Cardiac motion estimation and segmentation play important roles in quantitatively assessing cardiac function and diagnosing cardiovascular diseases. In this paper, we propose a novel deep learning method for joint estimation of motion and segmentation from cardiac MR image sequences. The proposed network consists of two branches: a cardiac motion estimation branch which is built on a novel unsupervised Siamese style recurrent spatial transformer network, and a cardiac segmentation branch that is based on a fully convolutional network. In particular, a joint multi-scale feature encoder is learned by optimizing the segmentation branch and the motion estimation branch simultaneously. This enables the weakly-supervised segmentation by taking advantage of features that are unsupervisedly learned in the motion estimation branch from a large amount of unannotated data. Experimental results using cardiac MRI images from 220 subjects show that the joint learning of both tasks is complementary and the proposed models outperform the competing methods significantly in terms of accuracy and speed.
Joint Motion Estimation and Segmentation from Undersampled Cardiac MR Image
Department of Computing, Imperial College London, London, UK(英国伦敦帝国理工学院计算机系);Division of Cardiovascular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, UK(英国牛津大学拉德克利夫医学院心血管医学系);NIHR Biomedical Research Centre at Barts, Queen Mary University of London, London, UK(英国伦敦玛丽女王大学巴茨国家健康研究所)
Automatic Segmentation and Disease Classification Using Cardiac Cine MR Images
Image Sciences Institute, University Medical Center Utrecht(乌得勒支大学医学中心影像科学研究所);Department of Radiology, University Medical Center Utrecht(乌得勒支大学医学中心放射科)
Learning Directional Feature Maps for Cardiac MRI Segmentation
School of Electronic Information and Communications, Huazhong University of Science and Technology(华中科技大学电子信息与通信学院);Department of Radiology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology(华中科技大学同济医学院附属协和医院放射科);Hubei Province Key Laboratory of Molecular Imaging(湖北省分子影像重点实验室);HUST-HW Joint Innovation Lab(华中科技大学-华为联合创新实验室)
Automatic Myocardial Segmentation by Using A Deep Learning Network in Cardiac MRI
CONICET - FCEN, Universidad Nacional de Cuyo(阿根廷国家科学与技术研究理事会 - 库约国立大学自然科学学院);Centro Integral de Medicina Nuclear y Radioterapia Bariloche(巴里洛切核医学与放射治疗综合中心)
心脏功能对于不同病理(如二尖瓣返流、缺血、不同步和心肌炎)的预后和治疗至关重要。心脏行为由结构和功能特征决定。在这两种情况下,医学成像研究的分析需要检测和分割心肌。目前,磁共振成像(MRI)是心脏结构和功能最相关和最准确的非侵入性诊断工具之一。在这项工作中,我们建议使用深度学习技术来辅助心脏MRI中心肌分割的自动化。我们提出了几个改进,在本文中以前的工作:我们提出使用Jaccard距离作为优化目标函数,我们集成了剩余学习策略的代码,我们引入了一个批量归一化层训练的全卷积神经网络。我们的研究结果表明,这种架构优于以前的方法基于类似的网络架构,并提供了一个合适的方法心肌分割。我们的基准测试表明,自动心肌分割所需的时间小于22 seg。对于3.1 GHz Intel Core i7中128 x 128 x 13像素的体积。
Left Ventricle Segmentation and Quantification from Cardiac Cine MR Images via Multi-task Learning
National Library of Medicine, National Institutes of Health(美国国立卫生研究院国家医学图书馆);Rochester Institute of Technology(罗切斯特理工学院);TAJ Technologies Inc.(TAJ科技公司)
