Facial Recognition Using Eigen Faces

advertisement
Facial Recognition Using Eigen Faces [1]
Introduction
Face Recognition is one of the popular areas of research in Image processing. Face recognition has
potential applications in engineering and medical fields. Face Recognition can be used for:
1. Person Identification
2. Surveillance
3. Human-Computer interaction
Face recognition is different from other image detection and recognition because of the fact that faces are
complex, multidimensional and that almost all faces have a similar overall configuration. Face
recognition involves the following steps:
1. Facial location detection
2. Feature extraction
3. Facial Image Classification.
One of the issues associated with facial recognition is to determine the features to be extracted. Some
popular approaches use specific individual features from a single image, but these approaches can
produce incorrect results when multiple views are presented. An approach to solve this problem would be
to work on several images rather than a single image [1]. i.e., we attempt to extract the relevant
information in a face image, encode it and compare it with a database of existing encoded models. The
steps involved could be described as:
a) Initialization
Acquire the training set and calculate eigenfaces (using PCA projections) which define
eigenspace.
b) When a new face is encountered, calculate its weight.
c) Determine if the image is face.
d) If yes, classify the weight pattern as known or unknown.
e) (Learning) If the same unknown face is seen several times incorporate it into known faces.
Implementation
Implementation of the above proposed method would involve
a) Calculating the Eigenfaces [9]
1. Prepare a training set. The faces constituting the training set T should be already prepared for
processing.
2. Subtract the mean. The average matrix A has to be calculated and subtracted from the original in
T. The results are stored in variable S.
3. Calculate the covariance matrix.
4. Calculate the eigenvectors and eigen values of this covariance matrix.
5. Choose the principal components.
Mathematically, we wish to find the Principal Components of the distribution of the faces. This can be
achieved by a method know as Principal Component Analysis. Principal component analysis (PCA) is a
vector space transform often used to reduce multidimensional data sets to lower dimensions for
analysis.[10]
b) Using the eigenfaces to detect faces [1] [4]:
Image faces when projected onto the face space do not change radically, but non-face images do. This
fact is used to recognize the presence of faces in an image. The distance from the face space is calculated
at every point in the image. The points where the distance is very minimal represents a face.
c) Using the Eigenfaces to classify a facial image
The image, containing the face to be recognized, is projected onto the face space to obtain the weights
associated with the eigenfaces. These weights are compared with the weights of the known faces(obtained
during the training phase). This comparison is based on the Euclidean distance between the image
projection and known projections. The face image is then classified as the face with the minimum
Euclidean distance.
References
[1] M. Turk, A. Pentland. “Face Recognition Using Eigenfaces”, In Proc. IEEE Conf. on Computer
Vision and Pattern Recognition, pp. 586-591, 1991
[2] Pranab Mohanty, Sudeep Sarkar, Rangachar Kasturi, "From Scores to Face Templates:A Model-Based
Approach",IEEE Transactions On Pattern Analysis And Machine Intelligence, Vol. 29, NO. 12,
December 2007
[3]Xudong Jiang, Bappaditya Mandal, Alex Kot."Eigenfeature Regularization and Extraction in Face
Recognition", IEEE Transactions On Pattern Analysis And Machine Intelligence, Vol. 30, NO. 3, March
2008
[4]M. Turk, A. Pentland. “Eigenfaces for Recognition”, Journal of Cognitive Neuroscience, Vol. 3, pp.
72-86, 1991
[5]R.
Cutler.
“Face
Recognition
citeseer.nj.nec.com/456378.html, 1996
Using
Infrared
Images
and
Eigenfaces”,
[6]A. Pentland, B. Moghaddom, T. Starner. “View-Based and Modular Eigenfaces for Face Recognition”,
Proc. of IEEE Conf. on Computer Vision and Pattern Recognition (CVPR'94), 1994
[7]P. Belhumeur, J. Hespanha, D. Kriegman. “Eigenfaces vs. Fisherfaces: Recognition Using Class
Specific Linear Projection”, IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI19
(7), pp. 711-720, 1997
[8]Thomas Heseltine, Nick Pears, Jim Austin. "Evaluation of image pre-processing techniques for
eigenface based face recognition", The Proceedings of the Second International Conference on Image and
Graphics, SPIE vol. 4875, pp. 677-685 (2002)
[9]http://en.wikipedia.org/wiki/Eigenface
[10] http://en.wikipedia.org/wiki/Principal_components
Download