EE368 PROJECT PROPOSAL
----------------------

Title : Male/Female Face Recognition

People: Jeffrey DiCarlo ( dicarlo@white.stanford.edu )
        Juan Santos     ( jmsantos@stanford.edu      )

Description:

The main goal of our project is to build an algorithm for determining the 
gender of a person based on an image of the person's face (frontal view) 
that is taken in a controlled environment.  We will begin the project by 
collecting or taking many pictures of different people's faces.  A 
majority of the images obtained will be used for training and the 
remainder of the images will be used for testing the accuracy of the 
developed algorithm.  Secondly, we will build a lower-dimensional 
representation of the face images using the eigenfaces technique.  Next, 
we will look at different ways of partitioning the lower-dimensional 
"face" space into two categories: male and female.  We plan on first 
looking a linear techniques and if there is time non-linear techniques. 
Finally, we will determine the accuracy of the partitioning techniques
by using face images that were not a part of the original training set.