一种神经网络的快速学习算法及其在图象边缘检测中的应用

A FAST LEARNING ALGORITHM FOR NEURAL NETWORKS AND ITS APPLICATION IN IMAGE EDGE DETECTOR

摘要: 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）．经图象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的成效

Abstract: This paper presents a neural network learning algorithm based on Kalman filter algorithm. The results of image edge detector show that the algorithm has better convergence properties than the conventinal backpropagation learning technique.