diff --git a/.DS_Store b/.DS_Store new file mode 100644 index 0000000..0ce8880 Binary files /dev/null and b/.DS_Store differ diff --git a/404.html b/404.html new file mode 100644 index 0000000..e233e1b --- /dev/null +++ b/404.html @@ -0,0 +1,1016 @@ + + + +
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +Perhaps you were looking for one of these?
+ + + + +Anna Reithmeir is a PhD student at the Chair of Computational Imaging and AI in Medicine at TU Munich. She received her B.Sc. and M.Sc. in Informatics from TU Munich with a focus on computer vision and high performance computing. In her Master’s thesis at the Munich Institute for Robotics and Machine Intelligence (MIRMI), she developed a novel algorithm for human-robot manipulability domain adaptation. Her research interests lie in the analysis and development of robust and data-driven models for image registration, numerics of machine learning algorithms and Riemannian manifolds.
+ +M.Sc. in Informatics, 2022
+TU Munich
+B.Sc. in Informatics, 2019
+TU Munich
+Exploring SPD Feature Descriptors for Medical Image Classification, Master's Thesis
+open, 1.1.2024 | tba
+Anneliese Riess is a PhD student at the Institute of Machine Learning for Biomedical Imaging (IML) at Helmholtz Center Munich and Technical University Munich (TUM). She received her B.Sc. and M.Sc. in Mathematics at TUM and devoted a substantial part of her studies to the field of probability theory. In her Master’s thesis she investigated Majority Voting Processes, a class of interacting particle systems. The main focus of the thesis was the equilibrium behaviour of such stochastic models. Prior to her PhD, she worked on two different projects at the university in her final year of her Master’s degree. In the first project, she worked on creating a model that describes the behaviour of DNA methylation. The second project involved modelling and analysing the propagation of underground water. Her research interests lie in the mathematical foundations of privacy-preserving artificial intelligence.
+ +M.Sc. in Mathematics, 2023
+TU Munich
+B.Sc. in Mathematics, 2019
+TU Munich
+Cosmin Bercea is a Ph.D. Student at the Institute of Machine Learning for Biomedical Imaging (IML) at Helmholtz Center Munich and Technical University of Munich (TUM). He received his B.Sc. and M.Sc. in Computer Science from the FAU University in Erlangen, Germany, focusing on pattern recognition and medical image analysis. In his Master’s thesis at Siemens Healthineers in Erlangen, he built novel shared memory neural networks for medical imaging. Before his Ph.D., he worked as a research engineer at Bosch Corporate Research, developing deep learning algorithms for scene understanding for self-driving cars. His research interests lie in interpretable machine learning algorithms for anomaly detection.
+ +M.Sc. in Computer Science, 2018
+FAU Erlangen
+B.Sc. in Computer Science, 2015
+FAU Erlangen
+Diffusion Models for Counterfactual Pathology Synthesis, Master's Thesis
+running, 1.11.2023 | Malek Ben Alaya
+Diffusion Models for Fetal US Anomaly Detection, GRP
+running, 1.11.2023 | Hanna Mykula
+Unsupervised Representation Learning for Alzheimer’s Disease Quantification, GRP
+running, 1.10.2023 | Mehmet Yigit Avci
+Unsupervised Representation Learning for Alzheimer’s Disease Detection, GRP
+finished, 31.10.2023 | Mehmet Yigit Avci
+Diffusion Models for Unsupervised Anomaly Detection, GRP
+finished, 24.10.2023 | Michael Neumayr
+Unsupervised Anomaly Detection in Fetal Brain Ultrasound, Master's Thesis
+finished, 15.08.2023 | Ruxandra Petrescu
+Daniel Lang will be a postdoc at the Institute of Machine Learning in Biomedical Imaging at Helmholtz Munich. +His research interest focuses on the application of deep learning models for problem settings in the field of +medical imaging with a special focus on cancer management. +He is particularly interested in topics like transfer and selfsupervised learning, out of distribution problems and +domain adaptation, and survival analysis.
