基于像素平均值加权LBP算子的人脸图像识别算法  

Face Image Recognition Algorithm Based on Pixel Average Weighted LBP Operator

作　　者：马凤娟[1] 宋大伟[1] MA Fengjuan;SONG Dawei(Weifang Engineering Vocational College,Weifang Shandong 262500,China)

机构地区：[1]潍坊工程职业学院,山东潍坊262500

出　　处：《重庆科技学院学报（自然科学版）》2020年第6期82-85,共4页Journal of Chongqing University of Science and Technology：Natural Sciences Edition

基　　金：山东省职业教育与成人教育科研规划项目“基于‘威客’网站的高职院校实训模式研究”(2011ZCJG178)。

摘　　要：为了解决当前人脸识别算法的准确度和复杂度平衡问题,设计了一套基于像素平均值加权LBP特征的人脸识别解决方案。首先,对人脸图像进行基于灰度图像的LBP特征分析,建立数学模型,建立特征提取流程;然后,引入像素平均值加权计算,以达到放大LBP特征的目的,从而增强图像有效特征,为人脸特征识别提供数据基础;最后,融合像素平均值加权与LBP特征,进一步优化LBP特征提取过程,建立人脸识别精准识别机制。In order to solve the problem of the balance between the accuracy and complexity of current face recognition algorithms,a face recognition solution based on pixel average weighted LBP features is designed.Firstly,LBP feature analysis based on gray image is carried out for certain face image to set up mathematical model,and feature extraction process is established.Then,the weighted average pixel value is introduced to enlarge the LBP features,so as to enhance the effective features of the image and make a good data base for face feature recognition.Finally,the weighted pixel average and LBP features are combined to further optimize the LBP feature extraction process and establish the accurate face recognition mechanism.

关 键 词：人脸识别 像素平均值 LBP特征 加权 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]