Research On Deformation And Cutting Of Soft Tissue In Virtual Surgery

In the traditional surgery training, trainees usually operate on human corpse, animal andrubber human body model, where many constrains and problems lie, such as limited corpsesources, expensive cost, big difference from the real human soft tissue and moral issues likeanimals persecution. Meanwhile, the rapid development of computer technologies makesnumerical simulation computation the powerful method to solve the problem. Manyresearchers began to explore the possibility to apply the virtual reality techniques into surgicaltraining, by establishing integrated visual display, haptic reproduction, hearing and touchingetc multi-awareness virtual surgery environment, the surgery trainees can be given a veryvivid surgery scene to realize surgery training classes, preoperative planning, intraoperativeancillary support, post-operation analysis, repeated exercise of complex operation. Thecombination between virtual reality techs and medical field not only provides vast space forhigh tech medical care, but also brings new and big challenge to virtual reality field.During the real process, visual and haptic feedback information is necessary to conductthe operations. Thus, most virtual reality system is focused on visual and haptic simulation.Through visual feedback, a direct and realistic surgery scene is created; through hapticfeedback, virtual object deformation can be felt. Starting from these two aspects, this thesisselects the typical as well as difficult human soft tissue as the research object and conducts anin-depth research on two key techniques: virtual object cutting and simulation. The main workis as follows:Firstly, geometric models of liver and stomach with anatomical structure are3Dreconstructed from real human slice images, where the combination of manual and automaticslice images segmentation is proposed specific to the disadvantages of a large amount of workin manual way and low accuracy in automatic way. Then these models are further improvedwith visual rendering and texture mapping to get a result that is quite similar to real ones.Force feedback techniques and their computation and measurement methods are also studied,where PVDF (Polyvinylidene Fluoride) piezoelectric thin film sensor based force measuringmethod is put forward, which are then applied in soft tissue deformation and cuttingsimulations.Secondly, a new Alterable and Local Mass-Spring/Damper Model (ALMSDM) model isproposed aiming at the short-comings of palpation operation and existing deformationsimulation methods. This model possesses features like changeable position and extendablearea, which improves local modeling area limitation and poor recovery ability of global surface model. From the real-time point of view, according to the application of ALMSDMmodel in surface model and volume model, the idea of vertex normal vector updating in localand the strategy of pre-computation are put forward respectively, making a big improvementin real-time characteristic of the system. The performance of system is assessed from recovery,feedback force and real-time under different models，which verifies feasibility and universalproperty of this proposed algorithm.Thirdly, meshless theory is innovatively applied into virtual reality deformationsimulation. On the basis of deep research in particle integral meshless Galerkin method,according to the large deformation feature of virtual soft tissue, an improved particleintegration element free Galerkin method in proposed, which is verified by2D beam theory.Then on the basis of this improved algorithm static and dynamic models of virtual soft tissuedeformation simulation are constructed. The static model has achieved the simulation ofdynamic deformation process using cumulative static displacement, with a high speedcompared to dynamic model, but it has low accuracy without the inner continuous changingprocess, while the dynamic model has a large amount computation. Comprehensivelyconsidering the features of dynamic model and the requirement of real-time in virtual surgery,these models are approximately solved. In simulation and compare, the meshless methodperforms well both in stability and reliability, and avoids the problems such as twisting andlocking in mesh method, and also overcomes the big errors during the large deformation. Atlast, the practical feedback force is tested using PVDF, and is compared with the datacomputed from the model.Finally, the cutting algorithm in virtual surgery is studied. The scheme named “incisionboundary drawn independently” is proposed, which doesn’t dependent topologyreconstructing after the meshes breaking. It combines incision characteristics and thesuperiority of the spline method. Thus, the innovative Bézier curve cutting method isproposed. By adding spring constrain between curve boundary points and the curve itself, thismodel can control the open of incision flexibly. The simulated experiment of removingforeign object after surgery cutting is designed. And the fraction during the cutting operationis measured.