结合异质扩散和小波包分解的医学图像分割算法  

A Segmentation Algorithm of Medical Image Combining Anisotropic Diffusion and Wavelet Packet Decomposition

作　　者：熊祖坤[1] 李海云[1] 

机构地区：[1]首都医科大学生物医学工程学院,北京100069

出　　处：《计算机辅助设计与图形学学报》2008年第12期1647-1652,共6页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(30670576);北京市自然科学基金(3062006)

摘　　要：提出一种医学图像解剖结构特征提取算法,结合异质扩散和小波包分解实现了医学图像分割.在异质扩散基础上,提出一种新的FAB扩散方法对图像进行去噪,同时保持图像解剖结构的边缘和局部细节;然后用一个基于小波包分解的自适应边缘检测算法来检测图像的解剖结构的轮廓特征.分割结果表明,该方法能在有效地去除噪声的基础上完成医学图像解剖结构轮廓特征的提取.A new algorithm on extraction of anatomical contour in medical image is presented, by combining anisotropic diffusion and wavelet packet decomposition for edge detection of the image. By the algorithm, firstly, a novel FAB diffusion filter is used for both image smoothing and noise removal, as well as edge enhancement. Then, a self-adaptable edge detection algorithm based on wavelet packet decomposition is applied to detect edges. Finally, test examples and encouraging result are presented to demonstrate the efficiency of the approach for medical image processing.

关 键 词：异质扩散 小波包分解 边缘检测 噪声消除 

分 类 号：R318[医药卫生—生物医学工程]