Automated Objective Basic Surgical Skills Assessment: Overall Kinematic Performance Assessment Method

Abstract

As an essential part of medical training, assessment surgical skill is a time consuming, subjective, and complicated process. This paper adopted overall kinematic performance assessment method to identify the skill level of a subjective given motion data from three benchtop surgical tasks performed on robotic surgical devices. Firstly, we extracted global movement features by computing from the raw data of 39 (Suturing), 36 (Knot Tying) and 28 (Needle Passing) trials collected on da Vinci surgical system, respectively. Then, the discrimination ability of single feature with optimizing threshold were calculated. In following classification process, we applied support Vector Machine (SVM) to distinguish expert from novice on the basis of selected global movement features. The results showed that global movement features (GMFs) such as task completion time, velocity, and motion smoothness have superior discrimination ability between novice and expert performance for suturing, knot tying and needle passing task. SVM could classify surgeons' expertise as novice or expert with an accuracy of 77.99% for suturing, 83.71% for knot tying and 74.66% for needle passing, respectively. This study clearly demonstrated the ability of overall kinematic performance assessment method to distinguish between novice and expert performance in the performance of robotic surgical devices.