计算机工程与应用 ›› 2014, Vol. 50 ›› Issue (11): 163-166.

• 图形图像处理 • 上一篇    下一篇

基于多数投票的DBC人脸识别

师飞龙,李  岚,徐楠楠   

  1. 江南大学 物联网工程学院,江苏 无锡 214122
  • 出版日期:2014-06-01 发布日期:2015-04-08

Directional binary code based on majority voting for face recognition

SHI Feilong, LI Lan, XU Nannan   

  1. School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • Online:2014-06-01 Published:2015-04-08

摘要: 在定向二值编码(DBC)的基础上提出了一种基于Haar小波变换和多数投票的V-HaarDBC方法。先对人脸图像进行Haar小波去噪,然后通过DBC进行特征提取,最后以多数投票的方式分类。该方法不仅解决了原DBC方法中特征维数过高的问题,而且有效地提高了识别率。随后进一步提出了融合全局特征的加权WV-HaarDBC方法。基于ORL人脸库的实验表明了该方法的有效性。

关键词: 人脸识别, 定向二值编码, Haar小波变换, 多数投票

Abstract: This paper presents a Directional Binary Code(DBC) method based on Haar wavelet transformation and majority voting strategy, termed as Voting Haar Wavelet DBC(V-HaarDBC). Firstly, the noise is eliminated by Haar wavelet transformation. And then the feature fetch is based on the DBC method. Finally the majority voting strategy is used for classification. This method not only solves the problem of high-dimensional feature in original DBC, but also improves the recognition rate. And then the WV-HaarDBC(Weight Vote HaarDBC) method which fusing global feature is proposed. The experimental results on ORL face database show that the method is effective.

Key words: face recognition, directional binary code, Haar wavelet transformation, majority voting