Automated assessment of laparoscopic pattern cutting skills using computer vision and deep learning.

Area of Science:

• Surgical Education
• Medical Simulation
• Computer Vision in Medicine

Background:

• Current pattern cutting assessment in Fundamentals of Laparoscopic Surgery (FLS) is manual, leading to time-consuming evaluations prone to errors.
• Objective and automated assessment systems are needed to improve the efficiency, reliability, and standardization of surgical skills evaluation.

Purpose of the Study:

• To develop and evaluate a machine learning-enhanced computer vision system for automated pattern cutting assessment in surgical training.

Main Methods:

• A computer vision system was developed using a You Only Look Once (YOLO)-based deep learning model for specimen segmentation.
• The system analyzes digital images of cut specimens, comparing them against circular targets and measuring area-based and radial deviations.
• Performance was validated using synthetic test samples and real surgical specimens.

Main Results:

• The segmentation model achieved 98.32% accuracy on synthetic samples with a mean absolute error of 34.7 mm².
• The system accurately measured deviations in real surgical specimens, with results closely matching expert evaluations.
• Quantitative feedback is provided via color-coded visualizations and radial deviation analysis.

Conclusions:

• The automated system offers objective, quantitative evaluation for pattern cutting skills, suitable for standardized implementation in FLS testing.
• Despite challenges with specimen distortion, the system's speed and reduced reliance on human assessment make it a valuable tool for surgical skills evaluation and training.