+ +PhD in Physics, 2022
+Helmholtz Munich and Technical University Munich
+MSc in Physics, 2018
+University Regensburg
+BSc in Physics, 2016
+University Regensburg
+Emily Chan is a postdoctoral researcher at the Institute of Machine Learning for Biomedical Imaging at Helmholtz Munich. She received her PhD in 2022 from King’s College London, where she worked on utilising classical machine learning and deep learning techniques with limited and imbalanced data for MR liver imaging, in collaboration with Perspectum. She is particularly interested in the automation of various clinically-relevant tasks in radiology, with her research at the IML focusing on deep learning for the early diagnosis and prognosis of Alzheimer’s disease.
+ +PhD in Biomedical Engineering, 2022
+King's College London
+MEng in Biomedical Engineering, 2016
+Imperial College London
+Georgios Kaissis is a principal investigator at the Institute of Biomedical Machine Learning (IML) at the Helmholtz Center Munich, a senior research scientist at the Institute of Artificial Intelligence and Informatics in Medicine and specialist diagnostic radiologist at the Institute for Radiology at TUM, a postdoctoral researcher at the Department of Computing at Imperial College London and leads the Healthcare Unit at OpenMined. His research concentrates on biomedical image analysis with a focus on next-generation privacy-preserving machine learning methods as well as probabilistic methods for the design and deployment of robust, secure, fair and transparent machine learning algorithms to medical imaging workflows.
+ +PostDoc in AI for medical imaging
+Imperial College London, UK
+Specialist Radiologist
+Technical University of Munich, Germany
+Doctorate in molecular medicine and systems biology (Dr. med.)
+LMU Munich, Germany
+Medical Degree
+LMU Munich, Germany
+Master of Healthcare Business Administration
+FAU Erlangen-Nuremberg, Germany
+Hannah Eichhorn is a PhD student at the Institute of Machine Learning in Biomedical Imaging (IML), Helmholtz Munich. She received her B.Sc. in Physics from Heidelberg University and her M.Sc. in Bio- and Medical Physics from University of Copenhagen. In her Master’s thesis at the Neurobiology Research Unit, Copenhagen University Hospital, she worked on prospective motion correction for brain magnetic resonance imaging (MRI). Her doctoral research focuses on deep-learning based reconstruction and motion correction of multi-parametric brain MRI, in collaboration with the Neuroscientific MR-Physics research group at Klinikum rechts der Isar (TUM).
+ +MSc in Bio- and Medical Physics, 2021
+Niels Bohr Institute, University of Copenhagen, Copenhagen, Denmark
+BSc in Physics, 2018
+Heidelberg University, Heidelberg, Germany
+Johannes Kiechle is a Ph.D. Student at the Technical University of Munich. He received his B.Eng. from Munich University of Applied Sciences and M.Sc. from Technical University of Munich. In his Master’s thesis, he investigated the shape change of the human hippocampus in the course of ageing within a population of healthy individuals using graph neural networks. For his PhD project, he works in collaboration with the department of Radiation Oncology at the University Hospital rechts der Isar. Therein the focus is on the development and validation of histology-specific AI-based decision support systems for soft-tissue-sarcoma patients.
+ +M.Sc. in Electrical and Computer Engineering, 2023
+Technical University of Munich
+B.Eng. in Electrical Engineering and Information Technology, 2020
+Munich University of Applied Sciences
+Julia A. Schnabel is Professor of Computational Imaging and AI in Medicine at Technical University of Munich (TUM Liesel Beckmann Distinguished Professorship) and Director of a new Institute of Machine Learning in Biomedical Imaging at Helmholtz Center Munich (Helmholtz Distinguished Professorship), with secondary appointment as Chair in Computational Imaging at King’s College London. She graduated in Computer Science (equiv. MSc) from Technical University of Berlin, Berlin, Germany, and was awarded the PhD in Computer Science from University College London, UK. In 2007, she joined the University of Oxford, UK as Associate Professor in Engineering Science (Medical Imaging), where she became Full Professor of Engineering Science by Recognition of Distinction in 2014. She joined King’s College London as a new Chair in 2015, and in 2021 joined TUM and Helmholtz Munich for her current positions. Her research interests include machine/deep learning, nonlinear motion modeling, as well as multimodality and quantitative imaging, for cancer imaging, cardiac imaging, neuroimaging and perinatal imaging. Dr. Schnabel has been elected Fellow of IEEE (2021), Fellow of ELLIS (2019), and Fellow of the MICCAI Society (2018). She is an Associate Editor of the IEEE Transactions on Medical Imaging on whose steering board she serves since 2021, the IEEE Transactions of Biomedical Engineering, on the Editorial Board of Medical Image Analysis and Executive/Founding Editor of MELBA. She currently serves as elected Technical Representative on IEEE EMBS AdCom, as voting member of the IEEE EMBS Technical Committee on Biomedical Imaging and Image Processing (BIIP), as Executive Secretary to the MICCAI board, and as member of ELLIS Health and ELLIS Munich.
