造影中冠状动脉的三维序列重建和运动估计

3D sequential reconstruction and motion estimation of coronary artery from Angiograms

提出采用snake模型对多角度造影图像序列中的冠状动脉血管轴线进行三维序列重建和运动估计的方法。snake能量函数综合考虑了曲线本身的特性、图像数据以及冠脉形态和运动的先验知识,确保三维snake曲线在内外力的共同作用下,直接在空间中变形,最终停留在三维血管轴线上,从而同时完成血管轴线的二维提取、三维序列重建、运动跟踪与估计,提高了重建的精度和速度。采用临床图像的实验验证了算法的可行性。

The coronary arterial vessd axes are sequentially reconstructed and tracked from X-ray angiogram sequences based on snake model. The energy function of snake combines the intrinsic property of the curve, the acquired image data and in priori knowledge of coronary arterial configuration and dynamic performance. Snake curve deforms in 3D directly under the constraints of internal and external forces,and it ultimatdy stops at the real axis of vessel. As a result,the 2D extraction,3D sequential reconstruction and motion tracking and estimation are synchronously implemented. Both recovery accuracy and computation cost are improved. The method is validated on clinical data and experimental results are encouraging.