Intracranial Hemorrhage Segmentation Using Deep Convolutional Model
Florida Atlantic University(佛罗里达大西洋大学);University of Technology(理工大学);Babylon Health Directorate(巴比伦卫生局)
Murtadha D. Hssayeni, M. S., Muayad S. Croock, Ph. D., Aymen Al-Ani, Ph. D., Hassan Falah Al-khafaji, M. D., Zakaria A. Yahya, M. D., Behnaz Ghoraani, Ph. D
Traumatic brain injuries could cause intracranial hemorrhage (ICH). ICH could lead to disability or death if it is not accurately diagnosed and treated in a time-sensitive procedure. The current clinical protocol to diagnose ICH is examining Computerized Tomography (CT) scans by radiologists to detect ICH and localize its regions. However, this process relies heavily on the availability of an experienced radiologist. In this paper, we designed a study protocol to collect a dataset of 82 CT scans of subjects with traumatic brain injury. Later, the ICH regions were manually delineated in each slice by a consensus decision of two radiologists. Recently, fully convolutional networks (FCN) have shown to be successful in medical image segmentation. We developed a deep FCN, called U-Net, to segment the ICH regions from the CT scans in a fully automated manner. The method achieved a Dice coefficient of 0.31 for the ICH segmentation based on 5-fold cross-validation. The dataset is publicly available online at PhysioNet repository for future analysis and comparison.
Accurate and Efficient Intracranial Hemorrhage Detection and Subtype Classification in 3D CT Scans with Convolutional and Long Short-Term Memory Neural Networks
University of Bucharest(布加勒斯特大学);Romanian Young Academy(罗马尼亚青年科学院);``Carol Davila'' University of Medicine and Pharmacy(“卡罗尔·达维拉”医药大学)
Vision Transformer for Intracranial Hemorrhage Classification in CT Scans Using an Entropy-Aware Fuzzy Integral Strategy for Adaptive Scan-Level Decision Fusion
School of Electrical and Computer Engineering, College of Engineering, University of Tehran(塔里斯坦大学电气与计算机工程学院);Department of Computer Science and Engineering, Sejong University(世宗大学计算机科学与工程系)
Deep-Learning Tool for Early Identifying Non-Traumatic Intracranial Hemorrhage Etiology based on CT Scan
Capital Medical University(首都医科大学);China National Clinical Research Center for Neurological Diseases(国家神经系统疾病临床医学研究中心);Tencent(腾讯);Shandong Jining Medical College(山东济宁医学院)
背景资料:开发一种人工智能系统,该系统可以基于非造影CT(NCCT)扫描准确识别急性非创伤性颅内出血(ICH)病因,并研究临床医生是否可以在诊断环境中从中受益。材料与方法:深度学习模型是使用2011年1月至2018年4月期间收集的1868例符合条件的非创伤性ICH NCCT扫描结果开发的。我们在2018年4月之后收集的两个独立数据集(TT 200和SD 98)上测试了该模型。该模型的诊断性能与临床医生的性能进行了比较。我们进一步设计了一项模拟研究,以比较临床医生在有和没有深度学习系统增强的情况下的表现。结果如下:所提出的深度学习系统实现了0.986的受试者工作曲线下面积(95% CI 0.967-1.000)动脉瘤,0.952(0.917-0.987)高血压出血,0.950(0.860-1.000)动静脉畸形(AVM),0.749 TT 200数据集中Moyamoya病(MMD)为0.586-0.912,海绵状血管畸形(CM)为0.837(0.704-0.969),其他原因为0.839(0.722-0.959)。在90%的特异性水平下,我们的模型对动脉瘤和AVM诊断的敏感性分别为97.1%和90.9%。该模型在独立数据集SD 98中也显示出令人印象深刻的普遍性。临床医生通过拟议的系统增强在某些出血病因诊断的灵敏度、特异性和准确性方面实现了显著改善。结论:提出的深度学习算法可以成为基…
Zero-Shot Multi-modal Large Language Model v.s. Supervised Deep Learning: A Comparative Study on CT-Based Intracranial Hemorrhage Subtyping
Image Processing Center, Beihang University(北京航空航天大学图像处理中心);School of Mechanical Engineering and Automation, Beihang University(北京航空航天大学机械工程与自动化学院);Department of Neurology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology(华中科技大学同济医学院神经内科,同济医院)
