Methods for choosing the regularization parameter and estimating the noise variance in image restoration and their relation
Methods for choosing the regularization parameter and estimating the noise variance in image restoration and their relation
复制标题
DOI：
10.1109/83.148606
复制
发表时间：
1992-07-01
影响因子：
10.6
通讯作者：
Katsaggelos, Aggelos K.
中科院分区：
文献类型：
--
作者：
Galatsanos, Nikolas P.;Katsaggelos, Aggelos K.
正则化的病态问题的应用需要一个正则化参数的选择，交易的解决方案的平滑性与数据的保真度。正则化参数的值取决于数据中噪声的方差。本文研究了图像复原中正则化参数的选取和噪声方差的估计问题。基于客观均方误差（MSE）标准的误差分析被用来激励正则化。提出了两种新的正则化参数选择和噪声方差估计方法。建议和现有的方法进行了比较，并检查它们之间的关系，线性最小均方误差（LMMSE）滤波。实验验证了理论结果。
The application of regularization to ill-conditioned problems necessitates the choice of a regularization parameter which trades fidelity to the data with smoothness of the solution. The value of the regularization parameter depends on the variance of the noise in the data. In this paper the problem of choosing the regularization parameter and estimating the noise variance in image restoration is examined. An error analysis based on an objective mean square error (MSE) criterion is used to motivate regularization. Two new approaches for choosing the regularization parameter and estimating the noise variance are proposed. The proposed and existing methods are compared and their relation to linear minimum mean square error (LMMSE) filtering is examined. Experiments are presented that verify the theoretical results.