Eye-Controlled Cursor Stabilization Technique Based on Optimal Estimation

Abstract In this study, we propose a stabilization algorithm based on optimal estimation to improve the stability of the eye-controlled cursor. The main principle is that the optimal result of the cursor state at the previous moment is used to calculate the predicted value of the current cursor state (a priori estimate). Then, the actual state of the cursor given by the eye tracker is used to correct the predicted value so as to move the cursor smoothly and stably. To verify the performance of the proposed eye-controlled cursor stabilization technique, we designed two sets of control experiments to compare our technique with two classical eye-controlled cursor stabilization techniques: manual and gaze input cascaded and warping to target centre. The experimental results show our stabilization algorithm significantly improves the performance speed of the eye-controlled cursor and reduces the error rate. It is additionally more accessible for users and has good application potential.