Insufficient training data and severe class imbalance are often limiting factors when developing machine learning models for the classification of rare diseases. In this work, we address the problem of classifying bone lesions from X-ray images by increasing the small number of positive samples in the training set. We propose a generative data augmentation approach based on a cycle-consistent generative adversarial network that synthesizes bone lesions on images without pathology. We pose the generative task as an image-patch translation problem that we optimize specifically for distinct bones (humerus, tibia, femur). In experimental results, we confirm that the described method mitigates the class imbalance problem in the binary classification task of bone lesion detection. We show that the augmented training sets enable the training of superior classifiers achieving better performance on a held-out test set. Additionally, we demonstrate the feasibility of transfer learning and apply a generative model that was trained on one body part to another.
LeFusion: Controllable Pathology Synthesis via Lesion-Focused Diffusion Models
University of Science and Technology of China (USTC)(中国科学技术大学);Beihang University(北京航空航天大学);Swiss Federal Institute of Technology Lausanne (EPFL)(洛桑联邦理工学院);Stanford University(斯坦福大学)
Bi-parametric prostate MR image synthesis using pathology and sequence-conditioned stable diffusion
University College London(伦敦大学学院);City University of Hong Kong(香港城市大学);University College Hospital NHS foundation Trust(大学学院医院国民保健服务基金会信托)
我们提出了一种图像合成机制,用于以文本为条件的多序列前列腺MR图像,以控制病变存在和序列,以及生成以图像为条件的配对双参数图像,例如用于从配对数据的T2加权MR生成扩散加权MR,这是病理图像合成中的两个挑战性任务。我们提出的机制利用并建立在最近的稳定扩散模型,提出基于图像的条件配对数据生成。我们验证我们的方法使用2D图像切片从真正的疑似前列腺癌患者。合成图像的真实性通过盲态专家评估进行验证,以识别真实图像与假图像,其中具有4年泌尿科MR阅读经验的放射科医生在所有测试序列中仅达到59.4%的准确度(概率为50%)。对于第一次,我们通过盲专家识别疑似病变的存在来评估所生成的病理的真实性,我们发现临床医生对真实图像和合成图像的表现相似,真实图像和合成图像之间的病变识别准确性差异为2.9个百分点,证明了放射学培训目的的潜力。此外,我们还表明,经过病变识别训练的机器学习模型在使用合成数据增强的真实数据进行训练时,表现出更好的性能(76.2% vs 70.4%,统计学显著改善),而不是仅使用真实图像进行训练,这证明了模型训练的有用性。
Bone Suppression on Chest Radiographs With Adversarial Learning
Imaging Biomarkers and Computer-aided Diagnosis Laboratory, Radiology and Imaging Sciences, National Institutes of Health Clinical Center(美国国立卫生研究院临床中心放射学与成像科学部成像生物标志物和计算机辅助诊断实验室);Ping An Technology Co., Ltd.(平安科技有限公司)
Generative Modelling of 3D in-silico Spongiosa with Controllable Micro-Structural Parameters
Consejo Nacional de Investigaciones Científicas y Técnicas(阿根廷国家科学技术研究委员会);Universidad Tecnológica Nacional, Facultad Regional Buenos Aires(阿根廷国立技术大学布宜诺斯艾利斯地区学院);Universidad Nacional del Sur(阿根廷国立南方大学)
BS-LDM: Effective Bone Suppression in High-Resolution Chest X-Ray Images with Conditional Latent Diffusion Models
Hangzhou Dianzi University(杭州电子科技大学);National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital&Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College(中国医学科学院国家癌症中心/国家临床医学研究中心/癌症医院及深圳医院);Central South University(中南大学);Yunnan University(云南大学);School of Biomedical Engineering, Shenzhen University(深圳大学生物医学工程学院);Shenzhen Research Institute of Big Data(深圳大数据研究院)
Cross-Source Supervision for Bone Infection Segmentation in Dual-Modality PET-CT
