基于标记点检测的视线跟踪注视点估计  

Eye Tracking Gaze Estimation Based on Marker Detection

作　　者：龚秀锋[1] 李斌[1] 邓宏平[1] 张文聪[1] 

机构地区：[1]中国科学技术大学电子科学与技术系,合肥230027

出　　处：《计算机工程》2011年第6期289-290,共2页Computer Engineering

基　　金：国家自然科学基金广东联合基金资助项目(U0835002)

摘　　要：传统的头戴式视线跟踪系统需要借助额外的头部位置跟踪器或其他辅助设备才能定位视线方向。针对该问题,提出一种基于标记点检测的注视点估计方法。该方法通过计算机视觉的方法检测标记点,建立场景图像与真实场景中计算机屏幕之间的空间关系,将场景图像中的注视点坐标映射到计算机屏幕中。实验结果表明,该方法简单易行,可以较好地估计出用户在真实场景中的注视点坐标。For head-mounted eye gaze tracking,additional head position sensors is needed to determine the gaze direction,a new method based on marker detection is proposed to estimate the gaze of point for head-mounted system in this paper.The markers are detected by computer vision method,and the relationship between scene image and computer screen is constructed with point correspondences in two views.The point of gaze in the scene image is translated to computer screen coordinate.Experimental result shows that this method can estimate the point of gaze in real scene easily irrespective of user＇s head position.

关 键 词：视线跟踪 注视点估计 头戴式视线跟踪系统 标记点检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]