Overview

In computer- and robot-assisted surgery, it is extraordinarily challenging for technical systems to transfer general anatomical models to the actual anatomy of individual patients with deformities present and organs covered by fat and connective tissue. Even for human experts, this is only possible through domain knowledge and experience.
The goal of the research project "KIRuRC": AI-based registration and robotic surgery is to autonomously perform surgical tasks robotically based on sensor data and an anatomical model. The challenge here is the continuous registration and deformation of the anatomical model to recorded sensor data by a learning software agent. The surgical task is thus planned interpretably on the anatomical model after it has been adapted by the learning agent to match the real environment. [Faculty News]