基于区域相似性测量的CT/PET图像配准

CT/PET Image Registration by Region Similarity Measure

图像配准在临床诊断中有重要意义 ,针对这一问题已经提出了许多方法。本文以区域相似性匹配测度 ,运用改进的分割方法 ,结合 Powell寻优算法实现了 CT/ PET多模医学图像配准。实验结果表明 ,该算法易于实现 ,配准速度快、精度高 ,鲁棒性较好。

Image registration is important in clinical applications and many techniques have been developed.In this paper,CT/PET multimodality medical image registration is performed by region similarity measure (RSM) accompanied with improved segment method and Powell optimization algorithm.RSM has been proved to be a quick,accurate and robust algorithm by the experimental results.