摘要
提出采用遗传算法作为人工神经网络学习算法,以优化神经网络结构的初值和权系数。实验结果表明,运用GA神经网络进行车牌识别与BP神经网络相比,其收敛性、学习速度和识别精度都有了显著提高。
This paper proposes to adopt the genetic algorithm （GA） as the study algorithm of artificial neural network to optimize the initial value and weight coefficient of ANN structure. It is indicated through the experimental results that the convergence and learning speed and the precision of recognition of using GA-ANN to recognise the Vehicles＇ License Plates, compared with BP- ANN, are improved remarkably.
出处
《机电产品开发与创新》
2005年第5期38-40,共3页
Development & Innovation of Machinery & Electrical Products
关键词
人工神经网络
遗传算法
车牌识别
artificial neural network
genetic algorithm
vehicles＇ license plates recognition
