Amanda Buczkowski’s Webpage:

Amanda

·        Attends Oakland University

·        Is a senior

·        Plans to graduate in May 2003

·        Major is Computer Science

 

My Interests

·        Computer programming

·        Knows C++ and Visual Basic

·        Plan to pursue a Master’s degree

 

My Hobbies: 

·        Playing video/computer games

·        Watching hockey

·        Shopping

·        Hanging out with friends

 

Congratulations to the 2002 Stanley Cup Champs: THE DETROIT RED WINGS!!!

 

 

 

Favorite game system: Playstation 2

 

 

 

Favorite Car:  New Beetle

 

 

Research Topic: 

 

Eye Gaze Tracking

 

The focus of my project has been to locate the eyes of a person given their picture.  To do this, I first use a simple skin detection algorithm to find the skin tone regions.  (For my project, I have assumed that the background contains no skin tone regions.  Next, I perform some image processing techniques on the face region.  These techniques include zero-crossing, morphological operations, intensity filters, and small and large size filters.

Next, I determine the eye regions based on the amount of overlap each region has with every other region.  Regions in the top ten percent and bottom forty percent of the picture are deleted as the eyes do not occur in those regions of the face.  After determining where the eyes are, I crop the eyes and enlarge them.  Then, I separate the eyes and crop them.

 

SCREENSHOTS OF MY PROGRAM:

 

Original Picture:

 

 

After skin tone filter:

 

 

Face location determined:

 

 

Cropped Face:

 

 

After zero-crossing and morph operation:

 

 

After intensity filter:

 

 

Possible eye candidates found

(after performing other image processing techniques)

lines indicate region to be searched

 

 

Possible eye candidate after deleting regions not in search area

 

 

Eye locations found:

 

 

Box draw around eye locations in original picture:

 

 

Enlarged Eyes:

 

 

Left Eye:

 

 

Right Eye: