基于稀疏散乱数据的实时心内膜三维重建方法

Real-time 3D endocardium surface reconstruction method based on sparse and scatter data

心内膜三维可视化可用于辅助心内科导管消融手术中靶点的准确定位和消融。为满足术中图像导航需求,针对由心内膜结构上动态采集到的三维坐标点的稀疏性和散乱性进行分析,提出一种心内膜三维模型实时构建方法。以采集到的稀疏采样点云的空间坐标数据构建心内膜网格模型,并以薄板样条(thin plate splines,TPS)作为网格模型弯曲能量的约束准则,通过对心内膜网格模型的形变逼近目标结构,快速准确达到收敛条件。实验结果表明,该方法重建模型对心内膜表面整体和局部具有良好的贴合性,重建的表面误差满足临床要求,模型形变时间满足由动态采集的坐标点实时更新对应心内膜表面的要求。

A three-dimensional visualization model of endocardium is extremely important for localization of target and ablation of tissue in catheter ablation procedure. This paper proposed a new method to reconstruct and sequentially captured update the object surface immediately after a spatial point on the endocardium based on the property of the 3D point cloud. It constructed a grid morphable model of endocardium to enclose the scatter and sparse points extracted from the target shape as a prior knowledge,then used a thin-plate spline（ TPS） rule to minimize the bending energy of the grid. Experiments result shows that the proposed method can provide an accurate global and local shape approximation,the reconstruction error can satisfy clinical requirement and can update the surface in real time.