Analyse von intraoperativen Sensordatenströmen zur Unterstützung minimalinvasiver chirurgischer Eingriffe
Minimally invasive mitral valve surgery offers many benefits for patients - including reduced pain, shorter hospital stays and faster recovery. Since physicians only see the heart valve during surgery via ultrasound and camera images, support in merging and interpreting the information from the various data streams is particularly important here. The aim of this project is to integrate ultrasound, vital signs and video data in real time using the Nebula Stream Framework and to automatically recognize surgical phases, surgical instruments, anatomical structures and the spatial correspondence between preoperative and intraoperative image data using resource-efficient multi-task deep learning models and geometric learning. On this basis, precise 3D models and assistance systems can be provided - an essential support for anesthetists and surgeons in navigation, risk prediction and informed decision-making in the OR.
Fellow
Director
Research Junior Group Lead
Doctoral Researcher