI Reconstructions using Neural Fields and Tensor Product Expansions
Computational Imaging Group for MRI Therapy & Diagnostics, Department of Radiotherapy, University Medical Center Utrecht（MRI治疗与诊断计算成像组，放射治疗系，乌得勒支大学医学中心）
Continuous K-space Recovery Network with Image Guidance for Fast MRI Reconstruction
The Shanghai Key Laboratory of Medical Image Computing and Computer Assisted Intervention（上海医学影像计算与计算机辅助介入重点实验室）；Academy of Engineering and Technology, Fudan University（复旦大学工程与技术学院）
Pixelwise Uncertainty Quantification of Accelerated MRI Reconstruction
Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, NYU Grossman School of Medicine（贝纳德与伊蕾娜·施瓦茨生物医学成像中心，放射学系，纽约大学格罗斯曼医学院）；Courant Institute of Mathematical Sciences, NYU（数学科学学院，纽约大学）；Center for Advanced Imaging Innovation and Research (CAI 2 R), Department of Radiology, NYU Grossman School of Medicine（先进成像创新与研究中心（CAI 2 R），放射学系，纽约大学格罗斯曼医学院）
Conditional Denoising Diffusion Model-Based Robust MR Image Reconstruction from Highly Undersampled Data
Department of Mathematics and Systems Engineering（数学与系统工程系）；Department of Biomedical Engineering & Science（生物医学工程与科学系）；Florida Institute of Technology（佛罗里达理工学院）；A. A. Martinos Center for Biomedical Imaging（A.A. Martinos生物医学成像中心）；Massachusetts General Hospital/Harvard Medical School（麻省总医院/哈佛医学院）；Department of Radiology（放射科）；Embry-Riddle Aeronautical University（艾姆伯-里尔航空大学）
The Challenge of Fetal Cardiac MRI Reconstruction Using Deep Learning
King’s College London（伦敦国王学院）；Technical University of Munich（慕尼黑工业大学）；Imperial College London（帝国理工学院）；Guy’s & St. Thomas’ NHS Foundation Trust（盖伊和圣托马斯国民保健服务基金会信托）；Evelina London Children’s Hospital（伊夫琳娜伦敦儿童医院）
Bayesian Uncertainty Estimation of Learned Variational MRI Reconstruction
Graz University of Technology（格拉茨技术大学）；Silicon Austria Labs（奥地利硅谷实验室）；University of Bonn（波恩大学）；Johannes Kepler University（约翰·开普勒大学）；Technical University of Munich（慕尼黑工业大学）；Imperial College London（伦敦帝国理工学院）；NYU School of Medicine（纽约大学医学院）
