Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

Our world is 3D and so is the patient. But visual camera observations are 2D. Advancements in digital imaging now enable capturing rich 3D information of our surrounding, transforming our ability to digitally perceive, present, and analyze data. By combining multiple data sources or frames in a video, we can create context-enhanced digital copies of the physical world as well as the patient - and apply data-driven interpretations and processing. Data sources reach from ceiling mounted cameras in the OR to endoscopic images or robot-mounted sensors. In this talk, we want to look into underlying core ideas of example 3D computer vision pipelines and investigate how these approaches can be used in clinical applications. As a basis for a discussion at this exciting interdisciplinary crossroad, we dive into the subtopics 3D digital reconstruction, data curation for XR, and robot-assisted surgery.

Speaker: Dr. Benjamin Busam

Biography

Benjamin Busam is a Senior Research Scientist with the Technical University of Munich. He coordinates the Computer Vision activities at the Chair for Computer Aided Medical Procedures, I16. Formerly Head of Research at FRAMOS Imaging Systems, he led the 3D Computer Vision & AI Team at Huawei Research, London from 2018 to 2020. Benjamin studied Mathematics at TUM (Germany), ParisTech (France) and at the University of Melbourne (Australia), before he graduated with distinction at TU Munich in 2014. In continuation to a mathematical focus on projective geometry and 3D point cloud matching, he now works on 2D/3D computer vision and sensor fusion and applications into the medical domain. For his work on adaptable high-resolution real-time stereo tracking, he received the EMVA Young Professional Award 2015 from the European Machine Vision Association and was awarded Innovation Pioneer of the Year 2019 by Noah's Ark Laboratory, London. He was given multiple Outstanding Reviewer Awards at 3DV 2020, 3DV 2021, and ECCV 2022.

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

Due to the rapid progress in developing experimental techniques, establishing and improving analysis methods is one of the major challenges in computational life sciences. For many analysis tasks, however, the limitations and performance of competing methods remain unknown, and there are no clear rules or guidelines for selecting the optimal analysis method. Benchmark studies have proven to be valuable tools for evaluating the performance and applicability of analysis approaches. However, they are often subject to methodological limitations and deficiencies, leading to potential bias in the results.

In my presentation, I will give an overview of novel approaches developed in my group in the context of mathematical modeling and omics analyses. In particular, I will summarize our ongoing efforts to improve the methodology of benchmark studies. By generally incorporating rigorous planning, design, and analysis principles in benchmark studies, we aim to promote the development of novel analysis approaches and the identification of decision rules for optimal method selection in practice. Especially in view of the enormous efforts to apply deep learning methods in all areas of research, reliable performance comparisons are of great importance.

Speaker: Dr. Clemens Kreutz

Biography

Dr. Clemens Kreutz is the Head of the Methods of Systems Biomedicine (MSB) group and Deputy Director at the Institute of Medical Biometry and Statistics, Medical Center, Faculty of Medicine, University of Freiburg. Since 2018, he has held a permanent position as Group Leader at the institute, and in 2021, he was appointed Deputy Director. Dr. Kreutz's research focuses on the mathematical modeling and analysis of high-throughput data. His work within the field of systems biology has significantly advanced statistical methods for parameter estimation, model selection, and experimental design, contributing to the understanding and modeling of complex biological processes.

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

The primary aim of this talk to develop and apply biostatistical methods to improve predictive modeling of data for precision medicine. In order to achieve such a solid framework for different type of datasets the selection of methods are important. However, currently their is no universal method which can be applied for data gathered from humans. To understand the given data at hand, first statistical tools which considers the distribution of the data namely Bayesian posterior distribution will be discussed. Second, to counter one of the most important problems in most of the clinical datasets is the demographics namely matching cohorts for sex would be considered with propensity score matching analyses. In order to model complex datasets, with several input and output variables the structural equation modelling will be delibrated. After understanding and modelling the datasets the prediction will be covered with some machine and deep learning approaches and finally some applications to datasets from multimodal, longitudinal and signal based analyses will be explored. For each methodological aspect an example will be provided with obtained results.  The applications will be highlighted with examples and corresponding results. Taken together, the integration of methods leading to the individualized prediction of each subject will be demonstrated.

