基于NAS-RIF算法和神经网络的图像盲复原

Blind Image Restoration Algorithm Based on NAS-RIF Algorithm and Neural Network

在图像盲复原中,NAS-RIF算法在无噪情况下,能够得到较好的复原结果,但是对有观测噪声的图像复原效果不理想。而Hopfield神经网络有利于缓解图像复原过程中的震铃效应,但前提是知道退化图像的点扩展函数。将二者相结合提出一种基于NAS-RIF算法和神经网络的图像盲复原新算法,首先由NAS-RIF算法先估计出退化图像的点扩展函数,再利用Hopfield神经网络算法对其进行复原。实验结果表明,该算法具有较好的盲复原效果。

In blind image restoration, the NAS- RIF algorithm can obtain the better restoration result while without noise. When observational noise exists, the restoration obtained by NAS- RIF algorithm isn＇t ideal, However, Hopfield neural network is good at alleviating ringing effect in the process of restoration. But the premise is that the PSF of the degraded image has been known. Gives a new blind image restoration algorithm which based on NAS- RIF algorithm and neural network. NAS - RIF algorithm estimates PSF from degraded image at first , and then neural network restores. The result of the experiment shows that this algorithm will produce better effect than others.