颈动脉B超估计血管壁搏动位移的干扰抑制

发布时间：2018-02-27 00:05

本文关键词： B型超声血管壁搏动检测刚性配准相位特征主轴质心法斑点跟踪互信息各向异性斑点减少　出处：《电子学报》2017年12期 　论文类型：期刊论文

【摘要】：超声检测时,由于测量者把握探头的不稳定、受试者的呼吸运动等造成的干扰,将对通过B超图像进行斑点跟踪提取到的动脉壁位移的精度造成影响.为此,本文提出基于B超序列相位图像的全局刚性特征进行配准消除干扰.相位特征通过对B超图像进行Riesz变换,再采用高斯差分法提取得到.为了更快地完成寻优,减少配准误差,采用了基于位置加权的主轴质心法与互信息相结合的分阶段的配准方法.结果表明,与直接对B超或相位图采用斑点跟踪的方法相比,该方法有效去除了干扰,管壁位移提取的精度有显著提高.本文方法还可用于心脑等医学图像的配准.
[Abstract]:In ultrasonic detection, the disturbance caused by the unstable grasp of the probe and the breathing movement of the subjects will affect the accuracy of the arterial wall displacement extracted from the spot tracking by B-ultrasound image. In this paper, the global rigid feature based on the phase image of B-mode ultrasound sequence is proposed to remove the interference. The phase feature is extracted by Riesz transform and Gao Si difference method. In order to achieve the optimization faster and reduce the registration error, the phase feature is obtained by using Gao Si difference method. A phased registration method based on position weighted spindle centroid method and mutual information is used. The results show that the proposed method is more effective than the spot tracking method for B-mode ultrasound or phase map. The accuracy of tube wall displacement extraction is significantly improved, and this method can also be used for the registration of medical images such as heart and brain.
【作者单位】：云南大学信息学院电子工程系;西南林业大学计算机与信息学院;昆明医学院第二附属医院;
【基金】：国家自然科学基金(No.61261007,No.61561049) 云南省自然科学基金重点项目(No.2013FA008)
【分类号】：R445.1;TP391.41

【相似文献】