【作者】 林杰灵；

【导师】 李桂清； 顾生华；

【作者基本信息】 华南理工大学 ， 计算机技术， 2010， 硕士

【摘要】 本文探讨基于小波变换的人脸识别。对介绍了传统小波变换和Gabor小波变换及其对比分析，并利用子空间分析（特征脸）技术作了几个人脸识别实验。具体内容如下：介绍了人脸识别的两步骤：①人脸检测和定位。包括：基于统计的人脸检测方法和基于知识建模的人脸检测方法。②人脸特征提取与识别的几种主要方法：基于几何特征的人脸识别、基于子空间分析的人脸识别、基于弹性图匹配的人脸识别、基于神经网络的人脸识别、基于隐马尔可夫模型的人脸识别和基于模板匹配的人脸识别等。其次，分别用传统小波变换和Gabor小波变换对人脸图像做处理，并做识别实验。最后，分别应用了在实验室中采集的人脸图像以及从ORL人脸库和Stirling人脸库中抽取的人脸图像进行具体实验，并对实验结果进行分析。更多还原

【Abstract】 This paper investigates face recognition based on wavelet transformation. It alsodescribes traditional wavelets and Gabor wavelets as well as their comparison. Several facerecognition experiments are performed based on subspace analysis (feature face). Our majorwork is as follows.Two steps of face recognition are discussed:①face detection and location which includestatistical method and knowledge based modeling methods;②facial feature extraction andrecognition including geometric features, subspace analysis, elastic graph matching,neural network, hidden Markov model, and template matching methods etc..Next, analysis of the wavelet transform in image processing applications, use traditionalwavelet transform and Gabor wavelet transform to processing pictures.Finally, we use the face images collected in our laboratory, the ORL face library and theStirling international authoritative face library to perform some specific experiments, thenanalyze the experimental results to show the advantages of wavelet transformation in facerecognition.更多还原