Abstract:

The Fundamentals of Laparoscopic Surgery (FLS) box trainer is a standardized simulator for training and assessing basic laparoscopic skills. This work is comprised from two studies that provide new ways to assess the performance of the practitioner using computer vision and statistical analysis.

The aim of the first study is to provide automated feedback on the intracorporeal suture exercise in the FLS simulator using deep learning algorithms. Different metrics were developed with the goal of providing informative feedback to the user. The automation of the feedback will allow students to practice at any time without the supervision of experts.  Object detection and semantic segmentation were used to collect statistics on the practitioner’s performance. Three task specific metrics were defined. Good agreement between the human labeling and the different algorithms was achieved.

The aim of the second study is to develop a system that measure how much force is applied on the graspers while performing the peg transfer exercise in the FLS simulator. We installed strain gages on the grasper that allows us to measure the forces the practitioner applies on the grasper. We measured the strains of residents and senior surgeons performing the exercise. Statistical analysis was performed on the results. The analysis shows significant differences of the way forces are applied on different materials, during transferring the triangles of the exercise from one hand to another and after repeating the exercise.