左心室的分割和各种心脏收缩功能的量化对于心血管疾病的及时诊断和治疗至关重要。传统上,这两项任务是独立处理的。在这里,我们提出了一个基于卷积神经网络的多任务学习方法,同时执行这两个任务,这样,网络学习更好的数据表示,提高泛化性能。问题的概率公式化使得能够在训练过程中学习任务的不确定性,这些不确定性用于自动计算任务的权重。我们对通过STACOM LV分割挑战获得的97个患者4维心脏电影-MRI数据集上从所提出的多任务网络获得的心肌分割与所提供的金标准心肌分割进行了五重交叉验证,获得了$0.849 \pm 0.036$的Dice重叠和$0.274 \pm 0.083$ mm的平均表面距离,同时估计心肌面积,平均绝对差误差为$205\pm198$ mm $^2$。
Using Foundation Models as Pseudo-Label Generators for Pre-Clinical 4D Cardiac CT Segmentation
Yale University(耶鲁大学);University of Pennsylvania(宾夕法尼亚大学);Washington University School of Medicine in St Louis(华盛顿大学圣路易斯医学学院);University of South Carolina School of Medicine(南卡罗来纳大学医学院);University of Colorado Boulder(科罗拉多大学波德分校)
Reverse Imaging for Wide-spectrum Generalization of Cardiac MRI Segmentation
Department of Imaging Physics, Delft University of Technology, Delft, The Netherlands(影像物理系,代尔夫特理工大学,代尔夫特,荷兰);National Heart Lung and Blood Institute, National Institutes of Health, Bethesda, Maryland, USA(国家心肺血液研究所,美国国立卫生研究院,贝塞斯达,马里兰州,美国);Microsoft Research(微软研究院);Cardiovascular Division, The Ohio State University Wexner Medical Center, Columbus, Ohio, USA(心血管科,俄亥俄州立大学韦克斯纳医学中心,哥伦布,俄亥俄州,美国)
Automatic segmentation with detection of local segmentation failures in cardiac MRI
Amsterdam UMC, University of Amsterdam(阿姆斯特丹大学阿姆斯特丹大学医学中心);Amsterdam University Medical Center, University of Amsterdam(阿姆斯特丹大学阿姆斯特丹大学医学中心);University Medical Center Utrecht(乌得勒支大学医学中心)
Improving Myocardium Segmentation in Cardiac CT Angiography using Spectral Information
Image Sciences Institute, University Medical Center Utrecht(乌得勒支大学医学中心影像科学研究所);Department of Radiology, University Medical Center Utrecht(乌得勒支大学医学中心放射科)
Neural Multi-Atlas Label Fusion: Application to Cardiac MR Images
Xi’an Jiaotong University(西安交通大学);National Engineering Laboratory for Big Data Algorithm and Analysis Technology(大数据算法与分析技术国家工程实验室);Shanghai Jiaotong University(上海交通大学)
多图谱分割方法是生物医学领域中应用最广泛的图像分割技术之一。在这类方法中存在两个主要挑战,即,图谱选择和标签融合。在本文中,我们提出了一种新的多图谱分割方法,该方法在深度学习框架中制定多图谱分割,以更好地解决这些挑战。所提出的方法,被称为深度融合网络(DFN),是一种深度架构,它集成了一个特征提取子网和一个非本地的基于补丁的标签融合(NL-PLF)子网在一个单一的网络。通过端到端训练来学习网络参数,以自动学习深度特征,从而在NL-PLF框架中实现最佳性能。所学习的深度特征被进一步用于定义用于图谱选择的相似性度量。通过对SATA-13和LV-09两个公共心脏MR数据集的左心室分割评估,我们的方法在SATA-13数据集上的平均Dice度量(ADM)达到0.833,在LV-09数据集上的心外膜分割ADM达到0.95,与其他自动左心室分割方法相比,我们的方法是有利的。我们还在MICCAI 2013 SATA Segmentation Challenge的Cardiac Atlas Project(CAP)测试集上测试了我们的方法,我们的方法在ADM中达到了0.815,在撰写本文时排名最高。
Automatic diagnosis of cardiac magnetic resonance images based on semi-supervised learning
School of Information and Electrical Engineering, Hunan University of Science and Technology(湖南科技大学信息与电气工程学院);Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences(中国科学院深圳先进技术研究院);Guilin Medical University(桂林医学院);Rucheng County Hospital of Traditional Chinese Medicine(汝城县中医医院)
