Fuzzy classification: towards evaluating performance on a surgical simulator.

Abstract

Computer-based surgical simulators such as the MIST-VR are able to provide scoring metrics such as time taken to complete a task, number of errors made, and economy of movement. Using MIST-VR's basic metrics, we explored the possibility of classifying skill levels using fuzzy logic. Our objective was to create a fuzzy classifier capable of classifying the performance of a subject training on a surgical simulator into 1 of 3 categories: Novice, Intermediate, and Expert. To accomplish this, we needed to establish a baseline skill level for each category. We had four laparoscopic surgeons, four surgical assistants/residents and four non-surgical staff/students with no laparoscopic experience perform two basic tasks on the simulator involving the placement of a ball into a box. We have found, through this preliminary study, that the results were inconclusive. We suspected a number of issues such as the size of our sample space used to train our classifier, and the difficulty of the chosen tasks adversely affected our results.