The KI-SIGS project successfully concluded last month after 3.5 years of intensive work. The project was focused on developing an "AI space for intelligent healthcare systems" in collaboration with AI institutes in northern Germany, located in Bremen, Hamburg, and Schleswig-Holstein, as well as medical technology companies and university hospitals. Our contribution was in the field of "AI for radiological imaging" within the context of stroke applications. We have previously informed you about the project's content and our work.
From September 12 to 14, 2023, the "KI-MED Connect – the northern AI Booster" conference took place on the Lübeck campus. The conference centered on the application of Artificial Intelligence in the healthcare sector. Experts from the industry, universities, and hospitals gathered to discuss the use of Artificial Intelligence in healthcare, covering topics from data collection to ethical considerations and transparency to High Performance Computing. A special part of the conference was the presentation of selected projects. We also had the honor to give a presentation about our subproject.
In collaboration with the University Hospital Schleswig-Holstein (UKSH), algorithms to support radiologists in interpreting stroke CT scans were developed. Specifically, algorithms for detecting brain hemorrhages and locating vascular obstructions in native CT scans were trained and tested. Detecting vascular obstructions is a crucial yet challenging task, where radiologists could greatly benefit from AI-based support. Our AI algorithm showed promising results in this regard. In a further collaboration with the UKSH, the potential of the algorithm to assist radiologists in stroke diagnosis is now being more closely examined.
The results and posters of the project can be found on the KI-SIGS project page.