Shanghai Maritime University(上海海洋大学);University College London(伦敦大学学院);Shanghai Sixth People’s Hospital(上海第六人民医院);Shanghai Sixth People’s Hospital Affiliated to SJTU School of Medicine(上海第六人民医院附属复旦大学医学院);Queen Mary University of London(伦敦女王玛丽大学)
Automated anomaly-aware 3D segmentation of bones and cartilages in knee MR images from the Osteoarthritis Initiative
The University of Queensland(昆士兰大学);Australian e-Health Research Centre(澳大利亚电子健康研究中心);Commonwealth Scientific and Industrial Research Organisation(联邦科学与工业研究组织)
PathoSyn: Imaging-Pathology MRI Synthesis via Disentangled Deviation Diffusion
Department of Radiology, Boston Children’s Hospital, Harvard Medical School(放射科,波士顿儿童医院,哈佛医学院);College of Science, Northeastern University(科学学院,东北大学);School of Medicine, Yale University(医学院,耶鲁大学);Department of Cardiac Surgery, The Second Affiliated Hospital of Jiangxi Medical College, Nanchang University(心脏外科科,江西医学院第二附属医院,南昌大学);College of Computing and Data Science, Nanyang Technological University(计算与数据科学学院,南洋理工大学)
Ocular Diseases Diagnosis in Fundus Images using a Deep Learning: Approaches, tools and Performance evaluation
ESIEE-Paris, University Paris-Est Marne-la-Vallée(巴黎东马恩拉瓦莱大学ESIEE巴黎);Faculty of medicine, University of Monastir(莫纳斯提尔大学医学院);ISITCom Hammam-Sousse, University of Sousse(苏斯大学哈马姆苏斯高等信息与通信技术学院)
Harmonizing Pathological and Normal Pixels for Pseudo-healthy Synthesis
Xiamen University(厦门大学);Westlake University(西湖大学);Chinese PLA General Hospital(中国人民解放军总医院);Deepwise AI Laboratory(深睿人工智能实验室);The University of Hong Kong(香港大学)
When does Bone Suppression and Lung Field Segmentation Improve Chest X-Ray Disease Classification?
University Medical Center Hamburg-Eppendorf(汉堡-埃彭多夫大学医学中心);Hamburg University of Technology(汉堡工业大学);Forschungszentrum Medizintechnik Hamburg(汉堡医学技术研究中心);Philips Research(飞利浦研究院)
AI-driven software for automated quantification of skeletal metastases and treatment response evaluation using Whole-Body Diffusion-Weighted MRI (WB-DWI) in Advanced Prostate Cancer
The Institute of Cancer Research(癌症研究院);The Royal Marsden NHS Foundation Trust(皇家马斯登NHS基金会信托);IEO European Institute of Oncology IRCCS(欧洲肿瘤研究院);University Hospital Basel(巴塞尔大学医院)
VinDr-SpineXR: A deep learning framework for spinal lesions detection and classification from radiographs
Vingroup Big Data Institute(Vingroup大数据研究所);Hanoi University of Science and Technology(河内理工大学);VinUniversity(Vin大学);Yale University(耶鲁大学)
在临床实践中,X线片被用作识别脊柱异常的最重要的成像工具。然而,脊柱骨病变的评价对放射科医生来说是一项具有挑战性的任务。这项工作旨在开发和评估一个基于深度学习的框架,名为VinDr-SpineXR,用于脊柱X射线异常的分类和定位。首先,我们建立了一个大型数据集,包括来自5,000项研究的10,468张脊柱X射线图像,每个图像都由经验丰富的放射科医生手动注释,并在13个类别的异常发现周围使用边界框。使用这个数据集,我们然后训练一个深度学习分类器来确定脊柱扫描是否异常,并训练一个检测器来定位13个发现中的7个关键发现。VinDr-SpineXR在来自1,000项研究的2,078张图像的测试集上进行评估,该测试集与训练集分开。它证明了图像级分类任务的受试者工作特征曲线下面积(AUROC)为88.61%(95% CI 87.19%,90.02%),病灶级定位任务的平均平均精度(mAP@0.5)为33.56%。这些结果可作为概念验证,并为该方向的未来研究设定了基线。为了鼓励进步,数据集,代码和训练的深度学习模型都是公开的。
On Interpretability of Deep Learning based Skin Lesion Classifiers using Concept Activation Vectors
