一种基于Cross-Validation的盲图像恢复方法

图像盲复原所面临的主要问题是可利用的信息不足，目前已有的图像盲复原算法一般都是有先验知识，如非负和有限支持域的限制。但在实际中目标的支持域是未知的。文中介绍了CV(cross-validation)的基本原理，给出了一种基于CV原理的支持域确定方法的详细步骤；并针对计算大的问题提出了一定的改进。最后给出了实验结果，证明CV确定支持域的方法对于图像盲复原是有一定价值的。

The Approach to the Blind Image Restoration Based on Cross-Validation

Insufficiency of available information about images is blind image restoration's main problem. Now, existing method's information is the nonnegativity of true image and the support size of the original object. Unfortunately, the support size of the original object is unknown. This article presents basic principle of cross-validation and introduces a support-finding approach based on CV principle. We provide some improvements on great computation. Simulation results presented in this article demonstrate the reliable performance of the procedure.