Eyegaze tracking method for HCI system

YANG Qinghua, ZHANG Dalei, XUN Yi, BAO Guanjun

(School of Mechanical Engeering, Zhejiang University of Technology, Hangzhou 310014, China)

Abstract: Aiming at the difficulty of interaction with computer for people with severe physical disabilities and special people with mobility impairments, study on an interactive system based on eyetracking was carried out. Wearable systems hardware is introduced as the solution: the CDD camera that connected with the PC is mounted on the spectacles via the carrier made through 3D printing. Through eye diagram capturing, utilizing the improved elipse fitting method to detect the center of the pupil, and calibrating the gaze point based on secondorder model, eventually, the eye gaze tracking and detection in natural light is achieved. The user tests show that the error of eyegaze detection on horizontal and vertical is 3.74% and 5.34% respectively, verifying the feasibility and effectiveness of the proposed interactive eyetracking system.

Key words: eye tracking；human computer interaction(HCI) eyegaze detection