Research Of Abdominal Aorta Segmentation Based On Arterial Phase CT Image

With the development of the medical imaging technology, a variety of imaging equipments such as CT, MRI, PET, etc. have been widely used for medical diagnosis, and provide a number of human anatomical and functional information. Medical imaging diagnosis has become one of the most important skills of today’s medical diagnostic. When the surgeons schedule abdominal operations, they need an application which is able of analysis, segmentation, reconstruction and 3-D visualization of the datasets.According to these needs, this paper studied the abdominal aorta segmentation methods based on arterial phase CT image, and introduced some classic vessel segmentation methods and basic medical image processing skills used in vessel segmentation. We proposed an improved abdominal aorta segmentation method based on 3-D region growing, and built a platform based on Microsoft Visual Studio MFC framework. We tested the algorithm on this platform, and the final results were quite satisfactory.The purpose of this paper is to build an application which is capable to do the segmentation and visualization of abdominal aorta, and eventually provide an accurate visualization of the entire abdomen for clinical surgery planning. In this paper, we presented an improved abdominal aorta segmentation method based on 3-D region growing. First, we did some image preprocessing, then we used two global thresholds to extract the initial vessels and bones; then a 3-D region growing was applied for aorta and adjacent large vessels extraction. The thinner blood vessels are extracted by counting the number (HAI) of the pixels in every component of each slice. HAI is used for crude extraction of the thinner blood vessels. The thinner blood vessels along the axial is continuous, so we used the continuity of the blood vessels to determine the final thinner vessels; Finally, the larger blood vessels is simply add to the thinner blood vessels to get the final result of the abdominal aorta. We also introduced the manual modification to ensure the accuracy of the final segmentation. The final results were evaluated by clinical surgeon and proved to be reliable. Compared with the visualization of professional platform, the result verified the accuracy of the algorithm.