电子学报 ›› 2018, Vol. 46 ›› Issue (7): 1593-1600.DOI: 10.3969/j.issn.0372-2112.2018.07.008

• 学术论文 • 上一篇    下一篇

抗年龄干扰的人脸识别

吴长虹, 苏剑波, 陈叶飞   

  1. 上海交通大学自动化系, 系统控制与信息处理教育部重点实验室, 上海 200240
  • 收稿日期:2017-01-04 修回日期:2017-06-22 出版日期:2018-07-25
    • 作者简介:
    • 吴长虹,女,1992年8月生于江苏.上海交通大学自动化系硕士研究生,研究方向为人脸识别、图像处理.E-mail:wuchanghong@sjtu.edu.cn;苏剑波,男,1969年11月生于江苏.上海交通大学自动化系教授,研究方向为机器视觉、机器学习与人机交互、多传感器信息融合、智能机器人等;陈叶飞,男,1988年生于上海.现为上海交通大学自动化系博士,研究方向为人脸识别、深度学习.
    • 基金资助:
    • 国家自然科学基金 (No.61521063,No.61533012)

Age Invariant Face Recognition

WU Chang-hong, SU Jian-bo, CHEN Ye-fei   

  1. Department of Automation, Shanghai Jiao Tong University, Ministry of Education Key Laboratory of System Control and Information Processing, Shanghai 200240, China
  • Received:2017-01-04 Revised:2017-06-22 Online:2018-07-25 Published:2018-07-25
    • Supported by:
    • National Natural Science Foundation of China (No.61521063, No.61533012)

摘要: 本文将人脸图像特征分解为身份特征部分和年龄干扰特征部分,并分别投影到两个独立的子空间.为了提高由字典张成的特征子空间的表征能力和区分能力,字典学习过程中同时引入了人脸图像的重构误差约束项和类别约束项.因此,任意的人脸特征都可以由学习到的身份字典和年龄字典投影到对应的身份子空间和年龄子空间,然后再基于身份子空间进行人脸识别,从而使年龄的干扰得到了有效的抑制.通过在MORPH和FGNET数据库上的实验,证实了基于年龄不变的身份特征子空间学习方法能提升人脸识别性能.

关键词: 人脸识别, 年龄干扰, 字典分解, 子空间学习

Abstract: In this paper,the face image is divided into two separated parts:aging effect feature part and identity related feature part.The identity dictionary and the age dictionary are introduced to encode the two feature parts into two separated feature spaces.To make sure the learned dictionaries are discriminative for different classes,the reconstruction error and label matrices constraints are added in the training.Face features can be encoded into identity and age space with the learned identity and age dictionaries.The identity space can be used for further classification.Extensive experiments are conducted on the MORPH and FGNET dataset,illustrating a great improvement over the state-of-the-arts.

Key words: face recognition, age invariance, dictionary decomposition, space learning

中图分类号: