Large medical imaging data sets are becoming increasingly available, but ensuring sample quality without significant artefacts is challenging. Existing methods for identifying imperfections in medical imaging rely on data-intensive approaches, compounded by a scarcity of artefact-rich scans for training machine learning models in clinical research. To tackle this problem, we propose a framework with four main components: 1) artefact generators inspired by magnetic resonance physics to corrupt brain MRI scans and augment a training dataset, 2) abstract and engineered features to represent images compactly, 3) a feature selection process depending on the artefact class to improve classification, and 4) SVM classifiers to identify artefacts. Our contributions are threefold: first, physics-based artefact generators produce synthetic brain MRI scans with controlled artefacts for data augmentation. This will avoid the labour-intensive collection and labelling process of scans with rare artefacts. Second, we propose a pool of abstract and engineered image features to identify 9 different artefacts for structural MRI. Finally, we use an artefact-based feature selection block that, for each class of artefacts, finds the set of features providing the best classification performance. We performed validation experiments on a large data set of scans with artificially-generated artefacts, and in a multiple sclerosis clinical trial where real artefacts were identified by experts, showing that the proposed pipeline outperforms traditional methods. In particular, our data augmentation increases performance by up to 12.5 percentage points on accuracy, precision, and recall. The computational efficiency of our pipeline enables potential real-time deployment, promising high-throughput clinical applications through automated image-processing pipelines driven by quality control systems.
Automated MRI Quality Assessment of Brain T1-weighted MRI in Clinical Data Warehouses: A Transfer Learning Approach Relying on Artefact Simulation
Sorbonne Université(索邦大学);Institut du Cerveau – Paris Brain Institute - ICM(大脑研究所——巴黎脑研究所 - ICM);CNRS(法国国家科学研究中心);Inria(法国国家信息与自动化研究所);Inserm(法国国家健康与医学研究院);AP-HP(巴黎公共医疗集团);Hôpital de la Pitié-Salpêtrière(皮蒂-萨尔皮特里医院)
Motion-related Artefact Classification Using Patch-based Ensemble and Transfer Learning in Cardiac MRI
University of Nottingham(诺丁汉大学)
心脏磁共振成像(MRI)在心功能分析中起着重要作用。然而,由于屏气困难,采集通常伴随有运动伪影,特别是对于急性症状患者。因此,有必要对心脏MRI的质量进行评估,以便进一步分析。耗时的人工分类不利于构建端到端的计算机辅助诊断系统。为了克服这个问题,在这项工作中提出了一个自动心脏MRI质量估计框架,使用集成和迁移学习。多个预先训练的模型被初始化,并在从训练数据中采样的二维图像块上进行微调。在模型推理过程中,来自这些模型的决策被汇总以做出最终预测。该框架已在CMRxMotion Grand Challenge(MICCAI 2022)数据集上进行了评估,该数据集较小,多类且不平衡。它在训练集(5倍交叉验证)和验证集上分别实现了78.8%和70.0%的分类准确率。CMRxMotion组织者还在独立测试集上对最终训练模型进行了评估,其分类准确率为72.5%,Cohen's Kappa为0.6309(在这项重大挑战中排名第一)。我们的代码可以在Github上找到:https://github.com/ruizhe-l/CMRxMotion。
Simultaneous Image Quality Improvement and Artefacts Correction in Accelerated MRI
Translational Radiomics, Luxembourg Institute of Health, Luxembourg(转化放射组学,卢森堡健康研究所,卢森堡);In-Vivo Imaging Platform, Luxembourg Institute of Health, Luxembourg(活体成像平台,卢森堡健康研究所,卢森堡);Faculty of Electrical Engineering and Communication, Brno University of Technology, Czech Republic(电气工程与通信学院,布拉格技术大学,捷克共和国);Institute of Scientific Instruments of the Czech Academy of Sciences, Czech Republic(捷克科学院科学仪器研究所,捷克共和国)
IMPORTANT-Net: Integrated MRI Multi-Parameter Reinforcement Fusion Generator with Attention Network for Synthesizing Absent Data
The Netherlands Cancer Institute(荷兰癌症研究所)
磁共振成像(MRI)对于乳房中的病变检测是高度敏感的。通过不同的设置获得的序列可以捕获病变的特定特征。这种多参数MRI信息已经被证明可以提高放射科医师在病变分类方面的表现,以及提高人工智能模型在各种任务中的表现。然而,获得多参数MRI使得检查在财务和时间方面都是昂贵的,并且对于特殊人群可能存在安全性问题,从而使得MRI序列的全谱采集不太持久。在这项研究中,不同于原始输入融合或现有MRI参数的特征拼接,开发了一种新的$\textbf{I}$集成MRI $\textbf{M}$多功能增强$\textbf{O}$规则融合生成器$\textbf{R}$ wi $\textbf{T}$ h $\textbf{A}$ tte $\textbf{NT}$ ion Network(IMPORTANT-Net)来生成缺失参数。首先,参数重建模块用于编码和恢复现有的MRI参数,以获得相应的潜在表示信息在任何尺度水平。多参数融合注意模块通过一组算法策略使不同参数的编码信息相互作用,并通过信息融合后的注意机制对信息赋予不同的权重,得到精细化的表征信息。最后,使用嵌入在$V^{-}$形状生成模块中的强化融合方案来组合分层表示以生成缺失的MRI参数。结果表明,我们的IMPORTANT-Net能够生成缺失的MRI参数,并优于同类最先进的网络。我们的代码可在https://github.com/Netherlan…
RSFR: A Coarse-to-Fine Reconstruction Framework for Diffusion Tensor Cardiac MRI with Semantic-Aware Refinement
Imperial College London(帝国理工学院伦敦分校);National Heart and Lung Institute(国家心肺研究所);Cardiovascular Research Centre(心血管研究中心);Royal Brompton Hospital(皇家布里蒙特医院);School of Biomedical Engineering, Shenzhen Campus of Sun Yat-sen University(中山大学深圳校区生物医学工程学院);Robotics and Autonomous Systems Thrust & Data Science and Analytics Thrust, HKUST (GZ)(香港科技大学(广州)机器人与自主系统 thrust 及数据科学与分析 thrust);Yau Mathematical Sciences Centre, Tsinghua University(清华大学姚数学科学中心);Department of Applied Mathematics and Theoretical Physics, University of Cambridge(剑桥大学应用数学与理论物理系);School of Biomedical Engineering and Imaging Sciences, King’s College London(伦敦国王学院生物医学工程与影像科学学院)
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)
JoJoNet: Joint-contrast and Joint-sampling-and-reconstruction Network for Multi-contrast MRI
University of Georgia(佐治亚大学);United Imaging Intelligence(联影智能);ShanghaiTech University(上海科技大学);Shanghai United Imaging Intelligence Co., Ltd.(上海联影智能有限公司)
Imperial College London(帝国理工学院);Royal Brompton Hospital(皇家布朗普顿医院);Sun Yat-sen University(中山大学);Beihang University(北京航空航天大学);University of the Basque Country UPV/EHU(巴斯克大学(UPV/EHU));TECNALIA(泰克纳利亚研究中心);Shenzhen Second People’s Hospital(深圳市第二人民医院);The First Affiliated Hospital of Shenzhen University Health Science Center(深圳大学医学部第一附属医院)
Flexible Amortized Variational Inference in qBOLD MRI
University of Sussex(萨塞克斯大学);Brighton and Sussex Medical School(布莱顿和萨塞克斯医学院);St. Vincent’s University Hospital(圣文森特大学医院);University of Nottingham(诺丁汉大学);Cardiff University(卡迪夫大学)
Fighting the scanner effect in brain MRI segmentation with a progressive level-of-detail network trained on multi-site data
Center for Cognitive Neuroimaging(认知神经成像中心);School of Psychology & Neuroscience(心理与神经科学学院);University of Glasgow(格拉斯哥大学);University of Brescia(布雷西亚大学);Department of Information Engineering(信息工程系)
Robust joint registration of multiple stains and MRI for multimodal 3D histology reconstruction: Application to the Allen human brain atlas
University College London(伦敦大学学院);Inria(法国国家信息与自动化研究所);Massachusetts General Hospital(麻省总医院);Harvard Medical School(哈佛医学院);Massachusetts Institute of Technology(麻省理工学院)
SRUG: A Fusion-Driven Generator Network for Medical Image Translation
School of Electronic and Electrical Engineering, Shanghai University of Engineering Science(上海工程技术大学电子电气工程学院);Department of Thoracic Surgery, Zhongshan Hospital of Fudan University(复旦大学中山医院胸外科);Clinical Research Unit, Institute of Clinical Science, Zhongshan Hospital of Fudan University(复旦大学中山医院临床科研部);Department of Medical Oncology, Cancer Center, and Fudan Zhangjiang Institute, Zhongshan Hospital of Fudan University(复旦大学中山医院医学肿瘤科、癌症中心及复旦张江研究院)
Predictive Enhancement Calibration for Latent Breast MRI Virtual Contrast Enhancement
The First Affiliated Hospital of Chongqing Medical University(重庆医科大学附属第一医院);Chongqing Municipal Health Commission(重庆市卫生健康委员会);Chongqing Translational Medicine Center(重庆市转化医学中心)
Spatio-Temporal Deep Learning-Based Undersampling Artefact Reduction for 2D Radial Cine MRI with Limited Data
Charité - Universitätsmedizin Berlin(柏林夏里特医学院);Berlin Institute of Health(柏林健康研究所);Physikalisch-Technische Bundesanstalt (PTB)(德国联邦物理技术研究院(PTB));King’s College London(伦敦国王学院)
Response monitoring of breast cancer on DCE-MRI using convolutional neural network-generated seed points and constrained volume growing
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, Antoni van Leeuwenhoek Hospital - The Netherlands Cancer Institute(荷兰癌症研究所安东尼·范·列文虎克医院放射科)
Predicting dynamic, motion-related changes in B0 field in the brain at a 7 T MRI using a subject-specific fine-tuned U-net
Medical University of Vienna(维也纳医科大学);University of Freiburg(弗莱堡大学);University of Freiburg – Medical Centre(弗莱堡大学医学中心);Karl Landsteiner Institute for Clinical Molecular MR in Musculoskeletal Imaging(卡尔·兰德施泰纳肌肉骨骼成像临床分子磁共振研究所);Carinthia University of Applied Sciences(克恩顿应用科学大学)
Task-Adaptive 3D Cross-Field MRI Translation via Field-Conditioned Content-Style Pretraining
School of Integrated Circuits and Electronics, Beijing Institute of Technology(北京理工大学集成电路与电子学院);School of Biomedical Engineering, Tsinghua Medicine, Tsinghua University(清华大学医学院生物医学工程学院);Department of Electronic Engineering, The Chinese University of Hong Kong(香港中文大学电子工程系)
Med-NCA: Robust and Lightweight Segmentation with Neural Cellular Automata
Darmstadt University of Technology(达姆施塔特工业大学)
在使用深度学习执行医学图像分割时,访问适当的基础设施至关重要。这一要求使得在资源有限的情况下,如农村地区和危机期间的初级保健设施,很难运行最先进的细分模型。最近新兴的神经元胞自动机(NCA)领域已经表明,局部相互作用的单细胞模型可以在低分辨率输入的图像生成或分割等任务中取得有竞争力的结果。然而,它们受到高VRAM要求和难以达到高分辨率图像的收敛的限制。为了克服这些限制,我们提出了Med-NCA,一个用于高分辨率图像分割的端到端NCA训练管道。我们的方法遵循两个步骤的过程。全局知识首先在缩小的图像中的细胞之间进行通信。之后,执行基于块的分割。我们提出的Med-NCA在海马和前列腺分割方面分别比经典的UNet好2%和3%,同时也小了500倍。我们还表明,Med-NCA是通过设计不变的图像尺度,形状和平移,经历只有轻微的性能下降,即使有很强的转变,是强大的MRI采集伪影。Med-NCA甚至可以在Raspberry Pi B+上实现高分辨率医学图像分割,Raspberry Pi B+可以说是能够运行PyTorch的最小设备,并且可以由标准电源供电。
Towards contrast- and pathology-agnostic clinical fetal brain MRI segmentation using SynthSeg
University Children’s Hospital Zurich(苏黎世大学儿童医院);ETH Zurich(苏黎世联邦理工学院);University of Zurich(苏黎世大学);Medical University of Vienna(维也纳医科大学);CIBM Center for Biomedical Imaging(CIBM生物医学成像中心);University of Lausanne(洛桑大学);Lausanne University Hospital(洛桑大学医院);King's College London(伦敦国王学院);University Hospital Erlangen(埃尔朗根大学医院);Spanish National Research Council (CSIC)(西班牙国家研究委员会);Universitat Politècnica de València(瓦伦西亚理工大学)
Data and Physics Driven Learning Models for Fast MRI -- Fundamentals and Methodologies from CNN, GAN to Attention and Transformers
National Heart and Lung Institute, Imperial College London(帝国理工学院国家心肺研究所);Imperial College London(帝国理工学院);Sun Yat-sen University(中山大学);Beihang University(北京航空航天大学);Brunel University London(伦敦布鲁内尔大学);University of Cambridge(剑桥大学);Technical University of Munich(慕尼黑工业大学);Weizmann Institute of Science(魏茨曼科学研究所)
Rensselaer Polytechnic Institute(伦斯勒理工学院);Southeast University(东南大学);Fudan University(复旦大学);University of Massachusetts Lowell(马萨诸塞大学洛厄尔分校);The University of Hong Kong(香港大学)