Evaluating Surgical Performance in Real Time Using Data Mining

Yun Zhou,I Ioannou,James Bailey,George J. Kennedy,Stephen O’Leary

2013-01-01

Abstract:Virtual reality simulators are becoming increasingly popular as adjuncts to traditional surgical training methods, but most simulators do not have the ability to evaluate performance on-the-fly and provide advice to trainees as they practice. Timely feedback on performance is a critical component of surgical training, therefore the ability to provide such evaluation is necessary if simulation is to be effective as a platform for self-guided learning. We propose an evaluation framework to automatically assess performance within a temporal bone simulator in real time. This evaluation framework uses data mining techniques to assess performance at different granularities. Drilling technique is analysed to deliver detailed short-term evaluation, while hidden markov models are used to evaluate the completion of small surgical subtasks and provide medium-term assessment. Finally, an analysis of drilled bone shape is used to evaluate performance at the completion of each stage of a surgical procedure. We demonstrate the effectiveness of the proposed methods by validating them on an existing simulation dataset.