为实现相同个体在不同呼吸状态下产生较大形变的三维肺部医学影像配准，提出一种基于改进Demons算法的精确有效配准方案。对待配准影像进行全局非刚性配准，通过尺度不变特征变换算法对影像进行特征点提取与匹配，根据匹配结果计算变换参数，完成全局配准；利用改进Demons算法对全局配准后的影像进行局部非刚性配准。使用改进的方案实现了人体肺部影像配准，并且肺部整体轮廓以及内部组织的配准结果较理想。配准前，影像间的均方误差值为25835．3，经配准后影像间均方误差值降为3726．31，均方误差值下降率为85．58％。提出的方案能够有效配准三维肺部影像，对肺部呼吸运动估计以及呼吸功能分析提供了良好的基础。%To register three-dimensional pulmonary medical images of the same individual whose lung has large deformation under different respiration state,this paper put forward an accurate and effective registration method based on improved Demons algorithm.Firstly,it registered images globally and non-rigidly.It extracted feature points and matched by scale invariant fea-ture transform algorithm.It finished the global registration according to the transformation parameter computed based on matc-hing results.Afterwards,it registered images after global registration locally and non-rigidly utilizing improved Demons algo-rithm.It realized image registration of human lung employing the improved method,and the registration results of the overall profile and internal organization of the lung were ideal.The mean-square error between images before registration was 25 835.3 and it was reduced to 3 726.31 after registration.The descent rate of mean-square was 85.58%.The proposed method effec-tively registers three-dimensional pulmonary images,which provides a solid foundation for doctors to estimate pulmonary respir-atory movement and analyze respiratory function.
展开▼
