Neurosurgery Virtual Surgical Training System, Brain Model Three-dimensional Reconstruction

Neurosurgery virtual surgery system is an important application of virtual reality technology in neurosurgery. It can be used in surgical planning, training and further actual surgery, and promote the propagation of advanced surgical methods. Medical image 3D reconstruction is a crucial technology in virtual surgery system; it reconstructs the medical image data (such as CT or MRI image, etc). 3D model provides a vivid virtual scene in which the operator can make a precise diagnosis and plan a valid operation path. Therefore, Study on medical image 3D reconstruction has important significance on science and worthiness in practical application.At first, this dissertation introduces the medical digital image and communication standard (DICOM) briefly and analyzes the DICOM file format. Then, we introduce the preprocessing technologies including grey-scale transformation, interpolation and noise elimination, and the segmenting technologies including threshold segmentation and the fuzzy C-means clustering (FCM) segmentation. Because the FCM algorithm requires a long time to segment images, this dissertation proposes a modified method. This method includes two aspects. The one is taking cluster centers obtained by the K-means clustering algorithm as the initial cluster centers for the FCM algorithm. The other one is using the data set of gray scale instead of the data set of pixels in the original FCM algorithm. Experiments show that the algorithm can obtain segmentation results efficiently and accurately.Secondly, we discuss 3D reconstruction technologies. 3D reconstruction includes two methods: surface rendering and volume rendering. Because of volume rendering has a large scale of computing, and too depends on hardware, and surgical simulation system requires high real-time, so marching cubes algorithm (MC), one of the surface rendering methods, is used to reconstruct images in this dissertation. This dissertation introduces the principle and implementation of MC in detail, and analyzes the reason of ambiguous meaning of connection in MC, at the same time, implements the asymptote discrimination to eliminate ambiguous, and discuss the simplification of the triangular mesh. Lastly, according to the characteristics of neurosurgery virtual surgery, we analyze the requirement of software. At the same time, the functions of the system software was analyzed and defined in detail. We divide the system into five modules, including data management, preprocessing, segmentation, 3D reconstruction and surgical simulation. And we chose Visual C++ and OpenGL as the developing platform. Now, several modules have be finished, which include data management, preprocessing, segmentation, 3D reconstruction.