CAJ | 학술논문

探究基于磁敏感加权成像（SWI）的脑静脉中心线提取技术，在利用图像海森（Hessian）矩阵的几何特性提取出三维脑静脉的基础上，对得到的二值静脉血管图像进行快速三维欧几里德距离变换，得到静脉血管的距离图，再计算距离图中每一个体素的Hessian矩阵，利用Hessian矩阵的特征值和特征向量提取出静脉血管的中心点，并得到相应的血管半径大小。最后用Dijkstra最短路径连接法将中心点连接成线，便得到了静脉血管的中心线。
탐구기우자민감가권성상（SWI）적뇌정맥중심선제취기술，재이용도상해삼（Hessian）구진적궤하특성제취출삼유뇌정맥적기출상，대득도적이치정맥혈관도상진행쾌속삼유구궤리덕거리변환，득도정맥혈관적거리도，재계산거리도중매일개체소적Hessian구진，이용Hessian구진적특정치화특정향량제취출정맥혈관적중심점，병득도상응적혈관반경대소。최후용Dijkstra최단로경련접법장중심점련접성선，편득도료정맥혈관적중심선。
This paper studies the centerline extraction technique of the cerebral vein based on the susceptibility weighted imaging(SWI).After the vein was extracted using the geometry characteristic of Hessian matrix,the fast computation of the three dimensional Euclidean distance transform was used to get the distance image of the vein.The Hessian matrix of every voxel was calculated in the distance image. Then the eigenvalues and eigenvectors of the Hessian matrix were made to extract the center points of the cerebral vein, and the radius size of the corresponding vein as well.Last we used Dijkstra algorithm to connect the center points into the centerlines.