S. Bronte

This is a computer vision system which can estimate the aperture of both eyes from a person. With this information, some drowsiness parameters are measured, focusing on PERCLOS. My main contributions to this system were:

• Improvement of iris characterization
• Reduce computing time enough to reach real-time conditions
• Illumination independence using floodfill techniques

I am no longer working on this project since April 2010, but the system keeps in development.

This project is a part of my Master Thesis, which was included in Pedro Jiménez's PhD Thesis. I would like to thank Pedro for his help during the proccess of integrating all my work in his work.

For more information about this part, go to the master thesis results page.

This is one module of my Master thesis, in which a decimated model based on a scan-laser high resolution model is obtained.

For more information about this part, go to the master thesis results page.

This is my MsC Thesis, and it is basically a face recognition system based on 2DPCA database modified to be robust against illumination changes. This robustness is archieved using histogram equalization applied to the image database. Then, in the classification stage a very primitive coarse to fine clasifier is implemented. Face detection is implemented based on Viola & Jones algorithm.

It was my first contact with Computer Vision problems, so do not expect a lot of it.

First of all, I have to thank Ficosa International for the help at recording the video sequences.

This system is basically based on the difference between the vanishing point and a visual limit, which is obtained using floodfill techniques.

For more information about it, visit publications page.