摘要
文中采用均值偏移算法对加标记核磁共振(MR)图象的幅度谱进行模式分类构造低通滤波器,滤波后的频谱经傅立叶反变换后生成去网格标记的MR图象。然后运用Snake-GVF方法对结果图象进行分割,能够获得较好的左心室内膜边界。结果表明均值偏移算法应用于加标记MR图象分割是有效的。
Mean shift clustering algorithm is used to automatically design low-pass filter through partitioning the power spectrum of tagged magnetic resonance （MR） image, tag-removed MR image is obtained from filtered spectrum by inverse Fourier transform. Base on the tag-removed image, Snake-GVF method is used to extract the left ventricle endocardium. The result shows the mean shift al- gorithm is valid in the segmentation of tagged MR image.
出处
《微计算机信息》
2009年第21期254-256,共3页
Control & Automation
