Here is MyPSP Robotics version 5.2.1 (08th January 2008): 

MyEye implements Image processing using Intel Open Source Computer Vision Library.

Instructions to use "Principal Component Analysis" for face recognition

The folder .\WorkArea\faces is the work area for 
storing training 6 faces per indidual and debugging face recognition from new face input.

On first time, you cannot run " OpenCV PCA face recognition " because there aren't any individual face trained.

Before running "OpenCV PCA face recognition" please let run "OpenCV PCA individual face training"  for at least
two different individual faces or persons.

When running the training, please enter the name of the person using the On Screen Keyboard.

Any training on each individual face consists of detecting 6 time the frontal face possibly with different expressions.
At the end of the six detected faces MyEye voice will notice the end.
 
When at least two different people have been stored, PCA001.ipl PCA002.ipl etc... will be stored into
.\WorkArea\faces .

Now  with at least two or more individual sets of frontal faces stored you can run "OpenCV PCA face recognition".

A new face will be detected and the PCA statistical test will be performed to calculate the closer trained individual
face and a message by voice and on displey will give evidence of the closed result.

It is not a full proof recognition test, but it is a debugging procedure.

That's why you can find some files into .\WorkArea\faces with step by step results.

If your data becomes corrupted for some reason or if you want to repeat from scratch the tests, let delete
all files from .\WorkArea\faces .

In case of corrupted files or data the procedure could crash unexpectedly. In case, delete all files in
.\WorkArea\faces  and restart training and recognition in sequence.

Let enjoy
 