A5051295664- Machine Learning in Healthcare
- AI in cancer detection
- Advanced Graph Neural Networks
- Vestibular and auditory disorders
- Radiomics and Machine Learning in Medical Imaging
- Advanced Radiotherapy Techniques
- Bioinformatics and Genomic Networks
- Medical Imaging Techniques and Applications
- COVID-19 diagnosis using AI
- Dementia and Cognitive Impairment Research
- Functional Brain Connectivity Studies
- Advanced X-ray and CT Imaging
- Cerebral Venous Sinus Thrombosis
- Radiation Therapy and Dosimetry
- Anomaly Detection Techniques and Applications
- Computational Drug Discovery Methods
- Balance, Gait, and Falls Prevention
- Acute Ischemic Stroke Management
- Artificial Intelligence in Healthcare and Education
- Obstructive Sleep Apnea Research
- Digital Imaging for Blood Diseases
- Topic Modeling
- Botulinum Toxin and Related Neurological Disorders
- Hearing, Cochlea, Tinnitus, Genetics
- Time Series Analysis and Forecasting
Ludwig-Maximilians-Universität München
2018-2022
Technical University of Munich
2018-2022
German Center for Lung Research
2020-2021
The evaluation of automatic segmentation algorithms is commonly performed using geometric metrics. An analysis based on dosimetric parameters might be more relevant in clinical practice but often lacking the literature. aim this study was to investigate impact state-of-the-art 3D U-Net-generated organ delineations dose optimization radiation therapy (RT) for prostate cancer patients.A database 69 computed tomography images with prostate, bladder, and rectum used single-label U-Net training...
Abstract Background Diagnostic classification of central vs. peripheral etiologies in acute vestibular disorders remains a challenge the emergency setting. Novel machine-learning methods may help to support diagnostic decisions. In current study, we tested performance standard and approaches consecutive patients with or disorders. Methods 40 Patients stroke (19 21 without syndrome (AVS), defined by presence spontaneous nystagmus) 68 AVS due neuritis were recruited department, context...
In-vivo MR-based high-resolution volumetric quantification methods of the endolymphatic hydrops (ELH) are highly dependent on a reliable segmentation inner ear's total fluid space (TFS). This study aimed to develop novel open-source ear TFS approach using dedicated deep learning (DL) model.The model was based V-Net architecture (IE-Vnet) and multivariate (MR scans: T1, T2, FLAIR, SPACE) training dataset (D1, 179 consecutive patients with peripheral vestibulocochlear syndromes). Ground-truth...
Deep learning has been recently applied to a multitude of computer vision and medical image analysis problems. Although recent research efforts have improved the state art, most methods cannot be easily accessed, compared or used by other researchers clinicians. Even if developers publish their code pre-trained models on internet, integration in stand-alone applications existing workflows is often not straightforward, especially for clinical partners. In this paper, we propose an open-source...
Background: Multivariable analyses (MVA) and machine learning (ML) applied on large datasets may have a high potential to provide clinical decision support in neuro-otology reveal further avenues for vestibular research. To this end, we build base-ml, comprehensive MVA/ML software tool, it three increasingly difficult objectives differentiation of common disorders, using data from prospective patient registry (DizzyReg). Methods: Base-ml features full pipeline classification multimodal data,...
Clinical diagnostic decision making and population-based studies often rely on multi-modal data which is noisy incomplete. Recently, several works proposed geometric deep learning approaches to solve disease classification, by modeling patients as nodes in a graph, along with graph signal processing of features. Many these are limited assuming modality- feature-completeness, transductive inference, requires re-training the entire model for each new test sample. In this work, we propose novel...
Abstract Background The evaluation of the automatic segmentation algorithms is commonly performed using geometric metrics, yet an based on dosimetric parameters might be more relevant in clinical practice but still lacking literature. aim this study was to investigate impact state-of-the-art 3D U-Net-generated organ delineations dose optimization intensity-modulated radiation therapy (IMRT) for prostate patients first time. Methods A database 69 computed tomography (CT) images with prostate,...
Geometric deep learning provides a principled and versatile manner for the integration of imaging non-imaging modalities in medical domain. Graph Convolutional Networks (GCNs) particular have been explored on wide variety problems such as disease prediction, segmentation, matrix completion by leveraging large, multimodal datasets. In this paper, we introduce new spectral domain architecture graphs prediction. The novelty lies defining geometric 'inception modules' which are capable capturing...
In large population-based studies and in clinical routine, tasks like disease diagnosis progression prediction are inherently based on a rich set of multi-modal data, including imaging other sensor scores, phenotypes, labels demographics. However, missing features, rater bias inaccurate measurements typical ailments real-life medical datasets. Recently, it has been shown that deep learning with graph convolution neural networks (GCN) can outperform traditional machine classification, but...