Navigation auf uzh.ch
Navigation auf uzh.ch
The aim of this project is to use and develop novel (potentially ML-based) algorithms to obtain a deeper understanding of the graph-structured vasculature preserved in corrosion casts. To this end, potential projects include whole-brain quantitative analysis, graph neural networks, label propagation, and/or novel graph-based clustering approaches. A nice starting point would be to explore weakly supervised label propagation approaches (see following ICLR paper https://arxiv.org/abs/2210.03594). This project is mostly about the development of novel machine learning methods. The data we are working on has already been cleaned by me. So we can start right away with the exciting research part :).
Keywords: Neuroscience, Brain Vasculature, Medical AI, Machine Learning, Network Analysis, Biomedical Image Analysis, Whole Brain Analysis
Please send your CV and transcript to Thomas Wälchli (thomas.waelchli2@uzh.ch) and Bastian Wittmann (bastian.wittmann@uzh.ch). Links to previous work (e.g., your GitHub profile) are highly appreciated.
The aim of this project is to use and develop novel machine learning algorithms to segment high-resolution bone images and characterize cortical and trabecular bone structures. To this end, we aim to combine novel ideas tailored to cortical and trabecular bone structures with common segmentation practices and represent the trabecular bone structures as a graph representation with subsequent analysis.
Keywords: Segmentation, Deep Learning, Medical AI, Machine Learning, Computer Vision, Biomedical Image Analysis
Please send your CV and transcript to Bastian Wittmann (bastian.wittmann@uzh.ch) and Serena Bonaretti (serena.bonaretti@balgristcampus.ch)
In this master's thesis project, we are looking for a candidate to apply machine learning techniques to correct and predict signals of incomplete CT perfusion imaging for ischemic stroke. We hope to use machine learning techniques to de-noise and correct for the truncation in CT perfusion signals. In particular, we aim to infer the true attenuation curve after the truncation time-point.
Keywords: machine learning; CT perfusion imaging; ischemic stroke; contrast-media attenuation time-curves;
If you are interested, please apply through Kaiyuan Yang (kaiyuan.yang@uzh.ch) and Dr Til Schubert (https://www.usz.ch/team/tilman-schubert/)
The aim of this work is to develop a model for automatically predicting the cellular stage from single-cell microscopy images, as such a model would facilitate the personalization of treatments for patients suffering from rheumatic diseases. Therefore, the functional stage of synovial fibroblasts (SF) - the cells of interest - should be classified into biologically meaningful classes based on physiological processes such as mitochondrial activity, oxidative stress or apoptosis. Since some cells cannot be clearly assigned to a specific class, it may be interesting to use not only supervised but also semi- or unsupervised approaches. All in all, the final goal is an easy-to-use pipeline for single cell segmentation and classification that provides biologically meaningful outputs and visualizations.
Keywords: Deep Learning, Medical AI, Machine Learning, Computer Vision, Biomedical Image Analysis
Please send your CV and transcript to Paul Büschl (paul.bueschl@uzh.ch) and Eva Camarillo (eva.camarilloretamosa@usz.ch)
