Soft Tissue Modeling Method In Virtual Surgery Based On CT Image

As an emerging research direction, Virtual surgery is a crossing research field integratedwith medicine, biomechanics, optical engineering, materials science, computer technology,mechanics and other many other disciplines. It is an important achievement of thecombination of virtual reality technology and medical research, and also is one of the hotissues in the field of information technology. With the improvement of modern medicine, thesurgery is developing from traditional experience surgery to the precise surgery, and from thelarge area injured surgery to minimally invasive surgery. Then it is put forward that a new andhigher requirements to the interaction performance of virtual surgery system and the user’simmersion in sight and touch. In the meanwhile, the rapid development of computer softwareand hardware technology provides good support to the hardware platform for virtual surgerysystem. Therefore, to the soft envirorment, the soft tissue model as the operational object ofvirtual surgery system determines its performance especially in the user’s sense of vision andtouch, and becomes an important research topic in the virtual surgical field.As a high resolution optical image, the CT image is a veracious and digital expression ofthe structure and density of the soft tissue, and also is one of the most important image data oftoday’s medicine. In this paper，it is the object that improving the user’s sense of vision andtouch. According to the biological characters of the soft tissue, a study of human soft tissuemodeling is developed, which foucs on the key problems of the efficiency, reality andaccuracy of the virtual surgery system. The main content includes modeling of human softtissue surface meshing, optimization of the surface particle discretization and physicalmodeling based on CT images. According to the stress-strain relations of the tissue, thephysical model and mechanical model of the soft tissue are established separately. In addition,the model parameters are settled by the measurement of real soft tissue, and the accuracy andreality of the model are increased, so as to improve the sense of vision and touch for the user.The main work and innovations are as follows:(1) The major human soft tissue modeling methods are studied in depth, and the finite element method and the mass-spring system method are elaborated, and both advantage anddisadvantage of the two methods are analysed.(2) The physical model parameters of the soft tissue are determined by measuring thebiomechanical properties of real soft tissue. The biomechanics axial tensile and axialcompression experiment of real pig liver tissue is carried, and the stress-strain relationship ofsoft tissue is gained. In addition, the viscoelastic phenomenon of soft tissue during theexperiment is analyzed and discussed.(3) A three-dimensional surface reconstruction algorithm based CT images of soft tissueis put forward. The data source is the most common CT images, and the surface contourdiscrete algorithm based on the adaptive step is used to analyze the image layer by layer toextract the soft tissue contour feature points; then Delaunay triangulation principle is used toreconstruct three-dimensional soft tissue surface. According to the curvature of the surfacediscrete points, the three-dimensional reconstruction mesh is optimized to display thethree-dimensional features of soft tissue with smallest number of grid.(4) A new soft tissue discretization based on the surface triangulation is proposed. Themethod not only takes the distribution of the surface points into account but also combineswith the normal vector projection, so that the characteristics of the inner body of the modelcomponents is better reflected by the surface vertices. According to the model geometry, itsmass distribution is adjusted adaptively, and the overall distribution of the model features isreflected more objectively. Finally, the model comparison test was carried with three differentgeometric shapes. Compared with radiation-based body discretization method, a betterdiscretization effect is gotten by the proposed method in the same time complexity.(5) A new nonlinear physics soft tissue modeling method is proposed. Based on theclassical mass-spring system model, the concept of "undeformable shape" is put forward toreflect the volume characteristics of the model, and a physical model is established by the thenonlinear spring force function reflecting the soft tissue viscoelastic characteristics. Theaccuracy and reality of the model deformation is improved by this method.