> for an engineer it was an ultimate challenge
You won't find any argument from me; eye-tracking is extremely hard stuff to do correctly and reliably. There are many things that I still want to do to improve our implementation, but can't find the time to write.
(Insert a shameless request for interested readers to send me their resumes here!)
"There are many things that I still want to do to improve our implementation, but can't find the time to write." - so you did the main part of coding on Gazehawk?
What I did was quite a simple stuff - I combined opensource GazeTracker (http://www.gazegroup.org/downloads/23-gazetracker - that uses OpenCV) with FaceAPI from SeeingMachines. I also experimented with accelerometer module instead of FaceAPI and second webcam but that was a dead end :) In the end after calibration I managed to get about 1-2cm accuracy on 2 test subjects. After my dad came with his smaller eyes and accuracy went dramatically down - I gave up :)