血管横截面参数模型及其在血管重建中的应用

Parametrical Model of Vessel Cross-Section and Its Application in Reconstruction

为了改善传统重建方法在稀疏投影等病态情形下的重建效果 ,分析了描述血管弹性性质的微分方程 ,并利用离散余弦变换的能量压缩特性 ,建立了具有圆形分量、椭圆分量等 10个参数的截面模型 .与圆形、椭圆形截面模型相比 ,该模型能够更逼真地描述血管横截面 .利用该模型还可以对传统重建方法进行后处理 ,从而改善重建效果 .仿真结果表明 :该模型描述各种血管横截面的误差仅为 2 0 4 % ,在病态投影的情况下可以使重建误差减小到 5 %左右 .

In order to improve the reconstruction effect under ill conditions such as sparse data, a novel parametrical model is constructed through analyzing the differential equation of vessel's cross section and utilizing the energy compression of discrete cosine transform. The model has ten parameters such as circle deescriptor, ellipse descriptor and so on. Compared with circle or ellipse model, the parametrical model can not only express cross section more vividly but also be used as post processing to improve the reconstruction effect. Simulative experiments indicate that the model can reduce the error of describing vessel's cross section to 2 04%. And reconstruction errors can be significantly reduced to about 5% under ill conditions of sparse projections.