基于主成分分析和神经树的人脸识别方法研究-模式识别与智能系统专业论文.docxVIP

基于主成分分掇农毒孛经事孽懿入黢识别方法研究分析方法、分块主成分分析方法等，最后概括了这些方法的基本理论和应用方法。 基于主成分分掇农毒孛经事孽懿入黢识别方法研究 分析方法、分块主成分分析方法等，最后概括了这些方法的基本理论和应用方法。 (2)对神经霹络和神经树的基本理论进行7详细的综述。概述了神经网络的特 点、产生和发展历史，介绍了神经网络的基本思想、研究领域和应用发展情况，然后 从理论和应用研究上综述神经网络的研究现状，总结了神经网络设计中的难点问题。 研究了神经树的特点和实现方法。介绍了采用遗传编程的参数优化算法一粒子群优 化算法，阐述了建立神经树模型的方法。 (3)将主成分分孝厅秘神经树模型应用予人脸识别领域。在国际通用的人脸图像 库ORL和Yale数据库上实验的结果都表明基于生成分分析和神经树的人脸识别比其 他方法的实验结果更具有效性和可靠性。 关键字：人脸识别；主成分分析；灵活神经树；特征提取；粒子群优化算法 Ⅱ 济南大学硕士学位论文Ab 济南大学硕士学位论文 Ab stract The technology of face recognition is a technology that discriminate identify by using the computer to collect the feature of human face，it combines theories and ways of several subjects such as：paRem recognition，image processing，computer figure，computer vision， artificial neural networks，biography character,etc．Researchers of this area need wide-ranging knowledge system and rich experiences． The progress of face recognition mainly consists of three steps：image preprocessing， feature extraction and recognition，feature extraction and recognition are the most important of it．Feature extraction is the way to degrade high dimensions into low dimensions and maintain the main characters in the mean time，that is to say we hope describe the whole image by fewer characters．Principle components analysis is widely used in a lot of areas of science and engineering，is use fewer new variables which contain more information of old variables instead of old variables on the base of relationship of every variables together． How to choose the recognition way has the deep effect on the recognition effects．Now there are many ways of classification：Artificial neural networks(ANN)，Flexible neural tree(FNT)，Support vectors machine(SVM)，Fuzzy system(FS)，Bayesian，ere．Every of them have the both advantages and disadvantages on themselves．Among the research of artificial neural networks，the most important problems are how to select the appropriate architecture and relative parameters for all artificial neural network．Weights and biases of artificial