冠脉树三维重建优化方法的改进

Optimization Improvement for 3D Reconstruction of Coronary Arterial Tree

在冠脉造影中,两个不同角度成像坐标系间的几何变换矩阵是实现冠脉树三维重建的关键.然而,由于心脏的运动及平台的移动等,直接从造影系统中获得的几何变换矩阵不能很好地反映两个不同角度成像坐标系间的空间几何关系.几何变换矩阵中的平移向量包括两个不同角度成像坐标系间的平移向量和造影系统平台移动向量,由于冠脉树三维重建误差主要是由于平台移动影响造成的,因此,本文利用血管分支点和端点来优化几何变换矩阵中的平移向量.临床造影图像实验结果表明,该方法不仅能有效地提高冠脉树的三维重建精度,而且可获得重建的冠脉树在第三角度下的精确投影图像.

The transformation in the forms of a rotation matrix and a translation vector characterizes the spatial relationship between imaging coordinate systems in two different views, which is important for three-dimensional （3D） reconstruction of coronary arterial tree. However, the transformation directly calculated from imaging coordinate systems may not accurately define the spatial relationship between the two views due to cardiac cycles and table panning during the acquisitions. The errors of 3D reconstruction of coronary arterial tree are mainly caused by table panning. The translation vectors defining the spatial relationship of two views and the table bed panning are optimized by minimizing the errors of bifurcations and endpoints in terms of the square of Euclidean distance between the two-dimensional （2D） input data and the projection of calculated 3D method can yield more accurate data points. The experiments using 3D reconstruction of coronary arterial real coronary angiograms show the tree, and can obtain accurate projection of 3D reconstruction in the third view.