Self-Supervised Temporal Regularization for Landmark-Based Cardiac Segmentation with Automatic AHA Regional Mapping
Biomedical Image Technologies, ETSI Telecomunicación, Universidad Politécnica de Madrid, Madrid, Spain(生物医学图像技术,电信工程学院,马德里理工大学,马德里,西班牙);Centro de Investigación Biomédica en Red de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Instituto de Salud Carlos III, Madrid, Spain(生物工程、生物材料和纳米医学网络生物医学研究中心(CIBER-BBN),卡洛斯三世健康研究所,马德里,西班牙);APOLO Biotech, Ciudad Autónoma de Buenos Aires, Argentina(APOLO Biotech,布宜诺斯艾利斯自治市,阿根廷);Departamento de Computación, Universidad de Buenos Aires, Ciudad Autónoma de Buenos Aires, Argentina(计算系,布宜诺斯艾利斯大学,布宜诺斯艾利斯自治市,阿根廷);Instituto de Ciencias de la Computación (ICC), CONICET-Universidad de Buenos Aires, Ciudad Autónoma de Buenos Aires, Argentina(计算机科学研究所(ICC),CONICET-布宜诺斯艾利斯大学,布宜诺斯艾利斯自治市,阿根廷)
Joint Motion Correction and Super Resolution for Cardiac Segmentation via Latent Optimisation
Fudan University(复旦大学);Imperial College London(帝国理工学院);University of Edinburgh(爱丁堡大学);MRC London Institute of Medical Sciences(英国医学研究委员会伦敦医学科学研究所);National Heart Research Institute, Singapore(新加坡国家心脏研究所)
Convolutional Feature Noise Reduction for 2D Cardiac MR Image Segmentation
College of Computer Science, Sichuan Normal University(四川师范大学计算机科学学院);School of Computing and Artificial Intelligence, Southwest Jiaotong University(西南交通大学计算机科学与人工智能学院);Visual Computing and Virtual Reality Key Laboratory of Sichuan Province(四川省视觉计算与虚拟现实重点实验室)
Cardiac MRI Segmentation with Strong Anatomical Guarantees
University of Sherbrooke(舍布鲁克大学);Université de Bourgogne Franche-Comté(勃艮第-弗朗什孔泰大学);CArdiac Simulation & Imaging Software(心脏模拟与成像软件公司);Université de Lyon(里昂大学)
Compositional Segmentation of Cardiac Images Leveraging Metadata
School of Electronic Engineering and Computer Science, Queen Mary University of London(伦敦玛丽女王大学电子工程与计算机科学学院);Queen Mary’s Digital Environment Research Institute (DERI)(玛丽女王数字环境研究所);Department of Computer Science, University College London(伦敦大学学院计算机科学系);School of Engineering and Materials Science, Queen Mary University of London(伦敦玛丽女王大学工程与材料科学学院)
PnP-AdaNet: Plug-and-Play Adversarial Domain Adaptation Network with a Benchmark at Cross-modality Cardiac Segmentation
Department of Computer Science and Engineering, The Chinese University of Hong Kong(香港中文大学计算机科学与工程学系);Department of Computing, Imperial College London(帝国理工学院计算机系);Department of Electrical Engineering and Computer Science, University of Michigan(密歇根大学电气工程与计算机科学系);Imsight Medical Technology, Co, Ltd., China(中国医视医疗科技有限公司);School of Data Science, Fudan University, China(复旦大学数据科学学院)
Interpretability of a Deep Learning Model in the Application of Cardiac MRI Segmentation with an ACDC Challenge Dataset
School of Medicine, University College Dublin(都柏林大学学院医学院);School of Computer Science, University College Dublin(都柏林大学学院计算机科学学院);Department of Statistics, University of Wisconsin–Madison(威斯康星大学麦迪逊分校统计系)