基于CT图像的血管分割算法分析-计算机应用技术专业论文.docxVIP

哈尔滨工业大学工学硕士学位论文 哈尔滨工业大学工学硕士学位论文 - - II - Abstract Vascular diseases are important public health problems all over the world countries. In particular, an accurate description of blood vessel structures in 3 -D images is very important in many clinical applications, for example, disease diagnosis, surgical planning and monitoring of disease progress or remission, which reduces physical damage and alleviates the sufferings of patients , researchers have proposed a lot of blood vessels segmentation algorithm based on CT images. As the characteristics of CT images and complexity of blood vessels, blood vessels segmentation algorithms have great difficulty. In order to develop vascular segmentation algorithms of CT images that are easy and accurate to apply in engineering, I have done the following research in two aspects: (1) Select the appropriate blood vessel segmentation technology to segment the CT images accurately. (2) Select the right data structure and design algorithms to solve problem on space and time constraints in the case of a large amounts of data and better human-computer interaction to meet the clinical application. Analyzing the current vessel segmentation technology which is the application of engineering techniques including classical active contour models, region growing introducing the 4D Curves based on active contour and the competitive region growing. These two techniques have good features of good noise immunity by the segmenting the leg vascular based on statistical thinking. For 4D Curves algorithm, people need to give initial conditions on the approximate radius of the vessels. I propose an automatic algorithm to determine the radius and greatly increased the availability of this model in the practical application. For regional growth based fast marching algorithm, the initialization is needed to give the threshold, which needs prior experience, and meanwhile different data has different thresholds, in such cases, I proposed an automatic sele