Research On Three-dimensional Reconstruction Of The Blood Vessels In Medical Image And Its Application

Since the 21 st century, many medical experts’ researches have shown that cerebrovascular disease is one of the most harmful diseases to human health. With the rapid development of neuroimaging and computer technology, by using computer-aided diagnosis and treatment systems can effectively assist health workers to work better for patients. It can help doctors offer better treatment options. The techniques of reconstruction of vessels in medical image is very important. Whether for cerebrovascular disease diagnosis or surgical planning, it has important significance and practical value.In this thesis, many researches of three-dimensional reconstruction techniques of vessels in medical images have been done. All researches can be divided into three aspects. The first aspect is to study how to acquire information from Medical images. Then, the second aspect is about vascular segmentation techniques. Three-dimensional reconstruction is the last aspect. Through the study of these three aspects, the following work have been completed.Firstly, the medical image DICOM3.0 standard has been learnt, and DICOM image file structure has been analyzed. By manual analysis and program implementation, we have extracted information of the medical image files of DICOM format. The information contains the basic information and image information. These provide the data base of the whole researches.Secondly, a comprehensive overview of image segmentation algorithm has been done. After that an improved algorithm of medical image segmentation has been proposed. By introducing the fuzzy clustering algorithm to the original LBF algorithm based on active contour model, a fully automatic active contour model was put forward. Thesis proposed a new termination condition by using the absolute evolution of the pixel difference. Relying on the fuzzy clustering algorithm to initialize the level set function and the termination condition, we realize automatic image segmentation without manually setting the initial conditions and the end number of iterations. The experiment results show that the proposed algorithm is better than the traditional, such as CV, LCV, LBF.Then, we carry on the research on 3d reconstruction technology, mainly on 3d visualization technology, which including surface rendering technology and volume rendering technology. After achieving many different rendering technologies, we have contrasted and summarized the advantages and disadvantages of these technologies. It shows that ray casting is more effective in 3d medical image reconstruction.At last, by integrating these three aspects researches, we have developed a prototype system by using QT and VTK. Through the developing of the application system, we have verified the validity of the improved algorithm. And the system can realize the three-dimensional reconstruction drawing of blood vessels. What’s more important, this is the goal of our study and the goal have been achieved.