Segmentation And 3D Reconstruction Of CT Lung Vessel

In modern society, people’s living standards continue to improve, vascular disease’s incidence rate is rising all the time due to the unreasonable diet and lack of exercise, However, the doctor is still using the traditional way(reading the CT slice by slice) to diagnose vascular diseases, which relies heavily on the personal knowledge as well as experience. In this paper, we focus on the segmentation of lung CT sequence, after that, we use three-dimensional technology to reconstruct the result, which provide a powerful tool for the diagnosis and treatment of lung disease.First, this paper studies the principle of X light and CT imaging as well as the specific content of the standard. The first step in pulmonary vascular segmentation is to segment the lung parenchyma, In order to accomplish the segmentation of lung parenchyma, this paper studies the common methods of image segmentation, and finally uses the method of region growing, the Confidence Connected RG and morphological operation to finish the segmentation of lung parenchyma.Then, this paper studies the commonly used algorithm for blood vessel segmentation, which mainly includes two categories, one is the traditional Hessian matrix based segmentation algorithm, and the other is the VED algorithm, this kind of algorithm use the isotropic diffusion in scale selection process. The disadvantage is that while removing the noise it also blurs the image characteristic. After analysis and comparison of the two algorithms, we put forward the corresponding solutions, the global gray selection function is added and the filtering equation is modified to make it more in line with the clinical needs, Then, through composite image enhancement, we solve the problem of fuzzy distortion after image filtering to a certain extent. In order to verify the correctness of the improved method in this paper, we makes a comparative analysis about the performance of several algorithms based on the data set of expert segmentation, we finally verifies the effectiveness of the proposed method.Finally, in order to facilitate the presentation of the segmentation results, this paper studies the 3D reconstruction technique, which includes surface rendering and volume rendering. The surface rendering technology is analyzed with the moving cube algorithm as an example. The volume rendering technique is analyzed with ray casting algorithm as an example. In addition, we compared the advantages and disadvantages of two kinds of visualization algorithms.Finally, we complete a practical vascular segmentation and visualization software, which consists of several vascular segmentation method as well as three-dimensional reconstruction technology with the help of ITK, VTK, CTK, QT and other popular open source tool set. With callback function and GPU volume rendering, we can achieve nearly real-time rendering, this software has a certain clinical practical value.