+ +PhD in Computer Science, 1998
+University College London
+Diplom (Msc. eq) in Computer Science, 1993
+Technical University of Berlin
+Jun Li is a Ph.D. Student at the Chair of Computational Imaging and AI in Medicine at TU Munich. She received her M.E. in Computer Technology from the University of Chinese Academy of Sciences, China. In her Master’s thesis, she developed a novel framework that combines supervised and unsupervised learning for ultrasound report generation. Her research interests lie in Vision and Language, Multi-Modal Learning, and Cross-Modality Generation.
+ +M.E. in Computer Technology, 2023
+University of Chinese Academy of Sciences, Shenzhen Institute of Advanced Technology
+B.E. in Traffic and Transportation, 2020
+Shenzhen University
+Maxime Di Folco is a PostDoctoral researcher at the Institute of Machine Learning for Biomedical Imaging at Helmholtz Center Munich. His research interest is the study of the cardiac function via machine learning methods, in particular representation learning methods, that aim to acquire low dimensional representation of high dimensional data, with a strong focus on cardiac remodelling (adaptation of the heart to its environment or a disease), notably the study of the deformation and shape aspects.
+ +PhD in Artificial Intelligence, 2021
+Université de Lyon, CREATIS Laboratory
+MEng in Image Processing, 2018
+CPE Lyon
+MSc in Image development and 3D technologies, 2018
+Université Lyon 1
+Sabine Franke supports the Lab for Computational Imaging and AI in Medicine as a member of the administrative staff at the TU campus in Garching. She graduated in 2014 from the University of Graz, Austria, with a degree in conference interpreting for German, English and Spanish. Before joining the team at the TU Munich, she spent several years working as a translator and interpreter in Germany as well as in the Netherlands, adding Dutch to her working languages.
+ +Conference Interpreter (MA), 2014
+University of Graz, Austria
+Sameer Ambekar is a Ph.D. Student at the Technical University of Munich (TUM). He received his Masters in Artificial Intelligence (MSc AI) from the University of Amsterdam (UvA), Netherlands. For his Master’s thesis (48 ECTS), he worked on ‘Test-Time Adaptation for Domain Generalization by generating models and labels through Variational meta-learning’ at the AIM Lab, UvA. Prior to his master’s, he worked as a Research Assistant (RA) at IIT Delhi (IITD) to address Unsupervised Domain Adaptation through methods such as Variational generative latent search. He is interested in solving problems in unsupervised learning through methods such as meta-learning and variational inference alongside learning efficient and transferable features.
+ +M.Sc. in Artificial Intelligence (MSc AI), 2023
+University of Amsterdam, Netherlands
+B.E. in Computer Science, 2018
+VTU, India
+tba. Stay tuned on website - https://ambekarsameer.com
+, |
+Sandra Mayer supports the Lab for Computational Imaging and AI in Medicine as a member of the administrative staff at the Helmholtz Campus in Neuherberg.
+ +Simona Bottani is a PostDoctoral Fellow at the IML where she works on deep learning applied to big research medical imaging cohort. She received her PhD in computer science from Sorbonne University in April 2022. She worked at the ARAMIS Lab and her thesis focused on the application of deep learning models for neuroimaging studies using a large scale clinical data warehouse of the Paris Great Area Hospitals (AP-HP). From 2017 to 2018 she worked as research engineer in the ARAMIS Lab. She received a Master Degree in 2016 and a Bachelor degree in 2014 in Biomedical engineering from Politecnico di Torino.
