KimBi: On the way to AI-supported intelligent magnetic resonance imaging

The AI Center for Health Care project On the way to AI-supported intelligent magnetic resonance imaging aims to develop an application-oriented language for the development of imaging techniques in magnetic resonance imaging that enables the support of efficient machine learning processes and thus automatically selects the best possible imaging.

From left to right: Dr. Daniel Hoinkiss, Christina Plump, Jörn Huber, Ilseok Lee. © Shanice Allerheiligen / U Bremen Research Alliance

Activation required

By clicking on this thumbnail, you consent to the downloading of content from YouTube (USA). This provides Google (USA) with the information that you have accessed our site as well as the technically necessary data in this context. We have no influence on the further data processing by Google. Please note that Google and YouTube do not provide an adequate level of data protection.
Further information / data protection

In this project, the University of Bremen, Fraunhofer MEVIS and the German Research Center for Artificial Intelligence (DFKI) are collaborating to develop an application-oriented language for the development of imaging techniques in magnetic resonance imaging (MRI) that enables the support of efficient machine learning processes and thus automatically selects the best possible imaging.

The focus is on the development of a domain-specific language for MR sequences that makes it possible to describe the complex patterns and regularities of these sequences. This language allows MR sequences to be programmed automatically and more efficiently, which was previously done by laborious manual programming. The innovative method makes it possible to optimize MR sequences using machine learning processes in order to automatically implement the best possible MR sequence for specific requirements. This represents a revolutionary step away from a purely chronological, physical description towards a rule-based, application-oriented approach. This could significantly shorten the development times of new techniques, especially in patient-specific imaging, and facilitate the handling and testing of MR sequences.

For patients, this means shorter waiting times and faster diagnosis, as measurement times in clinics can be shortened. Intelligent systems that analyze the data during the MRI measurement and adapt the measurement have the potential to detect physiological abnormalities at an early stage, even before structural changes become visible. The combination with the gammaSTAR MR sequence development environment also promises a rapid transfer of the developed techniques into broad application, which represents a significant advance in medical imaging.

"A researcher or a clinician, without any prior MRI sequence programming knowledge, can go to an MRI, use our optimization tool for the perfect MRI sequence and then, gets a solution that has then only to be crosschecked by an expert."
- Christina Plump, Doktorandin DFKI