Using fuzzy set theory to objectively evaluate performance on minimally invasive surgical simulators

Objective surgical performance evaluation is a non-linear and ambiguous problem and hard to model with classic mathematical methods. This thesis explores employing fuzzy set theory as a novel approach to this problem, since the main strength of fuzzy logic is its ability to handle the vagueness and non-linearity of the everyday experiences. Using a commercial surgical simulator, data were collected from subjects who participated in user study of two surgical procedures. Half of these data were used to design four fuzzy models for surgical skills classification. The remaining data were used to test the constructed models and to investigate the effects of various fuzzy inference properties on their performances. Our results indicate satisfactory correlation between the surgical skill levels predicted by the fuzzy models and the actual skill levels of the user. Thus, fuzzy classifiers can be considered as effective tools to handle the fuzziness of objective performance evaluation.