计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (15): 180-183.

• 图形、图像、模式识别 • 上一篇    下一篇

维纳滤波和调整EM算法在COSM中的应用

王海凤,何小海,孙贵凡,贺可鑫   

  1. 四川大学 电子信息学院,成都 610064
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-05-21 发布日期:2011-05-21

Application of Wiener filter with regularized EM algorithm in COSM

WANG Haifeng,HE Xiaohai,SUN Guifan,HE Kexin   

  1. School of Electronics and Information Engineering,Sichuan University,Chengdu 610064,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-05-21 Published:2011-05-21

摘要: 用基于深度变化成像模型的调整EM算法进行三维显微图像复原,不能更好地复原图像细节,而且耗时长。为提高图像的复原质量,缩短时间,提出把维纳滤波和调整EM算法相结合的算法。该算法首先利用加权小波去除图像的部分离焦模糊,再用维纳滤波算法进行滤波复原,最后用基于深度变化成像模型的调整EM算法对序列图进行复原。实验表明复原效果得到了明显改善,并减少了迭代次数,效率明显提高。

关键词: 图像复原, 加权小波, 维纳滤波, 调整EM算法, 计算光学切片显微成像

Abstract: In the image restoration of Computational Optical Sectioning Microscopy(COSM),using regularized Expectation Maximization(EM) algorithm based on the depth-variant imaging model can not recover the more detail of image,and consume more time.In order to improve the restored results and reduce the time of the image restoration,the combination of the regularized EM algorithm and Wiener filter algorithm is proposed.Firstly,the weighted wavelet algorithm is used to reduce the information out of focus.Then,the Wiener filter algorithm and the regularized EM algorithm is used to restore the serial images in turn.Experiments show that the restoration result is improved and the iteration time is reduced.The efficiency is increased obviously.

Key words: image restoration, weighted wavelet, Wiener filter, regularized EM algorithm, computational optical sectioning microscopy