Slice-level Detection of Intracranial Hemorrhage on CT Using Deep Descriptors of Adjacent Slices
Vingroup Big Data Institute(Vingroup大数据研究院);VinBigData JSC(VinBigData股份公司);VinUni-Illinois Smart Health Center, College of Engineering & Computer Science(VinUni-伊利诺伊智慧健康中心,工程与计算机科学学院)
Examining Deployment and Refinement of the VIOLA-AI Intracranial Hemorrhage Model Using an Interactive NeoMedSys Platform
Computational Radiology & Artificial Intelligence Unit, Department of Physics and Computational Radiology, Clinic for Radiology and Nuclear Medicine(计算放射学与人工智能单元,物理与计算放射学部门,放射科与核医学诊所);Oslo University Hospital(奥斯陆大学医院);Department of Physics(物理部门);University of Oslo(奥斯陆大学);Department of Medical Biophysics(医学生物物理学部门);University of Toronto(多伦多大学)
AI-Driven Radiology Report Generation for Traumatic Brain Injuries
Higher School of Digital Economics, Manouba, University of Manouba, Tunisia(数字经济学高等学院,曼布巴,曼布巴大学,突尼斯);Higher Institute of Management of Tunis, University of Tunis, Tunisia(突尼斯管理高等学院,突尼斯大学,突尼斯);Emergency Department Charles Nicolle Hospital, Tunis El Manar University, Tunisia(查尔斯·尼科勒医院急诊部,突尼斯埃尔曼纳大学,突尼斯)
Multi-task Learning Approach for Intracranial Hemorrhage Prognosis
Advanced Computing and e-Science Group, Institute of Physics of Cantabria (IFCA), CSIC - UC(先进计算与电子科学组,坎塔布里亚物理研究所(IFCA),CSIC-UC);AI Technology for Life, Department of Computing and Information Sciences, Utrecht University(人工智能技术与生命,计算与信息科学系,乌得勒支大学);Department of Radiology, The Netherlands Cancer Institute(放射科,荷兰癌症研究所);Servicio de Radiología, Hospital Reina Sofía(放射科,雷娜·索菲亚医院);Servicio de Radiología, Hospital Universitario Rey Juan Carlos(放射科,雷昂·卡洛斯大学医院);Departamento de Tecnologías de La Información, Universidad CEU San Pablo(信息技术部,CEU桑帕布洛大学)
SWDL: Stratum-Wise Difference Learning with Deep Laplacian Pyramid for Semi-Supervised 3D Intracranial Hemorrhage Segmentation
School of Computer Science and Engineering, Southeast University(东南大学计算机科学与工程学院);Research Center for Medical Artificial Intelligence, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences(中国科学院深圳先进技术研究院医学人工智能研究中心);Beijing Tiantan Hospital, Capital Medical University(首都医科大学北京天坛医院)
Smooth Attention for Deep Multiple Instance Learning: Application to CT Intracranial Hemorrhage Detection
Northwestern University(西北大学);University of Granada(格拉纳达大学);Research Centre for Information and Communication Technologies (CITIC-UGR)(信息与通信技术研究中心(CITIC-UGR))
Aneumo: A Large-Scale Comprehensive Synthetic Dataset of Aneurysm Hemodynamics
Artificial Intelligence Innovation and Incubation Institute, Fudan University(复旦大学人工智能创新与孵化研究院);Shanghai Academy of Artificial Intelligence for Science(上海科学人工智能研究院)
Extracting 2D weak labels from volume labels using multiple instance learning in CT hemorrhage detection
Henry Jackson Foundation(亨利·杰克逊基金会);National Institute of Health(美国国立卫生研究院);Middle Tennessee State University(中田纳西州立大学);Vanderbilt University(范德堡大学);Vanderbilt University Medical Center(范德堡大学医学中心);VA Tennessee Valley Healthcare System(弗吉尼亚州田纳西谷医疗系统);Department of Veterans Affairs Medical Center(退伍军人事务部医疗中心)
