ECEn 671: Project Proposal-
Face Recognition Using Eigen faces
By Farnaz Karimdady Sharifabad & Binh Tran
From accent to modern days, identification has taken an important factor in the community life of
human society. The ability to recognize and distinguish different appearances is mainly based on the
neuron intelligent which can analyze vision from eyes. One of the significant distinguishable
appearances is human face. Based on this factor, many researchers have been looking for a way to train
the computer or even robot to recognize and detect different people based on their faces. This research
promises a wide range of applications from security systems to credit card verification or even criminal
identification.
There are mainly three ways to recognize and distinguish different faces. The first way is naturally used
by human brain to detect individual faces based on memory of facial characteristics. The second ways is
based on feature vectors extracted from profile projection. The third one is dealt with vectors extracted
from a frontal view of the face. Behind these three methods, there are two main approaches or
mechanisms for recognizing and decoding. One is based on the entire face image which give the
eigenfaces of the person, the other one focus on separating and combining individual parts of the face
such as eyes, nose, mouth..
In this work, we tend to investigate recognition of human faces using the projection of the face images
onto a feature space, namely “face space”. The face space will be determined by the “eigenfaces”,
which are the eigenvectors of the training set of face images. Thus, at the first step, the eigenfaces are
extracted from the training set of images. The number of eigenfaces can be less than the number of
images inside the training set.
In order to classify a given unknown face image, it will be projected into face space, and norm of the
differences between this projection vector and the projection vectors of the face classes will be
compared, and the class with minimum distance, which is also below a threshold, will be chosen (face
recognition). In the case that the distance is too large, we may decide the given image is not a face (face
detection). In this project, we try to implement face recognition and face detection algorithms in
Matlab.
References
[1] Kirby, M., and Sirovich, L., "Application of the Karhunen-Loeve procedure for the characterization of
human faces", IEEE PAMI, Vol. 12, pp. 103-108, (1990).
[2] Manjunath, B. S., Chellappa, R., and Malsburg, C., "A feature based approach to face recognition",
Trans. of IEEE, pp. 373-378, (1992).
[3] Nakamura, O., Mathur, S., and Minami, T., "Identification of human faces based on isodensity maps",
Pattern Recognition, Vol. 24(3), pp. 263-272, (1991).
[4] Yuille, A. L., Cohen, D. S., and Hallinan, P. W., "Feature extraction from faces using deformable
templates", Proc. of CVPR, (1989).
[1] M. Turk and A. Petland, “Eigenfaces for Recognition”, Journal of Cognitive Neuroscience, Vol. 3, No.
1, March 1991, pp. 71-86.
[2] M. Turk and A. Petland, “Face Recognition Using Eigenfaces”, Proc. of the IEEE Conf. on Computer
Vision and Pattern Recognition, June 1991, pp. 586-591.
[3] B. Gunturk, A. Batur, Y. Altunbasak, M. Hayes, and R. Mersereau, “Eigenface-Domain Super-
Resolution for Face Recognition”, IEEE Transactions on Image Processing, Vol. 12, No. 5, MAY 2003, pp.
597-606;
[4] K. Delac, M. Grgic, and P. Liatsis, “Appearance-based Statistical Methods for Face Recognition”, 47th
International Symposium ELMAR-2005, 08-10 June 2005, Zadar, Croatia.
[5] T. Heseltine, N. Pears, J. Austin, and Z. Chen, “Face Recognition: A Comparison of Appearance-Based
Approaches”, Proc. VIIth Digital Image Computing: Techniques and Applications, Sun C., Talbot H.,
Ourselin S. and Adriaansen T. (Eds.), 10-12 Dec. 2003, Sydney.
