基于改进C-V模型的超声心动图像轮廓的提取方法  

Extraction Method of Contour in Echocardiographic Images Based on Improved C-V Model

作　　者：杨渝舟 严洪[1] 仲崇发[1] 姚宇华[1] 宋晋忠[1] 

机构地区：[1]中国航天员科研训练中心,北京100094

出　　处：《现代生物医学进展》2014年第32期6205-6209,6215,共6页Progress in Modern Biomedicine

基　　金：航天医学基础与应用国家重点实验室资助项目(SMFA13B03;SMFA12B09);中国航天医学工程预先研究项目(2012SY54B0601)

摘　　要：目的:通过超声图像预处理和对图像分割方法的改进,完成超声心动图中心腔轮廓的提取。方法:首先,运用基于斑点指数的滤波方法对超声图像进行去噪。其次,对超声图像进行分段非线性灰度变换,提高图像对比度。最后,利用改进的基于C-V模型的水平集算法对超声图像进行分割,得到精确的初始轮廓。结果:1基于斑点指数的图像滤波方法可以在不丢失细节的情况下对超声图像进行噪声滤除。2分段非线性灰度变换可以有效提高超声图像的对比度。3改进的C-V模型可以成功的对含有斑点噪声的超声图像进行分割。结论:本文的超声图像预处理方法和分割算法可以有效提取心腔轮廓,降低斑点噪声对图像分割结果的影响。Objective： To extract the contour in Echocardiographic images by preprocessing ultrasound image and improving im- age segmentation method. Methods： First, a filter based on speckle index is used to suppress speckle noise in ultrasound image. Then, piecewise non-linear gray-scale transformation is used to improve image contrast. At last, the level set method based on improved C-V model is used to segment ultrasound image. Results： （1）The filter based on speckle index can eliminate noise in ultrasound image without the loss of detail. （2）The piecewise non-linear gray-scale transformation can improve the image contrast effectively. （3）Ultrasound images can be segmented successfully by using improved C-V mode. Conclusion： The preprocessing and segmentation methods can extract the contour in Echocardiographic images effectively and reduce the influence of speckle noise on the ultrasound image segmentation results.

关 键 词：灰度变换 斑点指数 超声图像滤波 超声图像分割 C-V模型 

分 类 号：R318[医药卫生—生物医学工程] R445.1[医药卫生—基础医学]