基于Kalman滤波的神经网络快速学习算法及应用

Fast Learning Algorithm for Neural Networks Based on Kalman Filter and Its Application

本文提出一种基于Kalman滤波算法的神经网络快速学习算法(EKL).经图像边缘检测应用结果表明,该算法对于加快网络学习的收敛性有着显著成效.

This paper presents a neural networks learning algorithm based on Kalman filter. The results of image edge detection show that the algorithm has better convergence properties than the conventional backpropagation learning technique.