As I wrote my thesis and created the software for it I had a chance to look
deeper in the theory of face detection and recognition. After I completed the project and
accomplished my goals, I had an application which can run in real-time and detect and
recognize faces robustly and that way it can help to automate tasks.
I looked for different approaches in face detection and I tested them but at the
end I chose the Viola-Jones object detector algorithm because that was the best in all of
them. The same way, I started to search for an accurate face recognition method and I
ran tests to see the results. After the tests, I selected the so called Eigenfaces algorithm
because with that I reached up to 90% accuracy.
With this two, separate module (namely detection and recognition) I created a
system which can generate attendance sheets. With that we can save a lot of time in
different cases. In the generated Excel sheet, we can find who stepped in the door at a
After the system was created I had to make tests on it in order to determine if I
can use it in real life situations. So, I made test scripts for the more complex algorithms
and tested those for accuracy and running time. After I ran all the tests I can say that the
system can be used in real life situations.