基于投影图像和CT容积图像的三维冠状动脉运动跟踪新方法

目的利用双平面X射线投影图像序列和参考CT容积图像进行冠状动脉的三维运动跟踪建模。方法①提取投影图像序列中的冠状动脉树;②利用多尺度滤波和血管函数提取CT容积图像中的动脉血管;③采用基于B样条配准的方法进行三维运动建模。结果将双投影图像中血管的运动估计结果的三维重建形态与三维运动模型预测出的结果进行了量化比较,调整配准的B样条参数后,两者误差处于有效范围之内。结论由此表明采用投影图像和对应容积图像的联合先验知识进行冠状动脉的三维运动建模是有效的。

A new approach of 3D coronary artery motion tracking based on projection and CT volume image

Objective To establish 3D coronary artery motion tracking by the images of double plane X-ray projection and 3D CT volume sequence. Methods ①The arterial tree from projection image sequences were extracted; ②The 3D arterial vessel from CT volume with multi-scale filter and vascular function were extracted; ③The 3D motion tracking model based on B-spline image registration was established. Results Quantitatively compared the 3D reconstruction modality with the estimated image by registration, the mean error could be controlled within the effective range if B-spline parameters were adjusted to appropriate values. Conclusion It is demonstrated that the 3D motion tracking of coronary artery with the combined prior knowledge from projection and the corresponding volume image is an effective approach.