Modern assistance systems in the operating room should not only make individual steps easier, but also understand the course of an operation in its entirety. The automatic recognition of surgical phases aims to identify the current progress of an operation. Such systems provide contextual knowledge about the current phase of the procedure and thus enable intelligent assistance functions, such as the targeted provision of instruments or automated surgical documentation.
At the IMT, we are researching how artificial intelligence methods can be applied to video data from real operations in order to develop models that reliably and robustly classify surgical procedures - even with limited data. The aim is to use such systems to increase safety in the OR, reduce the cognitive load on the OR team and create the basis for automated assistance in the future.
