UROP Project

Deep Learning on psychiatric MR imaging data

The aim of our research is to develop and integrate artificial intelligence into the workflow of a psychiatric clinic. Suitably trained neural networks provide a powerful tool for computer assisted diagnosis and may present further insight into psychiatric diseases. In this regard, pathological changes in white matter brain substance have been identified to correlate with mental health status. Typically, those changes appear as white matter hyperintensities in T2-weighted MR images. We aim to utilize Deep Learning methods to automatically detect and evaluate those white matter hyperintensities. This will allow us to further investigate correlation between number, size and severity of white matter changes and clinical non-imaging parameters.

Task

During the internship, you will get an insight into our research in the field of machine learning. You will work with our software frameworks and you will have the opportunity to gain experience in many tasks involved in creating and training of artificial neural networks. Depending on your interests, you have the possibility to contribute by completing a small project, tailored around these tasks. These might include, but are not limited to: • Implementation and training of artificial neural networks • Preparation and preprocessing of input data for machine learning • Coding of experiment related tasks using the programming language Python

Requirements

programming skills (e.g. cpp, Python, Matlab) preferably experience with machine learning (SVM, DNN, etc) preferably experience with computer vision