3D Solid Model Reconstruction Based On Tomographic Images

With the wider application of the three-dimensional reconstruction based on tomographic images in the medical field, it is demanded that the three-dimensional reconstruction model not only can be used for three-dimensional visual display, but also edit, modification, measurement and quantitative analysis. This requires that the three-dimensional reconstruction model must be a solid model. Quantitative analysis based on the solid model makes medical research great progress. In this paper, the three-dimensional solid model reconstruction based on Tomographic Images is researched. For tomographic images generated by CT (Computed Tomography) scan, MRI (Magnetic Resonance Imaging) and simulation, a method of multiple images automatic processing and segmentation is proposed. Complex three-dimensional surface model and solid model with holes are reconstructed. These algorithms are achieved by Java,Java3D program. And with these algorithms as the core, the image processing and three-dimensional reconstruction system based on web is built, that is, Medical Image Engineer (MIE).For image processing and segmentation, we first research the method of a single image automatic processing and segmentation. Selecting the best threshold value based on the Otsu criteria, the initial image segmentation is achieved by the threshold method. The theory of mathematical morphology is used to separate the edge of the region tested. The GVF Snake model is used to segment the objective of the image with weak edges, get the initial contour and optimize the initial contour by the principle of energy minimization. For multiple images processing, the information of many adjacent images is used. First, clear images are segmented by using the method of single image segmentation.Then fuzzy images are segmented by using the method of GVF Snake, and the initial contour is the precise contour of clear image segmented. In this paper, multiple images automatic segmentation method can process many images automatically and the edge is clear,continuous and close. This high-quality binary image not only provide a guarantee for reconstructing three-dimensional solid model, but also can be applied to many other aspects, such as target tracking, medical clinical examination and so on.In order to reconstruct complex three-dimensional solid model with holes based on tomographic images, a new method is proposed, which integrates a local optimization approach based on contours and marching cubes. Through using the auxiliary contours method to enhance the accuracy of contours, classing the complex model, choosing reasonable methods, the difficulty of multi-branch contours topological reconstruction is overcame. And a Three-dimensional surface model formed by triangular patches is created. Furthermore, a solid model is constructed from the surface model by ANSYS Parameter Design Language (APDL) in the ANSYS software, which not only can be used for three-dimensional display, but also Finite Element analysis. Besides, it makes the three-dimensional reconstruction method more practical.In order to make other researchers use the outcome easy, the image processing and three-dimensional reconstruction system based on web is built by Java,Java3D program. Image browse, images automatic processing, manual cross-processing, three-dimensional reconstruction and solid model construction are integrated as a whole. After inputting images, you can browse every images easy, process and segment images by the automatic processing method or manual processing method, and reconstruct three-dimensional solid model by the outcome. The system use methods of integrated design, modular design and human design, and run in multifold platforms and network.Finally, a summary of research work and prospective plan for the future work are given in this paper.