Low-level Event Detection System for Minimally-Invasive Surgery Training

We present an event detection system in a laparoscopic surgery domain, as part of a more ambitious supervision by observation project. The system, which only requires the incorporation of two cameras in a laparoscopic training box, integrates several computer vision and machine learning techniques to detect the states and movements of the elements involved in the exercise. We compare the states detected by the system with the hand-labelled ground truth, using an exercise of the domain as example. We show that the system is able to detect the events accurately.