欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对医学图像配准有鲁棒性强、准确性高和速度快的要求,文中提出一种基于特征点Rényi互信息的医学图像配准算法.起初从模板图像与待配准图像中依次提取出多尺度特征点,其次使用其空间坐标计算特征点Rényi互信息目标函数,实现图像配准.该算法有效地避免了多模噪声图像间的灰度差异影响,减少了待处理的数据量,同时使用Rényi互信息来消除目标函数所受的局部极值的影响,进一步提高了配准精度.实验证明该算法适于单模和多模医学图像配准,速度较快、精度高、鲁棒性强,是一种有效的自动配准方法,并且具有较好的临床应用价值.
英文摘要：
This paper proposes a new image registration algorithm based on feature points and Renyi mutual information for the requirements of medical image registration robustness, high accuracy and speed. It firstly extracts multi-scale feature points from float and template images. Then, the space coordinates of feature points are employed to estimate the object function of feature points Renyi mutual information in order to implement registration. The method can effectively avoid the difference of grey levels between two multi-modality noise medical images, reduce the amount of data which needed to be processed, and also eliminate the influence of the local optimum by Renyi mutual information. The experimental results show that the proposed algorithm is suitable for mono-modality and multi-modality medical image registration and it can achieve a higher speed with better registration accuracy and robustness. Therefore, it is an effective automatical registration method and has better value of clinical applications.
同期刊论文项目
同项目期刊论文
期刊信息
位置：