This file explains the role of each necessary matlab file in this
directory for signature identification/verification

It may be adapted for any other type of signal.

A Quick example: Assuming that we want to learn from the first 25 files
of each user (uers are specified in the file readusr.m) and the rest of files
to use them for testing. Also for computing the FRR and FAR, we need to specify
the range of thresholds

M=25;
Learn(M);
getgenuinescore(M+1);
getimpostorscore;
theta=-20:0.1:10;
eer=EER(theta, 1);

If you want play ...
identify('aab',15)   <- identify un uknown who provides a signature (from aab
                     <- the 15th signature
verify('aab', 'ygs', 15)  <- verify if the unknown user ('aab', with signature
                          <- 15th is really the user 'ygs'

In case the users are modified we need to touch file: readusr.m
In case the directory of the data is modified we need to touch file: readdir


1. Run then Learn to build each user's model (GMM) and the world model
   % specify the number of files you want for learning 1:M (sequentially)  
   Learn(M);
      
        This function calls:
		readusr.m
                getfeat.m
                    readdir.m
                    readusr.m
                    jonarpar.m    <- this the function that computes the features
                the system to run: vq and em
                CreatWorld.m

2. For testing your approach, you have now to compute the EER and show the DET 
    (a kind of ROC) curve 

   2a. Compute the genuine scores and save them in a file
   getgenuinescore(M+1);

          Calls:
                readusr.m
                locatef.m
                   readdir.m
                   locateu.m
                mlike.m
                   jonaspar.m
                   readgf.m
                   calls the system to run logL
 
   2b. Compute the impostor scores and save them in a file
   getimpostorscore;

          Calls:
                readusr.m
                locatef.m
                   readdir.m
                   locateu.m
                mlike.m
                   jonaspar.m
                   readgf.m
                   calls the system to run logL
 

   2c. compute Equal Error Rate (EER) and plot DET-curve
   eer=EER(theta, 1);