+ +PhD in Computer Science, 2021
+Sorbonne Université
+MSc in Biomedical Engineer, 2016
+Politecnico di Torino
+BSc in Biomedical Engineer, 2014
+Politecnico di Torino
+Stefan Fischer is a Ph.D. Student at the Technical University of Munich (TUM). He received his B.Sc. and M.Sc. from FAU in Erlangen, Germany with a focus on medical image analysis. In his Master’s thesis at the Radiooncology department of the university hospital Erlangen, he build a generative approach for brain metastasis for data augmentation in MR Imaging. His research interest lies in deep learning based segmentation, transfer learning and curriculum learning.
+ +M.Sc. in Computer Science, 2022
+FAU Erlangen
+B.Sc. in Medical Engineering, 2019
+FAU Erlangen
+Veronika Spieker is a PhD student at the Institute of Machine Learning for Biomedical Imaging (IML) at Helmholtz Munich and Technical University of Munich (TUM). After completing her B.Sc. in engineering at TU Darmstadt and Virginia Tech, she pursued her interest in the medical domain with a M.Sc. in Medical Technologies at TUM. For her PhD project, she works on Physics-Based AI for Motion Correction in Abdominal MRI in collaboration with the Body Magnetic Resonance Group at the Klinikum rechts der Isar. Her research interests include concepts such as neural implicit representations and it’s application to MR reconstruction and motion estimation.
+ +MSc Medical Technologies and Asstistant Systems, 2021
+Technical University of Munich
+MSc Mechanical Engineering, 2021
+Technical University of Munich
+BSc Mechanical Engineering, 2017
+Technical University of Darmstadt / Virginia Tech
+Reducing Labeling Efforts in Segmentation-based Registration in Medical Imaging, Master's thesis
+finished, 1.11.2023 | Varsha Raveendran
+Deep Learning For Medical Image Registration, GRP
+finished, 1.11.2023 | Varsha Raveendran
+Veronika A. Zimmer is a principal investigator at the Institute of Computer Sciences at TUM and a visiting researcher at the School of Biomedical Engineering & Imaging Sciences at King’s College London. She received her PhD in Information and Communication Technologies from the Universitat Pompeu Fabra, Barcelona, Spain, in 2017. Her research focuses on image analysis and machine learning with a particular interest in robust and generalizable methods for multimodal registration and segmentation in medical imaging.
+ +Medical Image Computing (Ph. D.), 2017
+Universitat Pompeu Fabra, Barcelona, Spain
+Computational Life Science (M. Sc.), 2011
+University of Luebeck, Germany
+Computational Life Science (B. Sc.), 2008
+University of Luebeck, Germany
+Faculty of Informatics and Institute for Advanced Study
+Institute of Machine Learning in Biomedical Imaging
+Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis posuere tellusac convallis placerat. Proin tincidunt magna sed ex sollicitudin condimentum. Sed ac faucibus dolor, scelerisque sollicitudin nisi. Cras purus urna, suscipit quis sapien eu, pulvinar tempor diam.
+ + +Slides can be added in a few ways:
+slides
parameter in the front matter of the talk filestatic/
and link using url_slides
parameter in the front matter of the talk fileFurther event details, including page elements such as image galleries, can be added to the body of this page.
+ +The Institute for Computational Imaging and AI in Medicine (CompAI) at TUM and the Institute of Machine Learning in Biomedical Imaging (IML) at Helmholtz Center Munich focus on research to leverage machine learning for the grand challenges in biomedical imaging in areas of unmet clinical need. Novel and affordable solutions should empower clinics to make more accurate, fast and reliable decisions for early detection, treatment planning and improved patient outcome. We are looking for team members, please contact us.
+Deep Learning for Retrospective Motion Correction in MRI: A Comprehensive Review by Veronika Spieker and Hannah Eichhorn et al. has been accepted for publication at IEEE Transactions on Medical Imaging.
+Recent
+Featured
+Winter semester 2023. TUM Informatics. Master Seminar.
+Winter semester 2023. TUM Informatics. Master Seminar.
