Parameter Automatic Optimization Of Support Vector Machine Soft Tissue Mechanical Model And Its Application In Virtual Surgical Simulation System

The virtual surgery simulation system can be used for surgeon training,and the existing virtual surgery simulation can present the surgical process well visually,but the force feedback during the virtual surgery operation is not very realistic yet due to the complex soft tissue mechanics model of human body.Therefore,this paper conducts research on soft tissue mechanics modeling,and applies the established soft tissue biomechanical model to the virtual surgery simulation system to obtain accurate force feedback during the virtual surgery operation.Firstly,this paper investigates the mechanical properties of soft tissue deformation process,and builds a biological soft tissue mechanics data collection platform.Pig liver is selected as the experimental object to collect the data required for mechanical modeling,and the main factors affecting stress are analyzed according to correlation.The soft tissue exhibited different mechanical properties during stress loading and relaxation during compression by surgical instruments,so the soft tissue mechanics modeling was subsequently divided into two stages: stress loading and relaxation.A biological soft tissue mechanics data acquisition platform was built with a robotic arm and a force sensor to control the robotic arm to perform compression experiments on soft tissues with cylindrical pins of different areas and at different speeds,and record the data for subsequent soft tissue biomechanical modeling.Secondly,the soft tissue compression process of the surgical instrument was divided into two stages: stress loading and relaxation,and the support vector machine model was established separately,and the kernel function parameters g and penalty coefficient C of the support vector machine were automatically searched for using the particle swarm algorithm,and the searched parameters were used to construct the support vector machine to predict the soft tissue interaction force.The experimental results show that the rootmean-square error of the predicted interaction force is 1.5737 N.In the stress relaxation phase,the root-mean-square error of the model prediction is 0.37857 N.Finally,the established support vector machine model is applied to the virtual surgery simulation platform in this paper.Firstly,the geometric models of surgical instruments and human liver tissues are established.The point cloud-based 3-dimensional model of human liver tissue is established based on the computed tomography data of the real human liver,and the model is mapped and rendered using the Open GL function library.Secondly,the support vector machine based biomechanical model is applied to the feedback force calculation of the interaction between the surgical instrument and the liver in the virtual surgery system.The interaction force between the surgical instrument and the tissue is calculated based on the cross-sectional area of the surgical instrument,the soft tissue deformation and the interaction speed,and the interaction force is fed back to the operator through the haptic manipulator.The actual experiment proves that the virtual surgery simulation system can make the operator feel the interaction force between surgical instruments and soft tissues more accurately,and the model calculation can meet the realtime requirements of the virtual surgery simulation system with smooth operation and small force feedback error.