首页 | 官方网站   微博 | 高级检索  
     

基于小波变换的人脸特征定位
引用本文:汪妍,夏良正.基于小波变换的人脸特征定位[J].东南大学学报(自然科学版),2004(Z1).
作者姓名:汪妍  夏良正
作者单位:东南大学自动控制系 南京210096 (汪妍),东南大学自动控制系 南京210096(夏良正)
摘    要:提出一种基于离散小波变换 (DWT)的人脸检测与特征定位算法 ,算法针对人脸的不同特征 ,选择小波变换相应最有效的频段和分量 .运用椭圆检测算法、模板匹配算法、Hough变换等方法 ,并结合人脸各特征间的几何特征关系 ,采用由粗到细的方法分步地检测人脸范围 ,分割定位人脸的眼睛、虹膜、嘴、鼻尖等特征 .该算法对人脸姿态、表情变化和遮盖物具有一定的鲁棒性 ,同时具有计算量小、效率高的特点 .

关 键 词:人脸特征定位  小波变换  椭圆检测  模板匹配  Hough变换

Localization of facial features based on wavelet
Wang Yan,Xia Liangzheng.Localization of facial features based on wavelet[J].Journal of Southeast University(Natural Science Edition),2004(Z1).
Authors:Wang Yan  Xia Liangzheng
Abstract:A new method of facial features localization based on discrete wavelet transform ation (DWT) is presented. According to different facial features, different comp onents and spectrums from wavelet transformation are selected. Then face area an d facial features such as eyes, irises, mouth and nose, are localized step by st ep by using the geometric features among facial features, ellipse detection algo rithm, template matching algorithm and Hough transformation. This method is more robust for changes of facial pose, expression and covering, meanwhile it needs less computation.
Keywords:facial features localization  wavelet transformation  ellipse detection  templat e matching  Hough transformation
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号