小波变换在三维EM图像复原中的应用  

3-D EM image restoration based on wavelet transform

作　　者：何小海[1] 陶青川[1] 张敏[1] 张菊[1] 孟雷[1] 

机构地区：[1]四川大学电子信息学院图像信息研究所,成都610064

出　　处：《计算机工程与应用》2010年第15期169-171,238,共4页Computer Engineering and Applications

基　　金：国家自然科学基金No.60372079;教育部重点项目资助(No.107094)~~

摘　　要：提出了一种新的基于小波变换的三维EM(Expectation Maximum)图像复原算法,并将该算法用于显微光学切片的图像复原中。该算法分别在傅里叶域和小波域内交替进行,在傅里叶域进行解卷积,在小波域进行去噪。实验表明,通过对参数的合理选取,可很好地对三维图像进行复原。同时和已有的调整EM图像复原算法相比,迭代的次数少,效率明显提高。A novel 3D EM image restoration algorithm based on wavelet transform is proposed and applied in the image restoration of Computational Optical Sectioning Microscopy（COSM）.The algorithm is an iterative deconvolution process which alternates between the Fourier and wavelet domains by combining Fourier-based deconvolution with wavelet-based denoising.The results show that this algorithm can restore the sequential images slices with optimal parameter values.As compared with the existed regular-ized EM algorithm,the efficiency is increased significantly due to less iterative times.

关 键 词：图像复原 小波变换 EM算法 计算光学切片显微术(COSM) 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]