Research On The Liver Modeling And Deformation Simulation On Mobile Platform

The technology of the virtual surgery has been widely used in the field of educationand operation training. With the development of miniaturization of the medical equipmentand the popularization of the smart mobile platform, there is a broad prospect for a virtualsurgery system applied to the mobile platform. The key technology of the virtual surgery isthe precise modeling for organs and the accurate simulation for flexible deformation. Theorgan segmentation of CT images is still a hot topic in the field of medical image analysis.Moreover, due to the limitation of the mobile platform computing and the inefficiency ofthe drawing and rendering, there are a lot of urgent problems existing in the lightweightcollision detection and flexible deformation simulation.Based on the analysis of the key technology of the virtual surgery, this thesis mainlyfocuses on the segmentation modeling method of the liver, the collision detection methodapplied to a mobile platform, and the flexible deformation simulation of organs. A liversegmentation method based on a3D deformation model is proposed. The image gradient,Gabor and balloon force are utilized in the calculation procedure of the external force of thedeformation model. The liver segmentation is realized by iterating the initial model towardthe boundaries of the object. The collision detection between the virtual surgery instrumentand the liver is achieved by using the ray-casting method and the optimization of thebounding box tree. The flexible deformation simulation of organs are implemented with amass-spring model. An interactive feedback system of the liver and the virtual surgeryinstruments is developed and applied to a mobile platform.(1) The segmentation modeling method of the liver is proposed. The internal andexternal force of a deformation model is constructed. The external force is built by theimage gradient, Gabor and balloon force. And the internal force is built by the smoothnessof organs. The liver edge obtained from the binary image is used as the initial contour ofthe model. The deformation model is forced to cover the lever boundary to achieve thesegmentation and modeling.(2) A collision detection method applied to a mobile platform is proposed. The motionmodels of the virtual instruments are established and the parametric description of the virtual motion is defined. The contact points of the virtual surgery instruments aresimplified as a single virtual particle. The collision detection between the virtual particleand the model is realized by using a ray-casting method. Meanwhile, the collision detectionand structure determination between arbitrary surface models are achieved by theoptimization of the bounding box tree.(3) The deformation simulation of livers applied to a platform is developed. Based onthe mass-spring model, the deformation field of the liver is built and further optimized toreduce the number of elements for deformation calculation and improve the efficiency ofthe deformation model. The3D liver rendering applied to a mobile platform, collisiondetection and flexible deformation simulation are all accomplished. The softwareframework of deformation interaction is designed on the mobile platform, which is thenintegrated with a tactile feedback hardware system by Bluetooth technology to construct apractice interaction system.