基于小波变换和LBP算子的人脸识别研究  

Research on Face Recognition Based on Wavelet Transform and LBP Operator

作　　者：李鹏 LI Peng(Computer Information Engineering College,Guangzhou Huali Science and Technology Vocational College,Guangzhou 511325,China)

机构地区：[1]广州华立科技职业学院,计算机信息工程学院,广东广州511325

出　　处：《微型电脑应用》2022年第5期11-14,共4页Microcomputer Applications

基　　金：教育部高等教育司产学合作协同育人项目(201901096003);广州华立科技职业学院教育教学改革项目(HLZ041910)。

摘　　要：为了提高人脸识别效果,针对当前人脸识别方法存在的局限性,如识别准确率低,耗时长等,提出了基于小波变换与LBP算子的人脸识别方法。首先采集待识别的人脸图像,并采用小波变换对人脸图像作多尺分解,通过对高频子图像进行处理,消除光照变化对人脸识别的干扰,然后采用LBP算子提取人脸图像纹理特征,将特征连接组合在一起产生特征向量,最后采用k近邻算法根据特征向量建立人脸识别的分类器。采用标准人脸识别数据集Yale-B与AR作为测试对象,测试结果表明,小波变换与LBP算子能够克服当前人脸识别方法的弊端,提高人脸识别的准确率,并且人脸识别效率得到了明显改善,整体人脸识别效果要优于当前其它方法,为后续人脸处理打下了良好的基础。In order to improve the effect of face recognition,aiming at the limitations of current face recognition methods,such as low recognition accuracy and long time consuming,a face recognition method based on wavelet transform and LBP operator is proposed.Firstly,the face image to be recognized is collected,and the wavelet transform is used to decompose the face image into multi scale.The high-frequency sub images are processed to eliminate the interference of illumination changes on face recognition.Then,the LBP operator is used to extract the texture features of the face image,and the features are connected and combined to generate the feature vector.Finally,the K-nearest neighbor algorithm is used to establish the classification model of face recognition according to the feature vector.Yale-B are AR are used as testing databases.The test results show that the wavelet transform and LBP operator overcome the shortcomings of the current face recognition methods,improve the accuracy of face recognition,and the efficiency of face recognition has been significantly improved.The whole face recognition effect is better than other current methods,which lays a good foundation for the subsequent face processing foundation.

关 键 词：小波变换 人脸图像 分类器设计 多尺度分解 LBP算子 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]