A New Calibration-Free Gaze Tracking Algorithm Based on DE-SLFA

摘要

Advanced remote gaze estimation systems use automatic calibration procedure without requiring active user involving into the estimation of subject-specific eye parameters. Though automatic calibration process can simplify the difficulty of calibration task, it still needs time to collect information for completing the eye parameters of users before the gaze tracking system is used. This paper proposes a novel method, free of calibration procedure to extract subject-specific eye parameters. To estimate the real-time angles between the optical and visual axes of each eye before calculating the direction of the visual axes of the both the left and right eyes, differential evolution and Shuffled Frog-leaping Algorithm (DE-SLFA) is used to minimize the distance between the intersections of the visual axes of the left and right eyes with the surface of a display while subjects look naturally at the display. As a consequence, the inconvenient calibration procedure which may produce possible calibration errors can be eliminated. Computer simulation have been performed to confirm the proposed method.