多实例学习(MIL)是一种监督学习方法,旨在允许模型从袋子类标签中学习实例类标签,其中袋子被定义为包含多个实例。MIL在从弱标签学习方面越来越受欢迎,但尚未广泛应用于3D医学成像。MIL非常适合临床CT采集,因为(1)高度各向异性的体素阻碍了传统3D网络的应用,以及(2)基于块的网络学习整个体积标签的能力有限。在这项工作中,我们将MIL与深度卷积神经网络一起应用,以识别临床CT头部图像体积是否具有一个或多个大尺寸(> 20 cm $^3$),从而在无需2D切片注释的情况下获得学习的2D模型。单个图像体积被视为单独的包,每个体积中的切片是实例。这样的框架为结合临床报告中获得的信息以帮助训练2D分割方法奠定了基础。在这种情况下,我们评估的数据要求,使MIL的泛化通过改变量的训练数据。我们的研究结果表明,需要至少400个患者图像体积的训练大小来实现精确的每层出血检测。在五重交叉验证中,使用最大训练卷数量的领先模型的平均真阳性率为98.10%,平均真阴性率为99.36%,平均精度为0.9698。这些模型已与源代码一起提供,以使MIL在CT神经成像中的持续探索和适应成为可能。
A Prospective Randomized Clinical Trial for Measuring Radiology Study Reporting Time on Artificial Intelligence-Based Detection of Intracranial Hemorrhage in Emergent Care Head CT
Self-supervised Skull Reconstruction in Brain CT Images with Decompressive Craniectomy
Research Institute for Signals, Systems and Computational Intelligence, sinc(i), CONICET, FICH-UNL(信号、系统与计算智能研究所 (sinc(i)),国家科学与技术研究理事会 (CONICET),卢汉国立大学工程学院 (FICH-UNL));University of Cambridge(剑桥大学);Imperial College London(帝国理工学院)
Automated anatomy-based post-processing reduces false positives and improved interpretability of deep learning intracranial aneurysm detection
Brigham and Women’s Hospital(布列根和妇女医院);Harvard Medical School(哈佛医学院);Northeastern University(东北大学);Dana-Farber Cancer Institute(丹娜-法伯癌症研究所);Washing University in St. Louis(圣路易斯华盛顿大学);First Affiliated Hospital of Guangzhou Medical University(广州医科大学附属第一医院)
Deep Learning for Brain Tumor Segmentation in Radiosurgery: Prospective Clinical Evaluation
Skolkovo Institute of Science and Technology(斯科尔科沃科学技术学院);Kharkevich Institute for Information Transmission Problems(哈尔克维奇信息传输问题研究所);Moscow Institute of Physics and Technology(莫斯科物理技术学院);Moscow Gamma-Knife Center(莫斯科伽马刀中心);Burdenko Neurosurgery Institute(布尔坚科神经外科研究所)
Uncertainty-Aware Ordinal Deep Learning for cross-Dataset Diabetic Retinopathy Grading
Department of Computer Science & Engineering, Mississippi State University, MS, USA(计算机科学与工程系,密苏里州立大学);Department of Mechanical Engineering, Mississippi State University, MS, USA(机械工程系,密苏里州立大学);College of Engineering & Architecture, International University of Rabat(工程与建筑学院,拉巴特国际大学)
Brain Lesion Synthesis via Progressive Adversarial Variational Auto-Encoder
King’s College London(伦敦国王学院);National Hospital for Neurology and Neurosurgery(国家神经内科和神经外科医院);Thomas Jefferson University(托马斯·杰斐逊大学);University of Washington(华盛顿大学)
Automated Thalamic Nuclei Segmentation Using Multi-Planar Cascaded Convolutional Neural Networks
一个级联的多平面计划与修改后的剩余U-Net架构被用来分割丘脑核的常规和白质无效(WMn)磁化准备快速梯度回波(MPWE)数据。一个单一的网络进行了优化,从健康对照和多发性硬化症(MS)和原发性震颤(ET),在3 T和7 T场强采集的患者的图像。采用骰子相似系数和体积相似指数(VSI)评价性状。通过应用该方法研究MS对丘脑核萎缩的影响,证明了临床实用性。在不到一分钟的时间里,每个丘脑被分割成12个核团。对于7 T WMn-MPaT,所提出的方法在ET患者上优于当前最先进的方法,其中五个核的Dice(在0.05-0.18的范围内增加)和四个核的VSI(在0.05-0.19的范围内增加)有统计学显著的改善,同时对健康和MS受试者进行了MRI。使用7 T WMn-MPaT数据获得的Dice和VSI与使用3 T WMn-MPaT数据获得的Dice和VSI相当。对于常规MPBAF,所提出的方法显示出在所有细胞核和疾病类型的FreeSurfer上在0.14-0.63的范围内的统计学显著的Dice改善。噪声对网络性能的影响显示出对SNR低至基线SNR一半的图像的鲁棒性。在控制了平行成像、颅内体积、性别和年龄的影响后,与健康对照受试者相比,观察到MS患者的四个丘脑核团和整个丘脑的萎缩(p<0.004)。所提出的分割方法是快速,准确,表现良好的疾病类型和场强,并显示出巨大的潜力,提高我们的理解丘脑核…
Dual Attention Residual U-Net for Accurate Brain Ultrasound Segmentation in IVH Detection
Chongqing Electric Power College(重庆电力学院);Chongqing Metropolitan College of Science and Technology(重庆市级科学研究院);Xinqiao Hospital, Army Medical University(新桥医院,军队医学大学)