Eye-Tracking Interface based on Artificial Vision for Robot Controlling

A. Úbeda, J.M. Azorín, E. Iañez, and J.M. Sabater (Spain)


Eye-Tracking, Robot Controlling, Man-Machine Interfaces, Computer Vision and Image Processing


This paper is about the development of a vision system which makes possible to know, through effective eye position identification, the direction the user is looking to (eye-tracking) and controlling a robot arm with this information. A CCD camera focuses the face and, with the suitable processing of the captured image, the direction of the eye is identified. A MFC graphic application in C++ has being programmed using specific image treatment libraries of Matrox (MIL). The usage of only one camera supposes a non-intrusive method of eye tracking suitable for disabled people who can only move their eyes and can barely move their head. When the system is correctly configured, eye direction can be robustly detected. These strategies will allow that people with a high degree of disability control a robot in order to helping them in their daily live activities.

Important Links:

Go Back