|本期目录/Table of Contents|

[1]赵文仓,王 倩,董瑞霞.基于多尺度分析的人脸识别算法研究[J].工业仪表与自动化装置,2017,(04):11-15.
 ZHAO Wencang,WANG Qian,DONG Ruixia.Research on face recognition algorithm based on multiscale analysis[J].Industrial Instrumentation & Automation,2017,(04):11-15.
点击复制

基于多尺度分析的人脸识别算法研究

《工业仪表与自动化装置》[ISSN:1000-0682/CN:61-1121/TH]

卷:
期数:
2017年04期
页码:
11-15
栏目:
出版日期:
2017-08-15

文章信息/Info

Title:
Research on face recognition algorithm based on multiscale analysis
文章编号:
1000-0682(2017)04-0000-00
作者:
赵文仓王 倩董瑞霞
(青岛科技大学 自动化与电子工程学院,山东 青岛 266042)
Author(s):
ZHAO WencangWANG QianDONG Ruixia
(College of Automation and Electronic Engineering,Qingdao University of Science & Technology,Shandong Qingdao 266042, China)
关键词:
人脸识别多尺度分析轮廓特征角点特征
Keywords:
face recognition multiscale analysis contour feature corner feature
分类号:
TP391
DOI:
-
文献标志码:
A
摘要:
针对传统的人脸识别算法易受光照等因素影响的缺点,提出了一种基于多尺度分析的人脸识别算法。首先,对采集到的人脸图像进行预处理,去除噪声,减弱无用信息,增强有用信息;然后,在粗尺度上采用形态学梯度边缘检测算法对人脸轮廓进行提取,缩小人脸库的搜索范围;最后,在细尺度上对人脸的不变特征进行提取,采用Harris角点检测算法对在粗尺度上得到的边缘图像进行特征提取,减少计算量,提高了识别速度。在自己拍摄的人脸库上对算法进行验证,实验表明,对易于进行轮廓提取的人脸图像的识别速度较快,精度较高。
Abstract:
The traditional face recognition algorithms are easily affected by some factor such as illumination, so theface recognition algorithm based on multiscale analysis is proposed in this paper. First, preprocessing on the detected face images is conducted. It can remove noise, reduce the useless information and enhance the useful information. Then, the face contour feature is extracted through the morphology gradient edge detection algorithm on the coarse scale. It can narrow down the search scope in the face database. Finally, the invariant feature is extracted on the fine scale. The feature of the edge images adopted on the coarse scale is extracted through the Harris corner detection algorithm. The calculation is reduced and the speed of recognition is improved. This algorithm is verified on the face database photographed by myself. Experimental results show that the algorithm has faster recognition speed and higher precision on the face images whose contour feature is easily extracted.

参考文献/References:

[1] Angshul Majumdar,Arusharka Bhattacharya. A Comparative Study in Wavelets, Curvelets and Contourlets as Feature Sets for Pattern Recognition[J].The International Arab Journal of Information Technology, 2009, 6(1): 47-51.

[2] 焦李成,谭山.图像的多尺度几何分析:回顾和展望[J].电子学报,2003, (S1): 1975-1981.
[3] 焦李成,侯彪,王爽,等.图像多尺度几何分析理论与应用—后小波分析理论与应用[M].西安:西安电子科技大学出版社,2008.
[4] Rafael C Gonzaez,, Richard E Woods.数字图像处理[M].3版.阮秋琦,阮宇智,译.北京:北京电子工业出版社,2011.
[5] 魏诗国.素描[M].北京:高等教育出版社,1999.[6] 孙即祥.图像处理[M].北京:科学出版社,2009.
[7] 何东健,耿楠,张义宽.数字图像处理[M].2版.西安:西安电子科技大学出版社,2012.
[8] 贺强,晏立.基于LOG和Canny算子的边缘检测算法[J].计算机工程,2011(3):210-212.?
[9] 杨帆,王志陶,张华.精通图像处理经典算法(MATLAB版)[M].北京:北京航空航天大学出版社,2014.
[10] Harris, M Stephens. A combined corner and edge detector [C]. In Alvey, 1988:147–152.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期:2016-10-18
基金项目:山东省科技计划项目(2013YD01033);山东省自然科学基金(ZR2015FL008)
作者简介:赵文仓(1973),男,山东聊城人,教授,博士(后),硕士生导师,研究方向为
更新日期/Last Update: 1900-01-01