Hochschule Pforzheim(普福尔茨海姆应用技术大学);Smart Data and Knowledge Services German Research Center for Artificial Intelligence GmbH (DFKI)(德国人工智能研究中心智能数据与知识服务部);Technische Universität Kaiserslautern(凯泽斯劳滕工业大学);Universitätsklinikum Münster(明斯特大学医院);Universitätsklinikum Düsseldorf(杜塞尔多夫大学医院);National University of Science and Technology (NUST)(巴基斯坦国立科技大学);National Center of Artificial Intelligence(国家人工智能中心)
BrainLesion Suite: A Flexible and User-Friendly Framework for Modular Brain Lesion Image Analysis
Department of Quantitative Biomedicine, University of Zurich, Zurich, Switzerland(苏黎世大学定量生物医学系);Helmholtz AI, Helmholtz Zentrum München, Germany(海德堡人工智能,慕尼黑研究中心);Department of Diagnostic and Interventional Neuroradiology, School of Medicine, Klinikum rechts der Isar, Technical University of Munich, Germany(医学诊断与介入神经放射学系,慕尼黑技术大学);TranslaTUM - Central Institute for Translational Cancer Research, Technical University of Munich, Germany(TranslaTUM - 转化癌症研究中央研究所,慕尼黑技术大学);Chair for AI in Medicine, Klinikum Rechts der Isar, Munich, Germany(医学人工智能系,慕尼黑Klinikum Rechts der Isar);Department of Radiation Oncology, Klinikum rechts der Isar, Technical University of Munich, Germany(放射肿瘤学系,慕尼黑Klinikum Rechts der Isar);Institute of Clinical Neuroimmunology, Faculty of Medicine, Ludwig Maximilian University, Munich, Germany(临床神经免疫学研究所,路德维希-马克西米利安大学);School of Computation, Information and Technology, Technical University of Munich, Germany(计算、信息与技术学院,慕尼黑技术大学);Division of Computational Pathology, Indiana School of Medicine, Indianapolis, IN, USA(计算病理学系,印第安纳医学院);Research Center for Digital Technologies in Dentistry and CAD/CAM, Department of Dentistry, Faculty of Medicine and Dentistry, Danube Private University, Krems, Austria(牙科数字技术研究中心,医学与牙科学院,多瑙私人大学);Medical Research Group, MLCommons, San Francisco, CA, USA(医学研究组,MLCommons,旧金山,美国);Department of Neurology, Clinical Neuroscience Center and Brain Tumor Center, University and University Hospital Zurich, 8091 Zurich, Switzerland(神经学系,临床神经科学中心和脑肿瘤中心,苏黎世大学及苏黎世大学医院);Department of Medical Radiation Physics, Stockholm University, Solna, Sweden(医学辐射物理学系,斯德哥尔摩大学);Department of Oncology-Pathology, Karolinska Institutet, Solna, Sweden, Sweden(肿瘤病理学系,卡罗林斯卡研究所);Division of Oncology and Children's Research Center, University Children's Hospital Zurich, Switzerland(肿瘤与儿童研究中心,苏黎世儿童医院);Wallace H. Coulter Department of Biomedical Engineering at Georgia Tech and Emory University, Atlanta, USA(Wallace H. Coulter生物医学工程系,佐治亚理工学院和埃默里大学,美国);Institute of Radiation Medicine (IRM), Helmholtz Zentrum München (HMGU), Germany(放射医学研究所(IRM),海德堡研究中心(HMGU));German Consortium for Translational Cancer Research (DKTK), Partner Site Munich, Munich, Germany(德国转化癌症研究联盟(DKTK),慕尼黑合作伙伴站点)
AI in radiological imaging of soft-tissue and bone tumours: a systematic review evaluating against CLAIM and FUTURE-AI guidelines
Erasmus MC Cancer Institute, University Medical Center Rotterdam(伊拉斯姆斯医学中心癌症研究所,鹿特丹大学医学中心);Leeds Teaching Hospitals NHS Trust(利兹教学医院 NHS 信托基金);Leeds Biomedical Research Centre, University of Leeds(利兹大学生物医学研究中心);Leeds Institute of Medical Research, University of Leeds(利兹大学医学研究所);Amsterdam UMC(阿姆斯特丹大学医学中心);Leiden University Medical Center(莱顿大学医学中心);Faculty of Medical Sciences, University of Groningen(格罗宁根大学医学科学学院)