具有简单约束的图像恢复正则化方法

Simple constraints regularization method for image restoration

图像恢复就是对退化的图像进行处理、尽可能恢复原图像的本来面目的过程.注意到图像的像素值一般都限制在一定的范围内,如8bit的灰度图像,其像素值在0到255之间,如果在图像恢复过程中充分考虑图像的这种特性,那么对图像恢复的效果将是有益的,因此,考虑这种带有简单约束的图像恢复问题.在一类基于分割的正则化模型的基础上,提出了利用ASCBB算法求解这种具有简单约束的图像恢复问题.数值实验结果表明,该算法是有效的.

The problem of recovering an original image x from a degraded observed version y, for the purpose of improving its quality or obtaining some type of information that is not readily available from the degraded image, is usually known as the restoration or reconstruction problem. Notice that the image pixel values are generally restricted to a certain range, such as the 8 bit gray image, its pixel values are between 0 and 255. It is beneficial for image restoration if we take full consideration to the feature of the image in the process of image restoration. As a result, we consider the image restoration with simple constraints. Based on the segmentation regularization, we propose to use the ASCBB algorithm to solve the recovery problem which has a simple constraint. The numerical experiment results show that the algorithm is effective.