Registration of planning CT (PCT) with daily cone beam CT (CBCT) images is an important component for adaptive radiation therapy in image guided radiation therapy system. This article uses a multiresolution wavelet representation for medical image registration to estimate deformation for the sake of improving the registration efficiency and accuracy. The wavelet, an inherently hierarchical representation of a signal, is naturally suitable for progressive optimization. The progressive optimization starts from the low resolution approximation of the signal ( represents the global deformation) , through the details in its different orientations and ends with finest details of the signal (represents the local deformation ). The wavelet coefficients are estimated by energy function which is derived from Navier partial differential equation.Furthermore, the external force is derived from the normalized mutual information which is served as the similarity measure. The experiments on PCT and CBCT images derived from image guided radiation therapy system show accuracy and robustness when compared to traditional B-spline based free form deformation (FFD)methods. In conclusion, the method can be used for re-contouring in radiation therapy planning.%计划CT图像与锥形束CT(CBCT)图像的配准是基于CBCT图像引导放射系统中实现自适应放疗的重要部分.为了提高系统中形变配准的精度和速度,提出一种基于正交小波变换的形变配准方法,此方法利用正交小波变换的多分辨率特性描述计划CT和CBCT图像的全局和局部形变,由Navier偏微分方程设计极小化能量函数来实现小波系数的能量估计.实验表明,所提出方法用于基于CBCT的图像引导放射系统时,可将日常放疗时的CBCT图像和计划CT图像进行准确且快速的配准,并且可用于放射计划系统中器官的自动分割,从而有效指导自适应治疗.
展开▼
