A New Method for Calibration of Kappa Angle in 3D Line of Sight Tracking System

Abstract

At present, the general idea of the 3D line of sight (LoS) estimation is as follows. First, the direction of the 3D line of gaze (LoG) can be reconstructed according to the visual characteristics of the eyeballs (pupil center coordinates, purkinje images coordinates, etc.). Then the transformation matrix of the direction of LoG and the direction of LoS can be calculated based on the Kappa angle between LoG and LoS. LoS can be estimated with known LoG. The transformation matrix between LoG and LoS is usually determined by the user calibration process of the gaze tracking system. Because the eyeball structure is unique and the Kappa is a space angle, the transformation matrix changes with the space location of the eyeball. This paper elaborates on the Kappa angle calibration problem in the gaze tracking system. A new calibration and calculation method of Kappa angle is proposed in this paper, which can solve the calculation of space Kappa angle in the case of head translation, pitch and rotation. Simulation and experimental results verify the effectiveness of the proposed method.