两种图像恢复的新方法

Two New Methods for Image Restoration

本文用人工神经网络实现极大熵约束下的图像重建,并提出综合图像恢复法,对从脉冲性噪声中恢复图像有较好结果。本文还用扩展的自相关函数(EAC)法改进普通的LPC自相关函数法恢复图像也有较好的结果。以上两种图像恢复领域中新方法对解决损失部份信息下的图像重建有一定意义。

In this paper we perform image reconstruction by an artificial neural network with a constrained condition of maximizing the image entropy, and propose a new synthetic method of image restoration, we get better restoration results in case of existing impulse noises we improve the traditional LPC methods by an extended autocorrelation method to perform image restoration and also get better results. The previous two new methods of image restoration are in a certain degree of importance for reconstructing an image in case of having lost a part of informations.