Understanding expertise in surgical gesture by means of Hidden Markov Models

Minimally invasive surgery (MIS) has became very widespread in the last ten years. Due to the difficulties encountered by the surgeons to learn and manage this technique, a huge importance has the improvement of training procedures, the improvement of surgical instrumentation and the robotic automation of surgical gesture. All these purposes require the analysis of surgical performance with the aim to understand it and to define what is expertise in surgical gesture. In this paper for the first time the Hidden Markov Models (HMMs) are used as a tool for the understanding of surgical performance and of the human factors that characterize it. In our experiments we used position data concerning the tools movements during exercises performed on a surgical simulator. Using Hidden Markov theory, we create a model of the expert surgeon performance able to evaluate surgical capability and to distinguish between expert and non-expert surgeons. By analyzing the trained model of the expert surgeon performance we show that it is possible to deduce information about features characterizing the surgical expertise.