作　　者： ; ; ;

机构地区： 西安电子科技大学电子工程学院

出　　处： 《中国图象图形学报》 2007年第6期1069-1078,共10页

摘　　要： 基于互信息的弹性图像配准是医学图像配准的重要方法之一。然而由于互信息在小样本图像配准中,会出现多局部极值和极值偏离问题,从而容易出现配准误差,进而造成整图的弹性配准误差。为减少这种配准误差,提出了一种基于特征分类的互信息医学图像弹性配准方法。该方法先采用图像的灰度和梯度特征训练自组织映射(self-organized mapping,SOM)神经网络特征分类器,将图像由高维灰度空间映射到低维特征类别空间;然后,在特征类别空间进行互信息图像弹性配准。实验结果表明,该方法大大提高了小样本图像配准的成功率,并可通用于有噪和无噪的医学图像弹性配准中。 Mutual information registration based on pixel intensity has been widely used in recent years. However, its application to sub-images with small samples is questionable, because many local maximums may happen or the global maximum may be away from the actual maximum value which causes unnecessary registration error. A new approach, mutual information registration based on feature-label （ MIF ） , is proposed to solve such problem. This method first uses image＇s intensity and gradient features to train the self-organized mapping（SOM） neural network, and then builds up the feature classifier for each modal image. Using such a classifier, images are project into a feature space with decreased dimensions. Finally mutual information is evaluated in the feature space to match images. Our results demonstrate that this method increases the success rate of the sub-image registration, and is optimal for the whole images（ either with or without noise） elastic registration.

关 键 词： 图像配准 弹性配准 互信息 特征分类 特征标识

领　　域： [自动化与计算机技术] [自动化与计算机技术]