An Improved BP Algorithm and Its Application in Classification of Surface Defects of Steel Plate

Abstract

Artificial neural network is a new approach to pattern recognition and classification. The model of multilayer perceptron (MLP) and back-propagation (BP) is used to train the algorithm in the artificial neural network. An improved fast algorithm of the BP network was presented, which adopts a singular value decomposition (SVD) and a generalized inverse matrix. It not only increases the speed of network learning but also achieves a satisfying precision. The simulation and experiment results show the effect of improvement of BP algorithm on the classification of the surface defects of steel plate.