作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2012, Vol. 38 ›› Issue (10): 154-156. doi: 10.3969/j.issn.1000-3428.2012.10.047

• 人工智能及识别技术 • 上一篇    下一篇

融合LTP与Fisherfaces的分块人脸识别

袁宝华 1,王 欢 2,任明武 2   

  1. (1. 南京理工大学泰州科技学院,江苏 泰州 225300;2. 南京理工大学计算机科学与技术学院,南京 210094)
  • 收稿日期:2011-07-15 出版日期:2012-05-20 发布日期:2012-05-20
  • 作者简介:袁宝华(1980-),男,讲师、硕士、CCF会员,主研方向:智能控制,图像处理;王 欢,讲师;任明武,教授、博士生导师
  • 基金资助:
    国家自然科学基金资助项目(60875010)

Block Face Recognition Combined with Local Ternary Pattern and Fisherfaces

YUAN Bao-hua 1, WANG Huan 2, REN Ming-wu 2   

  1. (1. Taizhou Institute of Science & Technology, Nanjing University of Science and Technology, Taizhou 225300, China; 2. School of Computer Science & Technology, Nanjing University of Science and Technology, Nanjing 210094, China)
  • Received:2011-07-15 Online:2012-05-20 Published:2012-05-20

摘要: 提出一种融合局部三值模式(LTP)和Fisherfaces进行人脸识别的方法,运用LTP算子提取分块人脸灰度图像的LTP直方图序列,采用Fisherfaces方法对采样后的特征进行特征选择,根据最近邻原则进行识别。该算法不仅能提取人脸纹理信息,大幅降低训练数据量,而且数据量的维数与原始图像大小无关。在ORL和YALE标准人脸数据库上的实验结果表明,该方法具有较高的识别率。

关键词: 局部三值模式, 局部二值模式, 直方图序列, 特征抽取, Fishfaces方法, 人脸识别

Abstract: This paper presents a face recognition method based on Local Ternary Pattern(LTP) and Fisherfaces. LTP operator is used to extract the LTP Histogram Sequence(LTPHS) from block grey-level face images. Fisherfaces based on feature selection method is applied to extract feature subspace. Face recognition is realized based on the nearest neighbor principle. The proposed method can effectively extract the face texture, and can greatly reduce the amount of training data, and the dimension of the amount of data has nothing to do with the original image size. Simulation experimental results illustrate that the proposed method obtains better recognition rate on both ORL and YALE face database comparing with classical methods.

Key words: Local Ternary Pattern(LTP), Local Binary Pattern(LBP), histogram sequence, feature extraction, Fisherfaces method, face recognition

中图分类号: