基于瞳孔检测的注视点跟踪系统

Gaze Tracking System Based on Pupil Detection

介绍了一种基于瞳孔检测的非强迫式注视点跟踪系统。该系统对被测对象观察不同注视点时的眼睛进行成像,通过图像处理得到瞳孔在眼睛中的相对位置数据,从而计算出所观察的注视点的位置。采用了二维定标算法进行系统定标,提出了一种角度修正的方法以提高系统精度。实验结果表明,在相同的实验条件下,系统估算点与实际观测点基本一致,视角平均误差为2.5°,可以满足一般对注视点测定的要求。

A non-compulsory gaze tracking system based on pupil detection was introduced. The system takes the image of the eye of the tested subject when he is gazing at a position, and processes the eye image to get the relative position of the pupil in the eye, then calculates out the gaze position where he is gazing at. A two-dimension calibration algorithm is employed to calibrate the system, and an angle correction method is proposed to improve the precision of the system. The experimental results under the same conditions show that the calculated gaze positions by the system mostly agree with the real gaze positions,the average error of the view angle is 2.5°,so the system can meet the requirement of common gaze tracking.