Evaluation And Prediction Of Virtual Surgical Skill Level Based On Machine Learning

The virtual reality surgical simulator is an important application of virtual reality technology in the medical field,which makes up for the shortcomings of traditional surgical training,including insufficient working hours of residents,insufficient training plans,and long training cycles.Surgical simulators provide physicians with a safe training environment,training feedback,and evaluation mechanisms while providing a standardized experience.With the rapid development of virtual reality surgical simulators,new assessment and training methods are being developed,such as incorporating machine learning algorithms to analyze and evaluate the performance of participants.However,machine learning models often have difficulty explaining the inner mechanisms of assessments,which makes it impossible to personalize training for participants.There are differences in the initial skill level and learning efficiency of doctors,and traditional training methods are difficult to ensure that trainees meet the training requirements.Therefore,this paper studies the potential of machine learning as a predictive tool for neurosurgery skill level,explains the internal mechanism of the machine learning model to predict surgical skill level and predicts the subsequent skill level based on the trainer’s previous training results to save training resources and ensure training effectiveness.The main work of this paper is as follows:(1)The study designed a set of virtual reality neurosurgery simulators,combining hardware and software for tactile and visual interaction during the operation.Soft tissue modeling in surgical scenarios using the Extended Position Dynamics algorithm.According to the brain tumor resection operation process in the real operating room,three surgical scenarios of skull drilling,meningeal resection,and tumor resection are designed,providing doctors with a fully immersive neurosurgery simulation in a virtual environment.(2)It is proposed to use machine learning to classify the doctor’s surgical skill level,and understand the doctor’s operation by explaining the machine learning model.This method first uses a virtual reality neurosurgery simulator to collect the trainer’s operation data and obtains the corresponding indicators through preprocessing,uses five commonly used machine learning models for training,and obtains the support vector machine model.The classification effect is the best,and it is proposed to use the Shapley value to explain the machine learning model,understand the shortage of trainers in brain tumor resection,and adopt a personalized training plan.(3)Proposed the use of machine learning to predict the characteristics of the learning curve in surgical training.By collecting the operation data during the doctor training process,using the confidence score in machine learning to obtain skill scores,defining the characteristics of the learning curve,and finally using a variety of machine learning methods to predict the characteristics of the learning curve,the results show that the SVM regression model is in the learning curve The prediction performance is the best among the features.This approach enables the development of individualized training programs,increasing efficiency and reducing costs.