计算机应用 ›› 2013, Vol. 33 ›› Issue (02): 507-514.DOI: 10.3724/SP.J.1087.2013.00507

• 多媒体处理技术 • 上一篇    下一篇

面向光照可变的人脸识别方法

李昕昕1,陈丹2,许凤娇2   

  1. 1. 四川大学锦城学院,成都 611731
    2. 重庆邮电大学 计算机科学与技术研究所,重庆 400065
  • 收稿日期:2012-08-23 修回日期:2012-10-06 出版日期:2013-02-01 发布日期:2013-02-25
  • 通讯作者: 李昕昕
  • 作者简介:李昕昕(1981-),女,四川成都人,讲师,硕士,CCF会员,主要研究方向:智能信息处理、模式识别、人工智能;
    陈丹(1985-),女,四川德昌人,硕士,主要研究方向:图像处理、人脸识别;
    许凤娇(1988-),女,河南汝南人,硕士研究生,主要研究方向:模式识别、人工智能。
  • 基金资助:
    国家自然科学基金资助项目;中央高校基本科研业务费专项资金资助项目

Face recognition method for scenario with lighting variation

LI Xinxin1,CHEN Dan2,XU Fengjiao2   

  1. 1. Jincheng College of Sichuan University, Chengdu Sichuan 611731, China
    2. Institute of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
  • Received:2012-08-23 Revised:2012-10-06 Online:2013-02-01 Published:2013-02-25
  • Contact: LI Xinxin

摘要: 传统Retinex算法在侧光严重的情况下难以消除阴影,为此提出一个对数形式的传导函数,取得了很好的光照补偿效果。为提高人脸识别率,将该问题看成一个典型的模式分类问题,提出基于局部二值模式(LBP)特征的支持向量机(SVM)人脸识别方法,使用“一对一”的方法将多类问题转化为SVM分类器可以解决的两类问题,实现了高效的人脸识别。在CMU PIE、AR、CAS-PEAL以及自行采集的人脸库上进行了仿真实验,结果表明该方法能够有效地去除光照影响,相对传统方法具有较优的识别性能。

关键词: 人脸识别, 光照, 局部二值模式, 支持向量机, 视网膜皮层

Abstract: With serious sidelight, it is difficult for the traditional algorithm to eliminate shadows. To improve the illumination compensation effect, a logarithmic transformation function was presented. In order to improve the performance of face recognition, by taking this problem as a classic pattern classification problem, a new method combining Local Binary Pattern (LBP) and Support Vector Machine (SVM) was proposed. One-against-one was used to convert multi-class problem to two-class problem, that can be used by SVM. Simulation experiments were conducted on the database of CMU PIE, AR, CAS-PEAL and one face database collected by the authors. The results show that lighting effects can be well eliminated and the proposed method performs better than the traditional ones.

Key words: face recognition, lighting, Local Binary Pattern (LBP), Support Vector Machine (SVM), Retinex

中图分类号: