| 基于双随机多变量高斯模型的Contourlet域贝叶斯图像估计 |
| |
| 引用本文： | 刘晓燕,于惠钧,谭兮. 基于双随机多变量高斯模型的Contourlet域贝叶斯图像估计[J]. 包装工程, 2010, 31(3): 58-61 |
| |
| 作者姓名： | 刘晓燕 于惠钧 谭兮 |
| |
| 作者单位： | 株洲职工大学,株洲,412008;湖南工业大学,株洲,412008 |
| |
| 基金项目： | 中国包装总公司2008年度科研计划学科建设项目(2008-XK09) |
| |
| 摘 要： | 提出了一种用双随机多变量高斯模型来刻画图像Contourlet变换系数统计行为的新方法。这种双随机高斯模型不仅考虑相邻系数间相关性,并且用一个Bernoulli随机变量来刻画系数的空间非平稳特性。利用这种双随机多变量高斯模型,导出了一种新颖的图像贝叶斯最小均方误差估计子。实验结果表明,估计子不仅能得到较满意的"线形结构",而且能获得比现有方法更高的信噪比。
|
| 关 键 词： | Contourlet变换 图像去噪 贝叶斯估计 双随机过程 |
Bayesian Image Estimation Based on Dual-random Multivariate Gaussian Model in Contourlet Domain |
| |
| LIU Xiao-yan,YU Hui-jun,TAN Xi. Bayesian Image Estimation Based on Dual-random Multivariate Gaussian Model in Contourlet Domain[J]. Packaging Engineering, 2010, 31(3): 58-61 |
| |
| Authors: | LIU Xiao-yan YU Hui-jun TAN Xi |
| |
| Affiliation: | 1.Zhuzhou Worker's College;Zhuzhou 412008;China;2.Hunan University of Technology;China |
| |
| Abstract: | A new method to characterize the statistical property of Contourlet coefficients of images using a dual-random multivariate Gaussian model was put forward.The dual-random multivariate Gaussian model not only takes into account the dependent property of coefficients,but also uses a Bernoulli random variable to characterize the non-stationary property of coefficients.Based on the dual-random multivariate Gaussian model,a novel Bayesian Minimum Mean Square Error estimator (MMSE) for image denoising was develop... |
| |
| Keywords: | Contourlet transform images denoising Bayesian estimation dual-random process |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
