Extraction Of Vessel Skeleton Based On Oriented Bounding Box Research And System Implementation

Vessel diseases present a substantial challenge in the medical field,and diagnostic imaging technology can aid in their diagnosis,treatment,and prognosis management.Vessel intervention surgery has emerged as an efficacious approach for diagnosing vessel diseases,with the vessel skeleton line serving as a useful guide.However,the automatic extraction of the vessel skeleton line remains a formidable task due to various impediments,such as calcification,occlusion,imaging artifacts,and inadequate contrast agent.This thesis is centered on the precise extraction of the vessel skeleton line from vessel imaging.It emphasizes the utilization of the M-OBB vessel skeleton extraction algorithm,which adeptly resolves the challenge of segmentation-based vessel skeleton extraction technology’s discontinuity issue.Additionally,this thesis suggests a swift level set-based vessel segmentation technique founded on the extracted vessel skeleton line.Lastly,a vessel skeleton line extraction system is proposed,based on the skeleton extraction and vessel segmentation algorithms.The specific research contents of this study are as follows:1.An automated technique for extracting the vessel skeleton from three-dimensional vessel imaging(CTA)is proposed in this study,which serves as a prerequisite for complete segmentation of the vessel tree.The method employs an oriented bounding box progression approach that generates coordinate points and a skeleton from the input image.The algorithm consists of three steps: first,an oriented bounding box encompassing a vessel segment is computed using two points in the vessel,and threshold segmentation is applied to the small vessel segment image.Next,the exit and middle surfaces of the vessel are estimated to determine the centroids of these two surfaces,and the oriention of the next bounding box is calculated based on the two points.Finally,these coordinate points within the vessel are utilized as nodes to construct the skeleton.2.A novel approach for automatic segmentation of vessel structures based on the horizontal set of the vessel skeleton is proposed.Initially,a three-dimensional vessel model is created from the initial skeleton line,which is then passed through the horizontal set propagation to regulate the growth of the vessel contour.Following the level set evolution,a new skeleton line is extracted and the vessel diameter is re-estimated to generate a new vessel model.This process of propagation and modeling is repeated until convergence.3.A one click processing vascular skeleton extraction and vascular segmentation system has been proposed,and the entire software has good system scalability and portability.