计算机与现代化

• 图像处理 • 上一篇    下一篇

基于NIB2DPCA的彩色图像过完整分块特征抽取方法

  

  1. 1.无锡职业技术学院物联网技术学院,江苏无锡214121;
    2.江南大学物联网应用技术教育部工程研究中心,江苏无锡214122;
    3.江南大学晓山股份联合实验室,江苏无锡214122
  • 收稿日期:2015-10-20 出版日期:2015-12-23 发布日期:2015-12-30
  • 作者简介:黄可望(1979-),女,江苏无锡人,无锡职业技术学院物联网技术学院讲师,硕士,研究方向:计算机应用,模式识别; 冯宗越(1990-),男,江南大学物联网应用技术教育部工程研究中心和 江南大学晓山股份联合实验室硕士研究生,研究方向:模式识别,人工智能; 朱嘉钢(1957-),男,副教授,博士,研究方向:模式识别,人工智能。
  • 基金资助:
    江苏省自然科学基金资助项目(BK20131106); 江苏省产学研联创项目(BY2013015-40)

NIB2DPCAbased Color Image Feature Extraction Method with Overcomplete Divided

  1. 1. School of the Internet of Things Technology, Wuxi Institute of Technology, Wuxi 214121, China;

    2. Engineering Research Center of Internet of Things Applied Technology, Ministry of Education, Jiangnan University, Wuxi 214122, China; 
    3. CoLaboratory in Hillsun Ltd. of Jiangnan University, Wuxi 214122, China
  • Received:2015-10-20 Online:2015-12-23 Published:2015-12-30

摘要:

在小空间占用的快速彩色图像的特征抽取方法和模块化FPCA(MFPCA)彩色图像特征提取方法的基础上,结合最新的过完整表示思想,提出基于NIB2DPCA的彩色图像过完整分块特征抽取新方法。该
方法对彩色图像进行过完整分块,然后对子图像模块从R、G、B三个信道用NIB2DPCA方法进行特征提取、重构,并进行多模块融合,最终获得分类的特征矩阵。该方法提取的信息量远大于原图像,提高了
彩色图像的识别率。通过在FEI和CVL标准彩色人脸数据库上的对比实验表明,所提出方法的人脸识别准确率比文献[14]中的小空间占用的快速彩色图像特征抽取方法提高约4%,比文献[19]中的彩色
图像MFPCA方法提高约5%。

关键词: 彩色人脸识别, 过完整分块, NIB2DPCA, 特征抽取

Abstract:

Based on color image feature extraction method with small space occupying and fast speed and color image feature extraction method with Modular Factored Principal
Components Analysis(color MFPCA), combined with the latest overcomplete representation idea, a novel method named NIB2DPCAbased color image feature extraction method with
overcomplete divided was proposed. This method can make the color image overcomplete divided into smaller modular images, then NIB2DPCA is employed to extract feature
information from three channels of a given sub color image module respectively. Then the three extracted feature matrices are reconstructed and multi module integrated. Finally,
the classification feature matrix is gotten. The information amount extracted by the novel method is much larger than the original color image by this method and it improves the
recognition accuracy of color image. The results of contrast experiments on CVL and FEI color face databases show that, the proposed method can obtain a higher accuracy by about
4% than color image feature extraction method with small space occupying and fast speed in the reference[14] and also obtain a higher accuracy by about 5% than color MFPCA
in the reference[19].

Key words: color face recognition, overcomplete divide, noniteration bilateral two dimensional principal component analysis(NIB2DPCA), feature extraction

中图分类号: