全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...

基于图像抽样重组的2维线性鉴别分析

DOI: 10.11834/jig.20100212

Keywords: 2DLDA,图像抽样重组,完全PCA,NLDA

Full-Text   Cite this paper   Add to My Lib

Abstract:

图像识别中的2维线性鉴别分析(2DLDA)实际上是将图像的各个列(或行)视为样本向量,但这些样本向量不能满足统计学中的独立同分布要求。为克服2DLDA的不足,提出了基于图像抽样重组的2DLDA(SR2DLDA),它对图像进行下抽样,并将抽样所得的不同小图像重组成矩阵,然后对这些矩阵实施2DLDA。由于抽样重组的矩阵改善了各个列向量的独立性与分布同一性,因而SR2DLDA的识别性能有可能优于2DLDA,也优于LDA。在ORL人脸库、UMIST人脸库和FERET人脸库上的实验验证了SR2DLDA的有效性。

References

[1]  Yang Jian,Yang J Y,An optimal FLD algorithm for facial feature extraction,Boston,MA,USA,2001.
[2]  张生亮.杨静宇 2DPCA及2DLDA相关研究综述 [J].-世界科技研究与发展2008(3)
[3]  UMIST Face Database 2009
[4]  Phillips P J,Moon H,Rizvi S,The FERET evaluation methodology for face recognition algorithms,IEEE Transac-tions on Pattern Analysis and Machine Intelligence,2000(10).
[5]  Fisher R A,The use of multiple measuroments in taxonomic prob-lems,Annu Eugenics,1936.
[6]  Turk M,Pontland A,Eignnfaces for recognition,Cognitive Neuroscionce,1991(1).
[7]  Liu Jun,Chon Song-can,Tan Xiao-yang,A study on three linear discriminant analysis based methods in small sample size problem,Pattern Recognition,2008(1).
[8]  Guo Yue-fei,Wu Li-de,Lu Hong,Null Foley-Sammontransform,Pattern Recognition,2006(11).
[9]  Yang Jian,Zhang D,Frangi A F,Two-Dimensional PCA:A new approach to appearance-based face representation and recogni-tion,IEEE Transactions on Pattern Analysis and Machine In-teiligence,2004(1).
[10]  Kongsontana S,Rangsanseri Y,Face recognition using 2DLDA algorithm,Sydney,Australia:IEEE,2005.
[11]  刘永俊,宋东兴,何世明,二维非线性鉴别分析及人脸识别,常熟理工学院学报(自然科学),2008(2).
[12]  ORL Face Database 2009
[13]  Belhumeur Peter N,Hespanha J P,Kricgman David J,Fisherfaces:Recognition using class specific linear projection,IEEE Transactions on Pattern Analysis and Machine Intelli-gence,1997(7).
[14]  Yang Jian,Yang Jing-yu,Why can LDA be performed in PCA transformed space,Pattern Recognition,2003(2).
[15]  Chen Li-fen,Liao M H Y,Ko M T,A new LDA-based face recognition system which can solve the small sample size problem,Pattern Recognition,2000(10).

Full-Text

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133