ABSTRACT
Our research team is working on the virtual surgery system. 3D geometric modeling of the human tissue or organ is the core module of the virtual surgery system. 3D geometric modeling can be employed to aid doctors in diagnosing patients accurately and efficiently so as to improve the success rate of surgery. In this paper, the implementation method of the geometric modeling software in the virtual surgery system is mainly introduced. 3D geometric modeling of human tissue consists of the preprocessing and three-dimensional reconstruction of medical images; the core operations of the preprocessing of images is the segmentation of medical images. Ramanujam Kayalvizh et al. 2010, proposed a new intelligent arithmetic, that is, the Particle Swarm Optimization (PSO). I. Cruz-Aceves et al. 2013, put forward an automatic image segmentation method based on active contour model theory and estimation of distribution algorithms. Nihar Ranjan Nayak et al. 2013, proposed an improved clustering algorithm which is applicable to gray images by evaluating three different kinds of clustering algorithms. For general medical images, the accuracy of computer automatic segmentation methods is difficult to meet the requirements. Therefore, the interactive operation in the image segmentation is required so as to obtain more accurate segmentation results. Mariofanna, Milanova et al. 2010, proposed an image segmentation method based on vision to determine its location in this paper, which applied the ChanVess model to the image segmentation. Liu Zaitao et al. 2011, put forward a new method of the complex medical image segmentation according to the
segmentation problem, but also to the type of medical data to be segmented. Growth criteria may be based on different principles. The difference in the growth criteria has an effect on the process of the region growing and segmentation results. Therefore, most of the growth criteria are based on the local characteristics of images. In this paper, two-dimensional medical images are regarded as the research object. Consequently, the growth criterion selected is based on the regional gray level difference. The flow of this algorithm is: a typical point in some region in the image is selected as the seed point according to the purpose of segmentation. After initializing the stack of computer memory, we will store the specified seed point into the stack. Afterwards, we take out a member of the top from the stack and scan its eight neighbourhoods. When there is a gray value of the pixel to meet the requirement, |gn − gc| < dv and |gn − gs| < cv, we store the value into the stack and repeat the cycle until there are no pixels that meet the condition. Ultimately, the process comes to an end. The flow diagram of the region growing algorithm is shown in Figure 1.
