20-12-2012, 06:31 PM
FACE RECOGNITION IN BIOMETRICS
[attachment=44654]
Introduction to Biometrics
The average adult working in a large business has 12 passwords to remember, and spends nearly a week in every year logging into systems.
The average cost to a large company for every password lost is $16.
FACE RECOGNITION
Importance
Security
Surveillance
This is mainly preferred over other biometrics as it can be carried out using available resources. Face recognition is a task performed by human with ease in day to day life. Human beings can recognize faces by considering many features such as view face at different angle, skin color, hair information, facial mark etc and tries to
compare with different expression.
Eigenfaces
Modeling
Given a collection of n labeled training images,
Compute mean image and covariance matrix.
Compute k Eigenvectors (note that these are images) of covariance matrix corresponding to k largest Eigenvalues.
Project the training images to the k-dimensional Eigenspace.
Recognition
Given a test image, project to Eigenspace.
Perform classification to the projected training images.
RESULT:
Classification is performed by comparing the feature vector(weight matrix) of the images in the training set with the feature vector (weight matrix) of the test image using Euclidean Distance.
[attachment=44654]
Introduction to Biometrics
The average adult working in a large business has 12 passwords to remember, and spends nearly a week in every year logging into systems.
The average cost to a large company for every password lost is $16.
FACE RECOGNITION
Importance
Security
Surveillance
This is mainly preferred over other biometrics as it can be carried out using available resources. Face recognition is a task performed by human with ease in day to day life. Human beings can recognize faces by considering many features such as view face at different angle, skin color, hair information, facial mark etc and tries to
compare with different expression.
Eigenfaces
Modeling
Given a collection of n labeled training images,
Compute mean image and covariance matrix.
Compute k Eigenvectors (note that these are images) of covariance matrix corresponding to k largest Eigenvalues.
Project the training images to the k-dimensional Eigenspace.
Recognition
Given a test image, project to Eigenspace.
Perform classification to the projected training images.
RESULT:
Classification is performed by comparing the feature vector(weight matrix) of the images in the training set with the feature vector (weight matrix) of the test image using Euclidean Distance.