Towards Automatic Skill Evaluation in Microsurgery
Shahram Eivazi, Michael Slupina, Wolfgang Fuhl, Hoorieh Afkari, Ahmad Hafez, Enkelejda Kasneci · 2017
In the past decade, eye tracking has emerged as a promising answer to the increasing needs of understanding surgical expertise. The implicit desire is to design an intelligent user interface (IUI) to monitor and assess the competency of surgical trainees. In this paper, for the first time in microsurgery, we explore the potential for a surgical automatic skill assessment through a combination of machine learning techniques, computational modeling, and eye tracking. We present primary findings from a random forest classification method where we achieved about 70% recognition rate for the detection of expert and novice group. This leads us to a conclusion that prediction of the micro-surgeon performance is possible, can be automated, and that the eye movement data carry important information about the skills of micro-surgeons.