BP神经网络在模式识别中的的研究.doc

人工智能科技论文 目录 一：摘要························································3 二：神经网络····················································4 2.1引言 2.2 BP神经网络 2.3模式识别的应用 BP神经网络在模式识别中的研究Research on pattern recognition with BP neural network 一、摘要： BP神经网络是目前研究最为成熟、应用最为广泛的人工神经网络模型之一被广泛应用于自动控制、模式识别、图像识别、信号处理、预测、函数拟合、系统仿真等学科和领域中。其中模式识别是神经网络最有前景的应用领域之一。本文首先详细介绍了应用神经网络进行模式识别的发展概况，通过与传统模式识别的比较，得出神经网络模式识别的优越性。针对已有的神经网络模式识别系统，本文介绍了该系统中使用最广泛的BP神经网络，通过对其关键技术及算法的研究，针对BP算法的不足，提出了BP算法的启发式改进。本文结合遗传算法对BP算法进行优化，建立了改进的GABP神经网络模型。然后具体研究3种模式识别提出了一种利用传统BP(Back-Propagation 误差逆传播)网，能够在合理的误差范围内以较好的效率成功识别字符基于BP神经网络原理对手写数字进行识通过对1000个手写数字的识别证实本方法在手写数字识别方面的有效性；利用BP神经网络的理论知识研究人脸识别技术对图像中人脸朝向判别进行实验研究。关键字：人工智能、神经网络、BP神经网络、遗传算法、模式识别、BP算法、人脸识别、字符识别、数字识别Abstract: BP neural network is the most mature, most of the artificial neural network models widely, is widely used in automatic control, pattern recognition, image recognition, signal processing, forecasting, function approximation, system simulation and other disciplines and fields. This thesis first introduces the development of pattern recognition using Artificial Neural Network in details to get the superiority of Neural Network Recognition over traditional pattern recognition methods．Then on the basis of existing Neural Network Recognition system ,the thesis introduces the BP Neural Network ,which is most widely used in the system．By studying the critical technology and algorithm and focusing on the disadvantage of BP Algorithm .we bring out the heuristic improvement of BP Algorithm．This thesis improves the BP Algorithm combined with Genetic Algorithm and establishes the modified GA-BP Genetic Algorithm Model．Then I study in detail three kinds of pattern recognition , a way to use BP network to recognize printed characters was proposed, after a lot of experiments, the printed characters can be recognized successful within the reasonable range of error ; For handwritten digit recognition principle based on BP neural network, through identi