Local Feature Extraction Models from Incomplete Data in Face Recognition Based on Nonnegative Matrix Factorization
American Journal of Software Engineering and Applications
Volume 4, Issue 3, June 2015, Pages: 50-55
Received: Apr. 21, 2015; Accepted: May 1, 2015; Published: May 13, 2015
Views 3276      Downloads 118
Authors
Yang Hongli, Science College, Shandong University of Science and Technology, Qingdao, Shandong, P. R. China
Hu Yunhong, Applied Mathematics Department, Yuncheng University, Yuncheng, P. R. China
Article Tools
Follow on us
Abstract
Data missing usually happens in the process of data collection, transmission, processing, preservation and application due to various reasons. In the research of face recognition, the missing of image pixel value will affect feature extraction. How to extract local feature from the incomplete data is an interesting as well as important problem. Nonnegative matrix factorization (NMF) is a low rank factorization method for matrix and has been successfully used in local feature extraction in various disciplines which face recognition is included. This paper mainly deals with this problem. Firstly, we classify the patterns of image pixel value missing, secondly, we provide the local feature extraction models basing on nonnegative matrix factorization under different types of missing data, thirdly, we compare the local feature extraction capabilities of the above given models under different missing ratio of the original data. Recognition rate is investigated under different data missing pattern. Numerical experiments are presented and conclusions are drawn at the end of the paper.
Keywords
Local Feature Extraction, Incomplete Data, Face Recognition, NMF, Model
To cite this article
Yang Hongli, Hu Yunhong, Local Feature Extraction Models from Incomplete Data in Face Recognition Based on Nonnegative Matrix Factorization, American Journal of Software Engineering and Applications. Vol. 4, No. 3, 2015, pp. 50-55. doi: 10.11648/j.ajsea.20150403.12
References
[1]
Hongli Yang. Nonnegative matrix and tensor factorization and their applications. Ph.D thesis, 2011.
[2]
Y. D. Kim and S. Choi. Weighted nonnegative matrix factorization. ICASSP 2009: 1541-1544.
[3]
H.Lee,J.Yoo and S.Choi. Semi-supervised nonnegative matrix factorization IEEE Signal Processing Letters, 2010, Vol (17) (1): 4-7.
[4]
S. Zhang W. H. Wang, J. Ford and F. Makedon. Learning from incomplete ratings using nonnegative matrix factorization. SIGCOMM 2006: 267-278.
[5]
E. Acar, D. M. Dunlavy, T. G. Kolda, and M. Morup. Scalable tensor factorization with missing data. Proceedings of the 2010 SIAM International Conference on Data Mining, 2010.
[6]
N. Srebro,T. Jaakkola. Weighted low rank approximation. IMCL2003: 720-727.
[7]
P. Paatero. Least squares formulation of robust, nonnegative factor analysis. Chemometrics and Intelligent Laboratory Systems, 1997, Vol (37) (1): 23-35.
[8]
A. M. Buchanan and A. W. Fitzgibbon. Damped Newton algorithms for matrix factorization with missing data CVPR2005, Vol (2): 316-322
[9]
V. D. Blondel, N. D. Ho, and P. V. Dooren. Weighted nonnegative matrix factorization and face feature extraction. Image and Vision Comput-ing, 2008.
[10]
G. Tomasi and R. Bro. Parafac and missing values. Chemometrics and intelligent laboratory systems.2005, Vol (75) (2): 163-180.
[11]
A. P. Dempster, N. M. Laird and D. B. Rubin. Maximum likelihood from incomplete data via the EM algorithm. Journal of Royal Statistical Society, 1977, Vol (39) (1): 1-38.
[12]
E. J. Candes and Y. Plan. Matrix completion with noise.arXiv: 0903.3131v1vl, 2009.
[13]
E. J. Candes and T. Tao. The power of convex relaxation: near-optimal matrix completion.arXiv:0903.1476vl,2009.
[14]
K. R. Gabriel and S. Zamir. Lower rank approximation of matrices by least squares approximation with any choice of weights. Technimetrics, 1979, Vol (21)(4): 489-498.
[15]
A. Ruhe. Numerical computation of principal components when several observations are missing. Technical report, University of Umea, Sweden, 1974.
[16]
R.J.A.Little and D.B.Rubin. The analysis of social science data with missing values. Sociological Methods and Research,1989,Vol(18)(2-3):292-326.
[17]
R. L. Carter. Solutions for missing data in structural equation modeling. Research and Practice in assessment, 2006, Vol (1)(1): 1-6.
[18]
D. Guillamet, J. Vitria and B. Schiele. Introducing a weighted nonnegative matrix factorization for image classification. Pattern Recognition Letters, 2003, Vol (24)(14): 2447-2454.
[19]
X. Li and K. Fukui. Fisher nonnegative factorization with pair wise weighting. MVA 2007, IAPR: 380-383.
[20]
P. J. B. Koeck. Missing data in image and signal processing: the case of binary objects. International journal for light and electron optics, 2004, Vol (115)(10): 459-472.
[21]
R. B. Kline. Principles and practices of structural equation modeling. Third edition, John Wiley &Sons, Inc., New York, 1988.
[22]
D. D. Lee and H. S. Seung. Learning the parts of objects by nonnegative matrix factorization. Nature, 1999, Vol (401)(6755): 788-794.
ADDRESS
Science Publishing Group
1 Rockefeller Plaza,
10th and 11th Floors,
New York, NY 10020
U.S.A.
Tel: (001)347-983-5186