The development of the data correction method and feature extraction for the skill evaluation of practical laparoscopic surgical training

Abstract

This paper describes the details of the data correction method and feature extraction for skill evaluation of practical laparoscopic surgical training. To analyze the surgical dexterity in practical laparoscopic surgical training, the motion capture (MoCap) based measurement system have been developed, and measurement experiment in cadaver surgical tranining were conducted. However, the measured data contains noise and missing values, skill analysis is not possible without data correction. For this background, outlier removal method using Kalman filter and other threshold-based method were applied. Linear interpolation and Savitzky-Golay filter were also applied to the data. After the data correction, feature extraction was performed and some kinematic indices of the surgical instruments were successfully calculated.