+Summer semester 2022. TUM Informatics. Master Seminar. Details
+Winter 2021. TUM Informatics. Lecture. Details.
+Summer 2022. TUM Informatics. Lecture. Details.
+If you are interested in one of these projects please contact us and attach a motivation letter, transcript of academic records and CV.
+Master Thesis. I’m interested
+Master Thesis. I’m interested
+PhD position starting July 2022. Position related to MCML Munich. I’m interested
+Institute of Machine Learning in Biomedical Imaging + + + + + + + + + + + + + + + +
+Faculty of Informatics and Institute for Advanced Study + + + + + + + + + + + + + + + +
+We are delighted to announce that our research on developing automatic diffusion models for anomaly detection has been accepted and will be published in the proceedings of the 3rd workshop for Interpretable Machine Learning in Healthcare, held at the International Conference on Machine Learning 2023. Congratulations to our dedicated student Michael for his outstanding contribution to this achievement!
+Curious about how to solve the noise paradox illustrated below? Check out our project page.
++ + + + + + + + + + + + + + + + +
+“What Do AEs Learn? Challenging Common Assumptions in Unsupervised Anomaly Detection and Reversing the Abnormal: Pseudo-Healthy Generative Networks for Anomaly Detection by Cosmin I. Bercea et al. have been accepted for publication at the 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023, which will be held from October 8th to 12th 2023 in Vancouver, Canada.
++ + + + + + + + + + + + + + + + +
++ + + + + + + + + + + + + + + + +
+“Generalizing Unsupervised Anomaly Detection: Towards Unbiased Pathology Screening” by Cosmin I. Bercea et al. has been accepted for publication at Medical Imaging with Deep Learning, Nashville, 2023. Cosmin Bercea will present his work on Monday, 10 July 2023 at 9:15 pm CEST.
++ + + + + + + + + + + + + + + + +
+Moving beyond hyperintensity thresholding: This paper analyzes the challenges and outlines opportunities for advancing the field of unsupervised anomaly detection. Our proposed method RA outperformed SOTA methods on T1w brain MRIs, detecting more global anomalies (AUROC increased from 73.1 to 89.4) and local pathologies (detection rate increased from 52.6% to 86.0%).
+Want to know more? Check the project site.
+Federated disentangled representation learning for unsupervised brain anomaly detection by Cosmin I. Bercea et al. has been published at Nature Machine Intelligence.
++ + + + + + + + + + + + + + + + +
+In this work, a federated algorithm was trained on more than 1,500 MR scans of healthy study participants from four institutions while maintaining data privacy with the goal to detect diseases such as multiple sclerosis, vascular disease, and various forms of brain tumors that the algorithm had never seen before.
+Check the project site for more information.
+Five papers have been accepted for publication at workshops associated with the 26th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2023, which will be held from October 8th to 12th 2023 in Vancouver, Canada.
+Interested to hear more about our work? Then join us at the following workshops:
+Veronika Spieker will be at the DGM4 workshop to talk about Neural Implicit Representations for Abdominal MR Reconstruction on October 8, at 10:25.
+Hannah Eichhorn presents her work on physics-aware motion simulation for T2*-weighted MRI at the SASHIMI workshop on October 8, at 14:40. Check out the preprint for more information!
+Maxime Di Folco presents at the STACOM workshop on October 12, at 11:15 the work of Josh Stein on “Sparse annotation strategies for segmentation of short axis cardiac MRI” (preprint).
+Cosmin Bercea will talk about Bias in Unsupervised Anomaly Detection at the FAIMI workshop on October 12, at 2:50 PDT.
+Daniel Lang will talk about Anomaly Detection in Non-Contrast Enhanced Breast MRI at the CaPTion workshop on October 12.
+Deep Learning for Retrospective Motion Correction in MRI: A Comprehensive Review by Veronika Spieker and Hannah Eichhorn et al. has been accepted for publication at IEEE Transactions on Medical Imaging.
+Federated disentangled representation learning for unsupervised brain anomaly detection by Cosmin I. Bercea et al. has been published at Nature Machine Intelligence.
