检索
E-mail
RSS
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 ›› 2013, Vol. 39 ›› Issue (8): 196-199,203.doi: 10.3969/j.issn.1000-3428.2013.08.042

• 人工智能及识别技术 • 上一篇    下一篇

复杂背景中的人脸识别技术研究

王金云1，周晖杰2，纪 政3   

  1. (1. 河北汉光重工有限责任公司，河北 邯郸 056028；2. 宁波大学科学技术学院，浙江 宁波 315212； 3. 上海交通大学计算机系，上海 200240)
  • 收稿日期:2012-02-25 出版日期:2013-08-15 发布日期:2013-08-13
  • 作者简介:王金云(1979－)，男，高级工程师、硕士，主研方向：软件工程，图像识别；周晖杰，硕士；纪 政，博士研究生
  • 基金项目:
    国家自然科学基金资助项目(61175054)

Research on Face Recognition Technology in Complex Background

WANG Jin-yun 1, ZHOU Hui-jie 2, JI Zheng 3   

  1. (1. Hebei Hanguang Heavy Industry Co., Ltd., Handan 056028, China; 2. College of Science and Technology, Ningbo University, Ningbo 315212, China; 3. Department of Computer, Shanghai Jiaotong University, Shanghai 200240, China)
  • Received:2012-02-25 Online:2013-08-15 Published:2013-08-13

PDF

1370

被引次数

16 | 1

摘要： 针对复杂背景下的人脸图像，提出一种快速人脸检测识别方法。包括基于肤色模型和OpenCV的综合方法进行人脸检测定位，并对图像重新保存、预处理，用以克服光照因素的干扰，剔除复杂背景对人脸识别不利因素的影响。采用二维主成分分析算法，对同一个人多幅不同表情的人脸图像进行采集和特征提取并归类。对ORL人脸库及实际外场背景下的人脸图像进行测试，结果表明，该方法可有效解决复杂背景下的人脸识别问题，具有快速、高效的实用性，正确识别率可达90%以上。

关键词: 复杂背景, OpenCV方法, 肤色模型, 二维主成分分析, 人脸识别技术

Abstract: This paper proposes a fast face detection and recognition algorithm based on face image in the complex background. The process includes complexion model and OpenCV integrated method to solve face detection and location, re-save image, pretreatment, and illumination influence will be conquered effectively, eliminates the disadvantage factor disturbance of complex background. Afterward, it adopts Two-dimensional Principle Component Analysis(2DPCA) algorithms to collect face image of different expression which comes from the same human and distill character, and classifies the similar face image. By ORL database and outfield face image test, experimental results show that this method resolves face location and recognition effectively in the complex background and has more fast and effective practicability than other means, over 90% correct recognition rate.

Key words: complex background, OpenCV method, complexion model, Two-dimensional Principle Component Analysis(2DPCA), face recognition technology

中图分类号: