A5065367041- Advanced Neuroimaging Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- Dementia and Cognitive Impairment Research
- Medical Image Segmentation Techniques
- Brain Tumor Detection and Classification
- Tensor decomposition and applications
- Industrial Vision Systems and Defect Detection
- Ferroelectric and Piezoelectric Materials
- Machine Learning in Healthcare
- Multiferroics and related materials
- Fetal and Pediatric Neurological Disorders
- Advanced MRI Techniques and Applications
- Magnetic and transport properties of perovskites and related materials
- Advanced Neural Network Applications
- MRI in cancer diagnosis
- Generative Adversarial Networks and Image Synthesis
- Microplastics and Plastic Pollution
ETH Zurich
2016-2020
Our ever-aging society faces the growing problem of neurodegenerative diseases, in particular dementia. Magnetic Resonance Imaging provides a unique tool for non-invasive investigation these brain diseases. However, it is extremely difficult neurologists to identify complex disease patterns from large amounts three-dimensional images. In contrast, machine learning excels at automatic pattern recognition data. particular, deep has achieved impressive results image classification....
We investigate the effect of chemical doping on electric and magnetic domain pattern in multiferroic hexagonal ErMnO3. Hole- electron are achieved through growth Er1−xCaxMnO3 Er1−xZrxMnO3 single crystals, which allows for a controlled introduction divalent tetravalent ions, respectively. Using conductance measurements, piezoresponse force microscopy nonlinear optics we study doping-related variations electronic transport image corrsponding ferroelectric antiferromagnetic domains. find that...
Microplastics pollution has been recognized as a serious environmental concern, with research efforts underway to determine primary causes. Experiments typically generate bright-field images of microplastic fibers that are filtered from water. Environmental decision making in process engineering critically relies on accurate quantification mi-croplastic these images. To satisfy the required standards, often analyzed manually, resulting highly tedious process, thousands fiber instances per...
Abstract White matter tractography, based on diffusion-weighted magnetic resonance images, is currently the only available in vivo method to gather information structural brain connectivity. The low resolution of diffusion MRI data suggests employ probabilistic methods for streamline reconstruction, i.e., fiber crossings. We propose a general model spherical regression Fisher-von-Mises distribution, which efficiently estimates maximum entropy posteriors local directions with machine learning...
Elaborate expert modeling has been the predominant approach to fiber tractography. It attempts invert measurement process of diffusion-weighted MRI reconstruct fibers. We present a purely data-driven neural network regression model for The sequentially takes as input local block data and incoming direction fiber. From this input, predicts outgoing direction. training can be provided by either automatic or human supervision. On both real, synthetic we observe that our produces smoother more...
Automatic diagnosis of Alzheimer's disease (AD) from MR images the brain promises to yield important information a patient's status or even early prediction onset. This work investigates deep learning based methods predict conversion Mild Cognitive Impairment (MCI) AD on widely available T1-weighted images. We present novel approach breaking up into generative and discriminative step. Using recently proposed Wasserstein-GAN model, we generate synthetically aged image given baseline image....