计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (25): 24-27.DOI: 10.3778/j.issn.1002-8331.2008.25.007

• 博士论坛 • 上一篇    下一篇

基于局部线性嵌入和最近特征线的人耳识别

谢朝霞1,穆志纯1,谢建军2   

  1. 1.北京科技大学 信息工程学院,北京 100083
    2.河南科技大学 机电工程学院,河南 洛阳 471003
  • 收稿日期:2008-04-14 修回日期:2008-05-12 出版日期:2008-09-01 发布日期:2008-09-01
  • 通讯作者: 谢朝霞

Ear recognition based on locally linear embedding and nearest feature line

XIE Zhao-xia1,MU Zhi-chun1,XIE Jian-jun2   

  1. 1.Information Engineering School,University of Science and Technology Beijing,Beijing 100083,China
    2.Mechatronics Engineering School,Henan University of Science and Technology,Luoyang,Henan 471003,China
  • Received:2008-04-14 Revised:2008-05-12 Online:2008-09-01 Published:2008-09-01
  • Contact: XIE Zhao-xia

摘要: 针对人耳生物特征,通过分析早期人耳识别方法的不足,提出了一种局部线性嵌入(LLE)和最近特征线(NFL)相结合的人耳识别方法。首先依据流形学习思想,采用局部线性嵌入算法提取人耳图像特征,然后采用最近特征线分类器进行人耳识别。实验结果表明,该方法在人耳姿态变化时能够取得非常理想的识别率,提高了人耳识别的鲁棒性,增强了人耳识别技术的实用性。

关键词: 人耳识别, 流形学习, 局部线性嵌入, 最近特征线

Abstract: Based on the simply analysis of the advantages of the early ear recognition methods,an ear recognition method combining locally linear embedding(LLE) and the nearest feature line(NFL) are proposed.The LLE algorithm which based on the manifold learning technique is applied for ear feature extraction,and the NFL-based classifier is used for ear recognition.Experiment results show that this method can obtain the satisfied recognition rate perfectly as for pose variation in ear recognition,improve the robustness of ear recognition,and enhance the practicability of the ear recognition technology.

Key words: ear recognition, manifold learning, locally linear embedding, nearest feature line