Machine Learning and Data Science Group

Research Topics

In our machine learning and data science group we develop novel ways of handling all data acquired in the context of radiological examinations to leverage recent advances in data science and computer vision to improve patient care and provide added value in various clinical settings.

Our interests focus on improving radiological reports through structured reporting as well as quantitative imaging and the development and deployment of deep learning techniques to improve the imaging workflow, from image acquisition to image interpretation. It is our belief that through openness and transparency with regards to the employed algorithms and methods we can strengthen radiology’s role as leader of innovation in medicine.

Teaching

We look to establish data science methods as an integral part of modern radiological education, as these emerging technologies will play a key role in the future of radiology and probably the whole of medicine.

Current Projects

  • Enhancement of current clinical infrastructure to allow for large-scale machine learning projects
  • Deep learning based image classification system for fully automated quality assurance with PACS integration
  • Fully automated detection of urgent findings in radiographs and CT with notification system to radiologists and referrers
  • Eye-tracking for reading of radiological reports
Collaborations

Multiparametric Imaging and Radiomics Group
Cancer & Immunometabolism Research Group, University Hospital of Cologne
Institute of German Language and Literature I, University of Cologne
Visual Computing Group, Institute of Computer Science, University of Mainz
Department of Radiology, University Medical Center Mainz
European Society of Medical Imaging Informatics
Working Group on Information Technology, German Radiological Society
Philips Healthcare, Best, The Netherlands

Group members

Daniel Giese, MD
Bettina Baeßler, MD
Jan Borggrefe, MD
Khaled Bhousabara, MSc

Nach oben scrollen