摘要
从两幅不同角度的造影图像实现血管树的三维骨架重建,传统重建方法常常需要较多的人工干预,才能为每一点找到准确的对应点.文中研究了一种基于Snake模型的重建方法.它在采用多尺度Gabor滤波提取造影血管中轴线的基础上,选取血管树分叉点,优化几何变换矩阵,提高用以初始化Snake的对应点的准确性;然后,针对血管的特性,文中采用GVF(Gradient Vector Flow)流量场作为三维Snake的外部能量场,并给出最小化血管Snake模型能量函数的表达式,使Snake保持自身平滑连续的同时在空间中发生形变以逼近真实轮廓.实验结果表明,该方法与传统的重建方法相比,不但减少了人工干预,而且有效地提高了重建精度.
For the 3D reconstruction of blood vessel＇s skeleton from biplane angiography system,the traditional way requires lots of user interaction to find corresponding points.A method based on Snake model is proposed in this paper.First,multiscale Gabor filter is used to extract the centerlines of blood vessels on the angiography images.Secondly,in order to find corresponding points correctly which are used to initialize Snake,bifurcations on blood vessel＇s skeleton are chosen to optimize GT（Geometry Translation）matrix.Thirdly,according to the feature of the blood vessels,GVF（Gradient Vector Flow）is utilized as external energy of 3D Snake and the expression of minimizing the external energy is deduced.At last,the Snake iteratively evolve towards the true configuration of the blood vessels,satisfying the constraint of being smooth.The result of experiments indicates：Compared with traditional way,user interaction is reduced and reconstruction performance is improved by this method.
出处
《计算机学报》
EI
CSCD
北大核心
2010年第3期603-612,共10页
Chinese Journal of Computers
基金
国家"九七三"重点基础研究发展规划项目基金(2003CB716105)资助~~
关键词
血管造影
三维重建
SNAKE模型
GVF
angiography
3D reconstruction
Snake model
GVF（Gradient Vector Flow）
