摘要
文章提出了一种基于混沌神经网络的图像复原新算法。在对退化图像进行复原的过程中,针对Hopfield算法易于陷入局部极小的缺点,在Hopfield神经网络中引入暂态混沌和时变增益,充分利用混沌理论的全局搜索性能进行"粗"搜索,当搜索到全局最优解附近时,再利用Hopfield算法进行局部搜索。通过对图像复原后的效果进行比较,证明基于混沌神经网络方法得到的图像复原的信噪比更高,目视效果更加。
This paper proposes a novel method to restore image by using chaotic neural network.Owing to the disadvantage of more easily trapped in local optimization,transiently chaos and time-variant gain are introduced into Hopfield algorithm.This paper makes full use of the characteristic of global search of chaos theory.First,it makes an approximately search,then it uses Hopfield algorithm for an exactly search.Comparing the performances of the restored images,experimental results demonstrate that this method of image restoration algorithm based on chaotic neural network can get higher PSNR of the restored images,and better views.
出处
《计算机与数字工程》
2012年第6期127-129,150,共4页
Computer & Digital Engineering