The New Jersey Institute of Technology's
Electronic Theses & Dissertations Project

Title:	A method of tracking eye motion for human-computer interface
Author:	Maranski, Martin
View Online:	njit-etd2000-082 (x, 54 pages ~ 3.6 MB pdf)
Department:	Department of Electrical and Computer Engineering
Degree:	Master of Science
Program:	Computer Engineering
Document Type:	Thesis
Advisory Committee:	Malik, Raashid Ahmed (Committee chair) Shi, Yun Q. (Committee member) Shih, Frank Y. (Committee member)
Date:	2000-05
Keywords:	Image processing --Digital techniques. Neural networks (Computer science).
Availability:	Unrestricted
Abstract:	It would be of value to create another device for interfacing with a computer besides the commonly used handheld devices; one option is to determine where the user is looking on the screen. The major obstacle to overcome in this task is to be able to reliably track the motion of the iris and pupil in the eye. A method to do this is revealed here, using a variation of the Rough transform to detect circular patterns in a close shot of the eye. The algorithm was developed in Matlab and analyzed over a series of 20 sets of 6 frames each; this analysis reveals that it is capable of producing 88.33 percent accuracy under normal conditions. The other obstacles to be overcome are also specified, and possible solutions outlined, with suggestions as to hardware implementation and applications for the handicapped and for a convenient hands-free interface method.