Speaker: Prof. Muthuraman Muthuraman

Biography

Muthuraman Muthuraman was born in Chennai, India, in 1980. He received the B.E. degree in electronics and communication engineering from the University of Madras, Madras, India, in 2002, and the MS in digital communications in Christian Albrecht’s University, Kiel, Germany in 2006. Ph.D. degree in biomedical engineering from the technical faculty and Department of Neurology of Christian Albrecht’s University, Kiel, Germany, in 2010. In 2010, he joined the Department of Neurology, University of Kiel, as a Post-doc, and in 2013 became a senior post-doc. Since December 2016, he has been with the Department of Neurology, Johannes Gutenberg University Mainz, where he is an Assistant Professor, and the head of the department biomedical statistics and multimodal signal processing unit. Currently from 2024 he is heading the group Informatics for Medical Technology (IMT) in Augsburg as an associate professor and second affiliation to Julius Maximilian university of Würzburg in the department of Neurology and head of the group Neural Engineering with Signal Analytics and Artificial Intelligence (NESA-AI). His current research interests include mathematical methods for time series analysis and source analysis on oscillatory signals, sleep, function of oscillatory activity in central motor systems, biomedical statistics, connectivity analyses, multimodal signal processing and analyses of EEG, MEG, fMRI and EMG, structural and network analyses on anatomical MRI and DTI, functional network analyses on PET imaging, machine learning and deep learning, network analyses on proteomic and genomic data, RNA, mRNA and Spatial transcriptomics.

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

Epigenetics, which involves heritable changes in gene expression without DNA sequence alterations, plays a crucial role in tumour classification, with DNA methylation serving as a key biomarker. Tumour classification based on DNA methylation patterns has emerged as a pivotal approach in precision diagnostics, offering insights into cell-of-origin as well as tumour progression. This talk will explore the evolution of methylation-based classifiers, with a focus on CNS tumours, and highlight recent advances in using methylation profiling to achieve accurate and scalable diagnostics. Additionally, the talk will introduce development and application of novel technologies like nanopore sequencing and AI-based computational histopathology to improve accessibility of cutting-edge molecular diagnostics. 

Speaker: Areeba Patel

Biography

Areeba Patel is a postdoctoral researcher at the German Cancer Research Center (DKFZ) affiliated to the departments of Neuropathology, Paediatric Neurooncology and AI in Oncology. She completed her masters in Cancer Biology at Imperial College London followed by PhD in computational neuro-oncology at the DKFZ. Her research revolves around developing novel solutions using nanopore sequencing as well as AI-based computational histopathology to make precision CNS tumour diagnostics swift, accessible and affordable.

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

In his talk, he will explore how computer vision tools are impacting the analysis of movement disorders and their potential going forward. In particular, he will discuss recent work on tracking the head movements of cervical dystonia patients to automate clinical scoring and identify novel kinematic biomarkers. Finally, he will also introduce some novel approaches for analysing whole-brain EEG data via the vector fields of phase-gradients.

Speaker:  Dr. Robert Peach

Biography

Robert Peach is an independent research group leader at the University Hospital Würzburg's Neurology Department and a Senior Research Fellow in the Department of Mathematics at Imperial College London. He earned his Ph.D. in Computational Biology from Imperial College London, where he explored protein dynamics using graph theory and single-molecule spectroscopy. His current research stretches from mathematics to neuroscience with a focus on applications in movement disorders. In particular, he specialises in using and developing methods in computer vision, geometry, and machine learning to automate the analysis of movement disorders, both in the kinematic and neural measurement domains. In his talk, he will explore how computer vision tools are impacting the analysis of movement disorders – with various examples - and their potential going forward. Finally, he will also introduce some novel approaches for analysing EEG data.

