使用多尺度LBP特征描述与识别人脸.pdfVIP

第16卷 第4期 光学 精密工程 VoI．16 No．4 2008年4月 Optics and Precision Engineering Apr．2008 文章编号 1004—924X(2008)04—0696—10 使用多尺度LBP特征描述与识别人脸 王 玮，黄非非，李见为，冯海亮 (重庆大学光电技术及系统教育部重点实验室，重庆400030) 摘要：提出了一种基于多尺度LBP特征的人脸描述与识别算法。对原始人脸图像进行二级小波分解，并采用LBP算子 分别计算两幅低频逼近图像的LBP特征谱；将LBP特征谱划分为若干个互不重叠的特征区域，然后分别进行直方图统 计。最后，将所有区域的LBP直方图序列连接起来得到多尺度LBP特征，将其作为人脸的鉴别特征用于分类识别。所 提出算法在ORL人脸数据库中取得了高达99 的人脸识别率。实验分析表明，多尺度LBP特征具有较强的人脸图像 描述能力和可鉴别性，且对人脸表情及位置的变化具有较高的鲁棒性。 关 键 词：人脸识别；多尺度分析；LBP算子；直方图 中图分类号：TP391．41 文献标识码：A Face description and recognition using multi_。scale LBP feature WANG Wei，HUANG Fei—fei，LI Jian—wei，FENG Hai—liang (Key Laboratory of Optoelectronic Technology and System of the Education Ministry， Chongqing University，Chongqing 400030，China) Abstract：In order to improve the accuracy and robustness of face recognition，a face description and recognition method based on multi—scale Local Binary Pattern(LBP)feature is proposed．The original face image is decomposed into two levels by wavelet analysis，and the LBP operator is applied to two approximate images respectively to extract LBP feature map．Then，the two maps are divided into several regions，in which the histograms are computed and linked to get a multi—scale LBP feature．Fi— nally，the multi—scale LBP feature is used as the face descriptor for classification and recognition．The experimental results on ORL face database show that the proposed method can achieve high face rec— ognition rate up to 99 ，which shows that the multi—scale LBP feature has highly descriptive and dis— criminable abilities for human face and is robust to face expressions and position variations． Key words：face recognition；multi—scale analysis；Local Binary Pattern(LBP)operator；histogram 收稿日期：2007-11-12；修订日期：2007—12—27． 基金