Surgical Skill Assessment System Using Fuzzy Logic in a Multi-Class Detection of Laparoscopic Box-Trainer Instruments

Abstract

The development of automated systems to evaluate the trainees’ laparoscopic surgery skills is a challenging problem. Box trainers use actual laparoscopic instruments and a video system that allows the monitoring of the execution of the test procedures. In this paper, we propose a fuzzy logic-based performance assessment system that takes inputs from a multi-class laparoscopic instrument detection system and a measurement algorithm. The laparoscopic instruments are tracked during the test by a deep learning algorithm using the SSD-MobileNet architecture. Based on the experimental result, the trained model could identify each instrument at a score of 70% fidelity while providing input information for the performance assessment system. The experimental results demonstrate that our proposed approach is a promising component for the development of a comprehensive laparoscopic surgery skill assessment system. This project is a collaborative research effort between the Department of Electrical and Computer Engineering and the Department of Surgery, of the Homer Stryker M.D. School of Medicine, at Western Michigan University.