Fabian Bongratz

Medical image data, for example recorded by MR or CT scanners, often form the basis for individual diagnoses and disease treatments. Furthermore, medical scans are recorded as part of population studies with the aim of better understanding and combating common diseases such as diabetes or Alzheimer’s disease. While the acquisition of images is now not too much of an obstacle due to technical advances in scanners, their analysis by trained radiologists is very limited. The goal of this dissertation project is to close this gap and enable fast, accurate, reliable and automated analysis of medical scans. For this purpose, new learning-based algorithms are being researched that include 3D surface models of the organs in addition to the images.