The Imaging Research Warehouse (IRW) at Mount Sinai continuously gathers routine medical images. I create an interactive dashboard which summarizes imaging metadata and medical concepts occuring in radiology reports to extract datasets for applied AI projects.
External validation of machine learning models is critical to test their use in clinical practice. We build a machine learning model store for medical imaging which enables researchers to share models and simplify their validation. Models trained on public datasets are used to test how well they generalize on IRW data.
Once a vision model is trained, it may be used in a different domain but fails to adapt to anatomical, acquisition or scanner differences. In this project, I research on methods using domain adaptation (given no target labels, no source data) and semi-supervision (given limited target labels, abundant unlabeled instances) to account for label scarcity.
Contact
Chair Representative:
Prof. Dr. Christoph Lippert Professor for Digital Health & Machine Learning Room: G-2.1.xx Tel.: +49-(0)331 5509-4850 E-Mail: christoph.lippert(at)hpi.de
Office:
Campus III, Haus G2 Room: G-2.1.22 Tel.: +49-(0)331 5509-4850 Fax: +49-(0)331 5509-4849 E-Mail: office-lippert(at)hpi.de
Visiting address:
Campus III Building G2 Rudolf-Breitscheid-Straße 187 14482 Potsdam, Germany
