The newly founded working group “Group for Applied AI in Neuroradiology” (GAAIN) of the University Clinic for Neuroradiology at the OVGU Magdeburg is dedicated to the exciting challenge of integrating artificial intelligence (AI) into everyday clinical practice. Our goal is to develop innovative AI technologies that revolutionize medical care and increase the efficiency of clinical processes.

Our current research focuses include:

1. More precise diagnoses with less radiation exposure:
   Development and application of advanced deep learning models, such as conditional Generative Adversarial Networks (cGAN) [1], to optimize diagnostic imaging while reducing patients' radiation exposure.
2. Automated segmentation of smallest structures:
   Use of modern AI technologies for automated and high-precision segmentation of the finest anatomical structures in high-resolution 7T MRI data..
3. Use of LLMs (Large Language Models) in clinical communication:
   Research into the potential of LLMs to improve patient education, optimize communication between doctors and patients and support healthcare professionals in everyday clinical practice..
4. Development of new approaches based on extensive medical data sets:
   Our group focuses on the creative use of large, multidimensional medical data sets to develop novel deep-learning methods that lead to improved diagnostic and therapeutic procedures.

A particular focus is on interdisciplinary cooperation:

Our working group combines expertise from medicine, natural sciences, computer science, and engineering to solve complex problems and create synergies between disciplines. This close cooperation also extends across institutes and faculties and thus enables unique innovation potential..

Our vision is to use our research to pave the way for patient-centered and AI-supported medicine of the future. The AG GAAIN invites all who share this mission to work with us at the interface of innovation and clinical application..

Interested in writing a thesis?

For interested students, we offer the opportunity to work on exciting and practice-oriented topics for final theses (Bachelor, Master, Dr. med., Dr. rer. medic., etc.) in a highly topical research environment. Together we design projects that are not only academically relevant, but also clinically significant. 

Figure 1 - (a) Basic concept of subtraction imaging and post-processing using cGAN; (b) Contrast-enhanced image (left), subtraction image with motion artifacts (center) and generated cGAN subtraction image (right). 

Figure 2 - Segmentation masks of PVS in 7T MRI data: manual segmentation (left), Frangi-based method (center) and deep learning-based method (right). 

Employees:

Publications: