计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (4): 190-192.DOI: 10.3778/j.issn.1002-8331.2009.04.054

• 图形、图像、模式识别 • 上一篇    下一篇

奇异值和RBF神经网络的彩色人脸识别

罗仁泽1,冉瑞生2,王汝言3   

  1. 1.电子科技大学 中山学院 电子工程系,广东 中山 528402
    2.中冶赛迪工程技术股份有限公司 自动化事业部,重庆 400013
    3.重庆邮电大学 通信与信息工程学院,重庆 400010
  • 收稿日期:2008-09-09 修回日期:2008-10-30 出版日期:2009-02-01 发布日期:2009-02-01
  • 通讯作者: 罗仁泽

Color face recognition based on singular values and RBF neural network

LUO Ren-ze1,RAN Rui-sheng2,WANG Ru-yan3   

  1. 1.Department of Electronic Engineering,Zhongshan Institute,University of Electronic Science and Technology of China,Zhongshan, Guangdong 528402,China
    2.Department of Automation,CISDI Engineering Co.,Ltd,Chongqing 400013,China
    3.Department of Communication and Information Engineering,Chongqing University of Posts and Telecommunication,Chongqing 400065,China
  • Received:2008-09-09 Revised:2008-10-30 Online:2009-02-01 Published:2009-02-01
  • Contact: LUO Ren-ze

摘要: 目前已有研究表明,相对于灰度图像,利用图像的彩色信息能改进人脸图像的识别率。但近年来的彩色人脸识别研究较少。提出了一种基于奇异值向量和RBF神经网络的彩色人脸图像识别方法。首先说明了彩色图像的奇异值向量具有代数和几何不变性,再将降维的奇异值向量作为图像的特征,然后应用RBF神经网络进行训练和识别。实验表明该方法的识别率为95%左右,是一种有效的彩色人脸识别方法。

关键词: 奇异值, 径向基函数神经网络, 彩色人脸识别

Abstract: As opposed to gray images,the color information of face images may be applied to improve the face image recognition rate.But only a few methods for color face recognition has been proposed in the recent years.Based on the singular value vector of color face image and RBF neural network,a method for color face recognition is presented.The algebraic and geometric invariance of singular value vector is firstly given,and then singular value vector compressed is extracted as the image feature,then RBF neural network is used to training and recognition.Experiment shows that the recognition rate of the proposed method is about 95% and it is an effective color face recognition method.

Key words: singular values, Radial Basis Function(RBF) neural network, color face recognition