一种基于PCA的医学图像特征提取与配准算法研究

Research on a Medical Image Feature Extraction and Registration Algorithm Based on PCA

为了进一步提高医学图像配准的准确性和速度,提出一种基于PCA的医学图像特征提取与配准算法。首先训练原始图像,生成样本矩阵,然后利用主成份分析的线性变换将样本图像的维度从高维空间降到低维空间,最后结合刚性变换的图像配准算法,对低维空间图像进行图像配准。临床实验数据表明,本研究算法具有较高速度与准确性,不仅可以显著提高图像的配准速度,而且还可以去除伪影和噪声,进而最大限度地保留临床意义上的图像特征信息,为研判病情与临床诊治提供了有效的帮助。

In order to improve the accuracy and speed of medical image registration,a PCA-based medical image feature extraction and registration algorithm was proposed.Firstly,the sample matrix is generated by training the original image.Secondly,this study use linear transformation of principal component analysis to reduce the dimension of sample image from high-dimensional space to low-dimensional space,and then a rigid transformation image registration algorithm is used to register low-dimensional space image.Clinical experimental data show that the algorithm has good speed and accuracy,not only can significantly improve the speed of image registration,but also artifacts and noise can be removed to maximize the retention of clinical image feature information,which provides an effective help for the research and diagnosis of disease and medical treatment.