肝脏介入治疗中基于B样条变形形变模型的配准

摘要

在肝脏介入治疗的图像导航系统中,术中的配准是极为关键的一个环节,配准的质量和速度直接影响着定位和手术效果.而对于肝脏这种软组织的配准,变形的处理是不得不考虑的因素.本文以B样条网格的变形为模型,采用L-BFGS算法优化模型参数,并通过缩小控制网格的方法对配准框架进行改进,提出采用交互式选取感兴趣区域的方法,缩小B样条网格控制的区域.通过采集同一患者不同时段CT图像进行配准实验,比较不同方法的配准时间和配准测度值;结果表明,相对于传统基于B样条的方法,该新方法可以快速、准确地处理软组织变形问题.%In the image navigation system of liver interventional treatment, intraoperative registration is critical. The quality and speed of registration directly influence the positioning and the outcomes of operation. For soft-tissue organs like the liver, deformation processing must be considered during registration. In this paper, B-spline grids were used as the model of deformation. L-BFGS algorithm was used to optimize the parameters of the model, and the registration framework was improved by reducing the control grids. The method of interactive choosing the region of interest was presented to decrease the control region of B-spline grids. By using CT images of a same patient in different times to examine the new method with traditional B-spline based method, the time cost and registration measure value were used to describe the effect. Results show that this new method can solve the problem of soft tissue deformation more rapidly and accurately.