+The MedtecLIVE Talent Award 2022 is given to bachelor’s and master’s theses that relate to an innovation, improvement, or new application in medical technology along with its entire value chain.
+After a first screening of her thesis abstract, Veronika was invited to the live finale in Stuttgart to present her thesis in an 8-minute pitch. The extensiveness of her work, her drive to clinical translation as well as visual and interactive presentation convinced the jury to award her the first prize.
+As part of her M.Sc. in Medical Technologies at TUM, Veronika conducted her master thesis at the Munich Institute of Robotics and Machine Intelligence (MIRMI) and published her results in Sensors (www.mdpi.com/1424-8220/21/21/7404).
+We are happy, that she is now pursuing her PhD at our lab at Helmholtz Munich!
+More information on the finale can be found here:
+ +Veronika Spieker’s and Hannah Eichhorn’s abstracts have been accepted to be presented as digital posters at the 2023 ISMRM & ISMRT Annual Meeting.
+Veronika Spieker will present her work on “Patient-specific respiratory liver motion analysis for individual motion-resolved reconstruction” on Monday, 05 June 2023 at 1:45 pm EDT.
+Hannah Eichhorn will present her work on “Investigating the Impact of Motion and Associated B0 Changes on Oxygenation Sensitive MRI through Realistic Simulations” on Tuesday, 06 June 2023 at 4:45 pm EDT. Check this GitHub repository for more information.
+ +Deep Learning for Retrospective Motion Correction in MRI: A Comprehensive Review by Veronika Spieker and Hannah Eichhorn et al. has been accepted for publication at IEEE Transactions on Medical Imaging.
+Motion remains a major challenge in MRI and various deep learning solutions have been proposed – but what are common challenges and potentials? Check out this review, which identifies differences and synergies of recent methods and bridges the gap between AI and MR physics.
+Even though auto-encoders (AEs) have the desirable property of learning compact representations without labels and have been widely applied to out-of-distribution (OoD) detection, they are generally still poorly understood and are used incorrectly in detecting outliers where the normal and abnormal distributions are strongly overlapping. In general, the learned manifold is assumed to contain key information that is only important for describing samples within the training distribution, and that the reconstruction of outliers leads to high residual errors. However, recent work suggests that AEs are likely to be even better at reconstructing some types of OoD samples. In this work, we challenge this assumption and investigate what auto-encoders actually learn when they are posed to solve two different tasks. First, we propose two metrics based on the Fréchet inception distance (FID) and confidence scores of a trained classifier to assess whether AEs can learn the training distribution and reliably recognize samples from other domains. Second, we investigate whether AEs are able to synthesize normal images from samples with abnormal regions, on a more challenging lung pathology detection task. We have found that state-of-the-art (SOTA) AEs are either unable to constrain the latent manifold and allow reconstruction of abnormal patterns, or they are failing to accurately restore the inputs from their latent distribution, resulting in blurred or misaligned reconstructions. We propose novel deformable auto-encoders (MorphAEus) to learn perceptually aware global image priors and locally adapt their morphometry based on estimated dense deformation fields. We demonstrate superior performance over unsupervised methods in detecting OoD and pathology.
+ + + + +Motion represents one of the major challenges in magnetic resonance imaging (MRI). Since the MR signal is acquired in frequency space, any motion of the imaged object leads to complex artefacts in the reconstructed image in addition to other MR imaging artefacts. Deep learning has been frequently proposed for motion correction at several stages of the reconstruction process. The wide range of MR acquisition sequences, anatomies and pathologies of interest, and motion patterns (rigid vs. deformable and random vs. regular) makes a comprehensive solution unlikely. To facilitate the transfer of ideas between different applications, this review provides a detailed overview of proposed methods for learning-based motion correction in MRI together with their common challenges and potentials. This review identifies differences and synergies in underlying data usage, architectures, training and evaluation strategies. We critically discuss general trends and outline future directions, with the aim to enhance interaction between different application areas and research fields.
+ + + + +At the end of the module students should be able to recall the important topics in the area of artificial intelligence in medicine, understand the relations between the topics, apply their knowledge to own deep learning projects, analyse and evaluate social and ethical implications and develop own strategies to apply the learned concepts to their own work.