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

Bioimpedance analysis is an established non-invasive method that is already in clinical use for body composition measurement or electrical impedance tomography. However, novel instrumentation and signal processing approaches enable significant improvements of these applications as well as new measurement approaches. This presentation will give a brief insight into the theory of bioimpedance analysis and discuss instrumentation approaches. Subsequently, examples of new scientific measurement approaches will be presented. The focus will be on the detection of skeletal muscle contractions (impedance myography) and pulse wave detection (impedance plethysmography).

Speaker: Prof. Dr. Roman Kusche

Biography

Roman Kusche has been a Professor in the Department of Computer Science at HAW Hamburg since 2023, where he leads the Medical Sensors Lab. His research interests include the development of novel biomedical measurement methods and related medical electronic devices. Before joining HAW Hamburg, he was the Head of the Department for Individualized Therapy and led the Medical Electronics research group at Fraunhofer IMTE. Roman Kusche has a background in electrical engineering and received his Ph.D. from the Institute of Medical Engineering at the University of Lübeck.

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

Artificial Intelligence (AI) is transforming our world and society with extraordinary speed. What does this mean for biomedical research and disease detection? What opportunities and risks arise from this? In my talk, I will present examples from the biomedical and clinical research conducted by my research group at the Helmholtz Center Munich. I will cover the fundamentals of machine learning and AI and showcase applications that allow individual cells to be characterized with unprecedented accuracy. These technological breakthroughs promise new perspectives for both basic research and the personalized medicine of the future.

Speaker:  Dr. Carsten Marr

Biography

Carsten Marr is the founding director of the Institute of AI for Health at Helmholtz Munich, a European center for applied artificial intelligence. His goal is to develop AI-based methods to improve the diagnosis, treatment, and understanding of diseases. After studying theoretical physics at the Technical University of Munich and completing his diploma thesis at the Max Planck Institute of Quantum Optics, Carsten switched from physics to theoretical biology. His PhD thesis at Technical University of Darmstadt focused on the architecture of biological networks and was awarded the best of its year in the Department of Biology. After postdoctoral research stays in Munich, Bremen and Edinburgh, he started his research group at the Helmholtz Munich in 2013 and became deputy head of the Institute of Computational Biology. In interdisciplinary projects with experimentalists, biomedical experts, and clinicians, he pioneered the training of deep neural networks on life science data for the prediction of stem cell decisions from microscopic images and the identification of leukemia from blood and bone marrow smears. He has received several awards for his research and analysis of single cell data as well as an ERC Consolidator Grant for the training of AI models for the automated analysis of blood diseases. 

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

Speaker: Prof. Ralf Seepold

Biography

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

Speaker: Anna Danese, M.Sc.

Biography

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract
Wearables have been the dominant consumer monitoring device for the past decades, however, not the ideal one. Battery dependency, correct placement, good contact with skin, etc., all impose requirements and restrictions that can’t be met by everyone (e.g., vulnerable populations, skin conditions, etc.). In light of this and in accordance with the ultimate idea of ubiquitous computing, the interaction between the user and sensor should be minimized or ideally removed entirely. This is being achieved with improvements and availability of sensors – namely RGB cameras, thermal cameras, radars, LIDARs, and others – accompanied by immense development of machine learning, specifically deep neural networks. In this talk we will overview existing methods for unobtrusive contact-free monitoring of physiological parameters and some psychological states that can be derived from them.

Speaker: Dr. Gašper Slapničar

Biography

Dr. Gašper Slapničar is the head of the ambient intelligence group at the department of intelligent systems, Jožef Stefan Institute. He finished his undergrad studies in computer science from the faculty of computer and information science, university of Ljubljana, and then pursued his PhD at Jožef Stefan international postgraduate school, focusing on contact-free physiological monitoring using radar and optical sensors. He is the author of over 30 scientific papers, many published in impact factor journals and at established conferences such as ICCV and BHI. He is also a regular program committee member for the computer vision for physiological measurement (CVPM) workshop. His ongoing research includes unobtrusive contact-free monitoring of complex states, including emotions and well-being.

Location: Lecture hall N2045 (Faculty of Applied Computer Science)

Abstract

Speaker:  Dr. Johannes Tran-Gia

Biography