Implementing neural networks for clinical use in medical applications necessitates the ability for the network to detect when input data differs significantly from the training data, with the aim of preventing unreliable predictions. The community has developed several methods for out-of-distribution (OOD) detection, within which distance-based approaches - such as Mahalanobis distance - have shown potential. This paper challenges the prevailing community understanding that there is an optimal layer, or combination of layers, of a neural network for applying Mahalanobis distance for detection of any OOD pattern. Using synthetic artefacts to emulate OOD patterns, this paper shows the optimum layer to apply Mahalanobis distance changes with the type of OOD pattern, showing there is no one-fits-all solution. This paper also shows that separating this OOD detector into multiple detectors at different depths of the network can enhance the robustness for detecting different OOD patterns. These insights were validated on real-world OOD tasks, training models on CheXpert chest X-rays with no support devices, then using scans with unseen pacemakers (we manually labelled 50% of CheXpert for this research) and unseen sex as OOD cases. The results inform best-practices for the use of Mahalanobis distance for OOD detection. The manually annotated pacemaker labels and the project's code are available at: https://github.com/HarryAnthony/Mahalanobis-OOD-detection.
MedNet: Pre-trained Convolutional Neural Network Model for the Medical Imaging Tasks
Queensland University of Technology(昆士兰科技大学);University of Jaen(哈恩大学);The University of Queensland(昆士兰大学);University of Sumer(苏美尔大学);University of Baghdad(巴格达大学);The University of Information Technology & Communications(信息与通信技术大学);University of Missouri(密苏里大学)
Benchmarking Dependence Measures to Prevent Shortcut Learning in Medical Imaging
Hertie Institute for AI in Brain Health, University of Tübingen(蒂宾根大学赫蒂脑健康人工智能研究所);University Hospital of Tübingen(蒂宾根大学医院);University of Stuttgart(斯图加特大学);University of Bern(伯尔尼大学);Inselspital, Bern University Hospital(伯尔尼大学医院 Inselspital);Stanford University(斯坦福大学)
Beyond Knowledge Silos: Task Fingerprinting for Democratization of Medical Imaging AI
National Center for Tumor Diseases (NCT)(肿瘤疾病国家中心);German Cancer Research Center (DKFZ) Heidelberg(德国癌症研究中心(Heidelberg));Faculty of Mathematics and Computer Science(数学和计算机科学学院);Heidelberg University(海德堡大学);HIDSS4Health - Helmholtz Information and Data Science School for Health(HIDSS4Health - 健康信息与数据科学学校);Medical Faculty(医学学院);Pattern Analysis and Learning Group(模式分析与学习小组)
CheXstray: Real-time Multi-Modal Data Concordance for Drift Detection in Medical Imaging AI
Microsoft Health and Life Sciences(微软健康与生命科学部);Stanford Center for Artificial Intelligence in Medicine and Imaging(斯坦福医学与成像人工智能中心)
临床人工智能(AI)应用正在全球范围内迅速扩展,并有可能影响到医疗实践的所有领域。医学成像应用占批准的临床AI应用的绝大多数。尽管医疗系统渴望采用人工智能解决方案,但仍然存在一个基本问题:\textit{what happens after the AI model goes into production?}我们使用CheXpert和PadChest公共数据集构建和测试医疗成像人工智能漂移监测工作流程,以跟踪数据和模型漂移,而无需同时获得地面实况。我们在多个实验中模拟漂移,将模型性能与我们新的多模态漂移度量进行比较,该度量使用DICOM元数据,变分自动编码器(VAE)的图像外观表示和模型输出概率作为输入。通过实验,我们展示了一个强大的代理地面真相的性能,使用相关元数据,预测概率和VAE潜在表示的无监督分布变化。我们的主要贡献包括(1)医学成像漂移检测的概念验证,包括使用VAE和特定领域的统计方法,(2)测量和统一漂移指标的多模式方法,(3)对观察部署的医学成像AI的挑战和解决方案的新见解,以及(4)创建开源工具,使其他人能够轻松运行自己的工作流程和场景。这项工作具有重要意义。它解决了在动态医疗环境中常见的连续医学成像AI模型监测中发现的相关翻译差距。
Deep reinforcement learning in medical imaging: A literature review
Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所);University of Arkansas(阿肯色大学);University of Houston(休斯顿大学);INRIA, Sophia Antipolis-Mediterranean Centre(法国国家信息与自动化研究所索菲亚·安蒂波利斯地中海中心)
Learning Personalized Representation for Inverse Problems in Medical Imaging Using Deep Neural Network
Gordon Center for Medical Imaging(戈登医学影像中心);Massachusetts General Hospital(麻省总医院);Martinos Center for Biomedical Imaging(马蒂诺斯生物医学影像中心);University of California, Davis(加州大学戴维斯分校)
Multi-environment lifelong deep reinforcement learning for medical imaging
Rice University(莱斯大学);The Johns Hopkins University(约翰斯·霍普金斯大学);McGovern Medical School, UTHealth Houston(休斯顿UTHealth麦戈文医学院);Johns Hopkins University School of Medicine(约翰斯·霍普金斯大学医学院);University of Maryland Medical Intelligent Imaging (UM2ii)(马里兰大学医学智能成像中心);University of Maryland School of Medicine(马里兰大学医学院)
Parallel Medical Imaging for Intelligent Medical Image Analysis: Concepts, Methods, and Applications
Sun Yat-sen University(中山大学);Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);University of Chinese Academy of Sciences(中国科学院大学);Xi’an Jiaotong University(西安交通大学);Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College(中国医学科学院北京协和医学院北京协和医院);University of Portsmouth(朴茨茅斯大学);Rensselaer Polytechnic Institute(伦斯勒理工学院);Qingdao Academy of Intelligent Industries(青岛智能产业技术研究院)
Towards Universal Unsupervised Anomaly Detection in Medical Imaging
Technical University of Munich(慕尼黑工业大学);Helmholtz AI and Helmholtz Center Munich(亥姆霍兹人工智能与亥姆霍兹慕尼黑中心);Klinikum Rechts der Isar(伊萨尔河右岸医院);Imperial College London(伦敦帝国理工学院)
Towards Data-Efficient Medical Imaging: A Generative and Semi-Supervised Framework
Dept. of Electrical and Electronic Engineering Imperial College London(电气与电子工程系帝国理工学院伦敦);Centre for Vision, Speech and Signal Processing University of Surrey(视觉、语音和信号处理中心萨里大学)
Transforming medical imaging with Transformers? A comparative review of key properties, current progresses, and future perspectives
Chinese Academy of Sciences(中国科学院);Johns Hopkins Medical Institutes(约翰斯·霍普金斯医学院);Vanderbilt University(范德堡大学);University of Science and Technology of China(中国科学技术大学)
Recent Advances in Medical Imaging Segmentation: A Survey
Junia, UMR 8520, CNRS, Centrale Lille, Univerity of Polytechnique Hauts-de-France(Junia研究院、法国国家科学研究中心、Centrale Lille、法国高等技术大学 Hauts-de-France);Sorbonne Center for Artificial Intelligence (SCAI), Abu Dhabi, UAE(索邦人工智能中心(SCAI))