Research Centre of Computational Health
The devision of Computational Health is specialised in the confluence of data-driven and mechanistic modelling approaches in medicine and biology. The methods encompass machine learning for image and signal analysis, graphical networks, optimization of stochastic models and physiological simulations.
The centre Computational Health addresses fundamental questions in biology and medicine using computer-assisted, data-driven methods. Important tools are machine learning for image and signal analysis, parameter estimation for differential equation systems and multiphysics simulation. Effective validation strategies are used to deal with the unknown.
The research group specializes in modeling biological and medical systems. New approaches are developed to simulate physiological processes and to predict pathological changes. In particular, in-depth knowledge of biological/physiological processes is incorporated into multi-physics simulations.
The group develops algorithms for parameter and uncertainty estimation of physically motivated stochastic models. In particular, machine learning methods are combined with Bayesian modeling for dimensionality reduction. These methods are widely used in medicine and life sciences.
The research group applies machine learning techniques to interpret medical image data. This way, features are extracted for the characterization of disease patterns and for use as diagnostic markers. Of particular interest are the radiomic and morphological analysis of diagnostic medical imaging data. The group pursues the goal of establishing reproducible, image-based biomarkers by means of explainable artificial intelligence and ensuring their clinical utility.
The research group uses statistical and machine learning methods to model and uncover causal relations in medical data, especially to study pathophysiological processes. Categorial patient data and imaging date, e.g. magnetic resonance imaging, are processed to extract clinical knowledge.
The research group studies data from wearables and biosensors using time series analysis and combines them with biological-physical models to robustly characterize physiological systems. These data sources are used for Patient Reported Outcomes in clinical practice and for the further development of patient-centered medicine.
OSR4H – Open Set Recognition for Hematology
Development of a Proof of Concept for visual Open Set Recognition (OSR) algorithms applied to a Hematology task, the classification of white blood cells.
Deep Brain Vessel Profiler
The architecture of the supplying brain blood vessels is believed to impact the occurrence and severity of common cerebrovascular diseases such as ischemic strokes or intracranial aneurysms. In this project, we study methods to efficiently quantify the variability in human cerebral vasculature and how this ...
Smart Hospital – Integrated Framework, Tools & Solutions (SHIFT)
Automated Release Testing System (ARTS)
Acthera Therapeutics AG and ZHAW propose to build an Automated Release Testing System (ARTS) for mechanotherapeutic drugs. The device is important for Acthera for testing new drugliposome formulations, controlling quality, keeping an edge over competition and collaborating with regulatory bodies. ...
Machine learning for NMR spectroscopy
The goal of this project is to make NMR spectroscopy available to a wider range of applications and to non-experts by the automation of data reduction and analysis steps, in particular by combining deep learning methods for the extraction and a Bayesian approach for the integration and refinement of information. ...
12(8), pp. 634.
Available from: https://doi.org/10.3390/bios12080634
Castelnovo, Anna; Amacker, Julian; Maiolo, Massimo; Amato, Ninfa; Pereno, Matteo; Riccardi, Silvia; Danani, Andrea; Ulzega, Simone; Manconi, Mauro,
(155), pp. 62-74.
Available from: https://doi.org/10.1016/j.cortex.2022.05.021
Computers in Biology and Medicine.
Available from: https://doi.org/10.1016/j.compbiomed.2022.105740
Frontiers in Neurology.
Available from: https://doi.org/10.3389/fneur.2022.809391
Available from: https://doi.org/10.21256/zhaw-24219