A5006032492- Cardiac Imaging and Diagnostics
- Intracranial Aneurysms: Treatment and Complications
- Advanced X-ray and CT Imaging
- Cardiac Valve Diseases and Treatments
- Cerebrovascular and Carotid Artery Diseases
- Traumatic Brain Injury and Neurovascular Disturbances
- Radiomics and Machine Learning in Medical Imaging
- Aortic aneurysm repair treatments
- Advanced MRI Techniques and Applications
- Medical Imaging Techniques and Applications
- Medical Image Segmentation Techniques
- Infective Endocarditis Diagnosis and Management
- Cardiovascular Function and Risk Factors
- Artificial Intelligence in Healthcare and Education
- Robotics and Sensor-Based Localization
- Cardiac tumors and thrombi
- Aortic Disease and Treatment Approaches
- Medical Imaging and Analysis
- AI in cancer detection
- Lung Cancer Diagnosis and Treatment
- Elasticity and Material Modeling
- Gas Dynamics and Kinetic Theory
- Cardiac and Coronary Surgery Techniques
- Retinal Imaging and Analysis
- Soft Robotics and Applications
Deutsches Herzzentrum der Charité
2023-2024
Charité - Universitätsmedizin Berlin
2020-2024
Fraunhofer Institute for Digital Medicine
2024
Humboldt-Universität zu Berlin
2023
Freie Universität Berlin
2023
Berlin Heart (Germany)
2021
PurposeAnalyzing the anatomy of aorta and left ventricular outflow tract (LVOT) is crucial for risk assessment planning transcatheter aortic valve implantation (TAVI). A comprehensive analysis root LVOT requires extraction patient-individual via segmentation. Deep learning has shown good performance on various segmentation tasks. If this formulated as a supervised problem, large amounts annotated data are required training. Therefore, minimizing annotation complexity desirable.ApproachWe...
The quality and acceptance of machine learning (ML) approaches in cardiovascular data interpretation depends strongly on model design training the interaction with clinical experts. We hypothesize that a software infrastructure for application ML models can support improvement provide relevant information understanding classification-relevant features. presented solution supports an iterative training, evaluation, exploration machine-learning-based multimodal methods considering cardiac MRI...
4D PC MRI of the aorta has become a routinely available examination, and multitude single parameters have been suggested for quantitative assessment relevant flow features clinical studies diagnosis. However, clinically applicable complex patterns is still challenging. We present concept applying radiomics characterization in aorta. To this end, we derive cross-sectional scalar parameter maps related to literature such as throughflow, direction, vorticity, normalized helicity. Derived are...
Radiomics has been applied in cardiovascular imaging with promising results. Still, the effect of parameters on reproducibility radiomics features needs to be well understood for these used clinical diagnostics. We conduct a retrospective study short-axis cine CMR images healthy volunteers assess impact different temporal sampling shape, first-order, texture, and deformation field-based features, contextualize findings inter- intra-observer variability. find that is dependent sampling,...
Motivation: HFpEF (heart failure with preserved ejection fraction) patients show similar diagnostic parameter values to healthy subjects in the quantitative evaluation of cardiac magnetic resonance imaging (CMR). Improving performance CMR-based HF characterization is desirable. Goal(s): Evaluation radiomics features from 2D flow MRI aorta for classification into healthy, HFpEF, HFmrEF (mildly reduced) and HFrEF (reduced). Approach: Training a decision tree classifier using 70 CMR datasets...
Abstract Minimally invasive surgery is increasingly utilized for mitral valve repair and replacement. The intervention performed with an endoscopic field of view on the arrested heart. Extracting necessary information from live video stream challenging due to moving camera position, high variability defects, occlusion structures by instruments. During such minimally interventions there no time segment regions interest manually. We propose a real-time-capable deep-learning-based approach...
Abstract Cardiac diseases manifest in a multitude of interconnected changes morphology and dynamics. Radiomics approaches are promising technique to analyze such directly from image data. We propose novel features specifically describe moving cardiac structures, an interactive 4D visualization method explore Prototypical tests with open data set containing different show that our approach can be fast useful tool for the analysis heterogeneous cohort