医学图像插值算法的研究  

Interpolation Algorithms of Medical Images

作　　者：许为华[1] 尹学松[2] 

机构地区：[1]浙江金融职业学院信息中心,浙江杭州310018 [2]浙江广播电视大学计算机科学与技术系,浙江杭州310012

出　　处：《计算机仿真》2006年第1期111-114,共4页Computer Simulation

摘　　要：在医学图像的显示、处理和分析中,常需要将图像进行插值处理。另外,图像插值也是三维重建中的一个重要步骤,且插值结果的好坏直接影响到重建工作的进行。文中在PC机上实现了最近邻插值、线性插值、Lagrange插值、匹配插值、基于形状的插值和小波插值等。插值试验结果表明,基于灰度插值方法的特点是计算简单,计算速度比较快,但在不同密度物质处存在边界模糊现象；基于对象插值方法的特点能够比灰度插值更为精确的轮廓边缘,但一次插值不能得到整个图像；基于小波插值方法能够克服以上两种插值方法的缺点,但这种方法的最大缺点是计算量太大,难以运用到实际中。The technique of image interpolation is usually required for medical image processing, analysis and display. In addition, image interpolation is an important step in 3 - D reconstruction. Moreover, the process of reconstruction is effected by the results of interpolation. In this paper, several interpolation methods such as nearest neighbor, linear, Lagrange polynomial, matching, shape and wavelet interpolation are implemented on PC. The experimental results show that grayl - based method computes simply and quickly, but blur of boundary is not overcome in the different density matter. Though shape - based method can overcome the blur of boundary, the whole image is not obtained by one time interpolation. Wavelet - based method can overcome the shortcomings of above methods, but much higher computation cost has limited its application.

关 键 词：医学图像 插值 小波 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]