首页 | 本学科首页   官方微博 | 高级检索  
     

基于DWT和ONMF的人脸表情特征提取
引用本文:杨丽瑞,张建明,王良民.基于DWT和ONMF的人脸表情特征提取[J].计算机工程与设计,2008,29(18).
作者姓名:杨丽瑞  张建明  王良民
作者单位:江苏大学计算机科学与通信工程学院,江苏,镇江,212013;江苏大学计算机科学与通信工程学院,江苏,镇江,212013;江苏大学计算机科学与通信工程学院,江苏,镇江,212013
基金项目:江苏大学校科研和教改项目
摘    要:提出了一种基于离散小波变换(discrete wavelet transformation,DWT)和标准正交非负矩阵分解(orthonormal non-negative matrix factorization,ONMF)的人脸表情特征提取方法,该方法先通过离散小波变换,降低图像的噪声,并将得到的图像的低频信息作为研究对象;再采用标准正交非负矩阵分解来提取表情特征.实验结果表明相对于PCA,NMF和LNMF方法,本文方法能够有效地减小类内与类间距离的比值,体现了很好的聚类效果;较高的识别率和较快的识别速度表明了该特征提取方法的有效性.

关 键 词:特征提取  离散小波变换  标准正交非负矩阵分解  类内类间距离比值  识别

Facial expression features extraction based on DWT and ONMF
YANG Li-rui,ZHANG Jian-ming,WANG Liang-min.Facial expression features extraction based on DWT and ONMF[J].Computer Engineering and Design,2008,29(18).
Authors:YANG Li-rui  ZHANG Jian-ming  WANG Liang-min
Affiliation:YANG Li-rui,ZHANG Jian-ming,WANG Liang-min(College of Computer Science , Communications Engineering,Jiangsu University,Zhenjiang 212013,China)
Abstract:A novel method based on discrete wavelet transformation and orthonormal non-negative matrix factorization for facial ex-pression recognition is presented.Firstly,discrete wavelet transformation is used to reduce the image noise and produce a representation in the low frequency domain.Then,orthonormal non-negative matrix factorization is used to extract the facial features.Experimental results show that this method can effectively reduce the ratio of the with-in distance and between-class distance compared t...
Keywords:feature extraction  DWT  ONMF  the ratio of the with-in distance and between-class distance  recognition  
本文献已被 CNKI 维普 万方数据 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