一种混沌神经网络及其在二维局部均值估计自适应滤波中的应用

Chaotic Neural Networks and Their Application in the 2-D Adaptive Local Mean Estimation Filtering

本文研究了具有模拟退火特性的混沌神经网络模型,给出了混沌神经网络的能量函数表达式及其搜索和优化过程,并将其应用于二维自适应局部均值估计滤波算法中。仿真结果表明,利用混沌神经网络进行二维自适应滤波是可行的,且比Hopfield神经网络具有更快的收敛效率。

A mathematical model of the chaotic neural networks （CNN） with simulated annealing characteristics is proposed in this paper. The energy function of CNN and its process of searching and optimizing are given,and CNN is applied to a 2-D adaptive local mean estimation filtering algorithm. The simulation results show that CNN is suitable for 2-D adaptive filtering and has better convergence property than the Hopfield neural network.