+Content
+Introduction and examples of advanced prediction and classification problems in medicine; ML for prognostic and diagnostic tasks; risk scores, time-to-event modeling, survival models, differential diagnosis & population stratification, geometric deep learning: point clouds & meshes, mesh-based segmentation, shape analysis, trustworthy AI in medicine: bias and fairness, generalizability, AI for affordable healthcare, clinical deployment and evaluation, data harmonization, causal inference, transformers, reinforcement learning in medicine, ML for neuro: structural neuroimaging, functional neuroimaging, diffusion imaging, ML for CVD: EEG analysis
+At the end of the module students should be able to recall advanced topics in the area of artificial intelligence in medicine, understand the relations between the topics, apply their knowledge to own AI projects, analyse and evaluate social and ethical implications and develop own strategies to apply the learned concepts to their own work.
+IN2403 Artificial Intelligence in Medicine
+ ++ + + + + + + + + + + + + + + + +
+Anomaly detection aims to identify patterns that do not conform to the expected normal distribution. Despite its importance for clinical applications, the detection of outliers is still a very challenging task due to the rarity, unknownness, diversity, and heterogeneity of anomalies. Basic problem formulations to recent advances in the field will be discussed.
+This includes, but is not limited to:
+Please register via the TUM matching system: https://matching.in.tum.de
+Check the intro slides here: + + + + + + + + + + + + + + + + +
+ + +Considering the manifold of medical imaging data, i.e. the underlying topological space, facilitates the analysis, interpretation, and visualization of the data. This seminar focuses on machine and deep learning methods that either learn the manifold from high-dimensional data or use manifold-valued data as input. Selected material of methods and applications from the field of medical imaging will be covered. Basic problem formulations to recent advances will be discussed. This includes, but is not +limited to:
+Please register to: https://matching.in.tum.de/m/jz0zflh/q/6wi1lmq4yx
+Check the intro slides here: + + + + + + + + + + + + + + + + +
+ + +Image registration is the process of aligning two or more images, and crucial for many image analysis pipelines. This seminar will cover selected material of image registration for medical imaging. Basic problem formulations to recent advances in the field will be discussed. This includes, but is not limited to:
+Abstract:
+This Master’s project aims to explore the use of covariance descriptors for disease classification with medical +images. First, the MedMNIST toy dataset will be explored. Then, the student will work with an open-source +medical dataset, e.g. of 2D chest x-ray or 3D cardiac MR images
+ +Abstract:
+In recent years, unsupervised and semi-supervised learning from populations of surfaces and curves has received a lot of attention. Such data representations are analyzed according to their shapes which open a broad range of applications in machine learning, robotics, statistics and engineering. In particular, studying the shape of surfaces have become an important tool in biology and medical imaging. The extraction of appropriate data representations, such as triangulated surfaces, is crucial for the subsequent analysis. These surfaces are for example obtained from binary segmentations or 3D point clouds. Using standard methods, such surfaces are often not very accurate and require several post-processing steps, such as smoothing and simplifications. +Deep learning based methods are of great interest in various fields such as medical imaging, com- puter vision, applied mathematics and are successfully used in the field of image segmentation. Gener- ally, a specific formulation requires a particular attention to representations, loss functions, probability models, optimization techniques, etc. This choice is very crucial due to the underlying geometry on the space of representations and constraints. we aim to develop a new set of automatic methods that can compute a triangulation and a normal field from a 3D dataset (binary image and/or 3D point cloud). +The goal of this project is to understand the-state-of-the-art methods (e.g., [?]) and to propose solutions in the context of constructing a mesh from 3D images/point sets. We are interested in learn- ing from a dataset of smooth surfaces and their corresponding 3D datasets to make the triangulation or resampling accurate. The application will be the extraction of a smooth surfaces from μ-CT and CT data of the cochlea and inner ear, whose shapes can then be analyzed subsequently for population studies. +To summarize, the key steps are : (i) Literature review and getting familiar with some state-of- the-art methods in the medical context; (ii) Implementing and testing the code before validation on real data; (iii) Optimizing the code and comparing with baseline methods. If successful, the method would be applied to analyze and classify surfaces.
+ +