论文部分内容阅读
针对存在自主运动器官所采集的医学影像,在利用传统的基于B样条的医学图像非刚性配准算法的基础上,提出一种多层B样条非刚性配准的方法,网格控制点由少到多变化,后一次配准在前一次配准的效果上进行,通过多次迭代降低了控制网格密度选取上的不准确性.针对多层次B样条非刚性配准方法计算量相当大导致配准时间过长,无法达到临床上实时配准要求的问题,提出一种力矩主轴法和局部更新策略相结合的分层B样条快速配准方法.该方法首先利用力矩主轴实现两幅待配准图像的粗配准,减少细配准时的计算量;然后采用局部更新策略的B样条非刚性配准方法实现细配准.实验结果表明,本文方法兼顾了配准的精度与速度,获得了较好的医学图像的非刚性配准结果.
Aiming at the medical images acquired by autonomous motor, a non-rigid registration algorithm based on B-spline is proposed based on B-spline non-rigid medical images. A grid control point From less to more changes, the latter registration on the effect of the previous registration, through a number of iterations to reduce the control grid density selection on the inaccuracy.For the multi-level B-spline non-rigid registration method calculation Aiming at the problem of too long registration time and unable to meet the requirement of clinical real-time registration, a fast stratified B-spline registration method based on moment master axis method and local updating strategy is proposed.Firstly, The coarse registration of the two images to be registered reduces the amount of fine registration, and then uses the locally updated strategy of B-spline non-rigid registration to achieve fine registration.The experimental results show that the proposed method takes into account the accuracy of registration With the speed, a non-rigid registration result of better medical images is obtained.