计算机科学 ›› 2019, Vol. 46 ›› Issue (8): 282-291.doi: 10.11896/j.issn.1002-137X.2019.08.047

• 人工智能 • 上一篇    下一篇

基于动态轨迹的眼动跟踪隐式标定方法

程时伟, 齐文杰   

  1. (浙江工业大学计算机科学与技术学院 杭州310023)
  • 收稿日期:2019-04-28 出版日期:2019-08-15 发布日期:2019-08-15
  • 通讯作者: 程时伟(1981-),男,博士,教授,CCF会员,主要研究方向为人机交互、普适计算、协同计算,E-mail:swc@zjut.edu.cn
  • 作者简介:齐文杰(1993-),男,硕士生,主要研究方向为图像处理
  • 基金资助:
    国家自然科学基金项目(61772468)

Dynamic Trajectory Based Implicit Calibration Method for Eye Tracking

CHENG Shi-wei, QI Wen-jie   

  1. (School of Computer Science and Technology,Zhejiang University of Technology,Hangzhou 310023,China)
  • Received:2019-04-28 Online:2019-08-15 Published:2019-08-15

摘要: 针对现有多点标定方案耗时长和简化标定方案注视精度较差的局限,提出一种眼动跟踪隐式标定方法,使得眼动跟踪系统只需采集少量样本即可建立准确的映射关系。该方法分为3个步骤:首先,标定数据采集,让用户视线跟随动态轨迹运动,记录这一过程中用户眼部图像特征和标定点之间的映射点对。然后,提出合理化的异常值去除方法以自动消除样本噪声,并选择最佳点对集合建立映射模型。对眼动跟踪数据的采集进行延时处理,减少了运动轨迹产生的误差。进一步对样本进行降噪时,排除瞳孔误差数据,并采用随机采样一致算法进一步筛选样本。最后,结合免标定和单点标定这两种方法,在后续标定过程中进一步简化隐式标定过程,并测试了隐式标定的最佳参数。实验表明,在视距为60 cm时,该方法的标定时间为8 s,平均精度为2.0°;在隐式标定原型系统中,对于已标定的用户,通过读取其映射模型,即可免标定地快速获取注视点坐标,所需时间为2 s,平均标定精度为2.47°;对于进行眼动跟踪的新用户,通过单点标定方法计算个体差异补偿模型,获取注视点坐标,所需时间为3 s,平均标定精度为2.49°,进一步提高了该方法的实用性。

关键词: 免标定, 人机交互, 眼动跟踪, 隐式标定, 注视点

Abstract: Aiming at the limitations of the existing multi-point calibration schemes,such as time-consuming and poor gaze accuracy of simplified calibration schemes,this paper proposed an implicit calibration method for eye tracking,which makes the eye tracking system only need a few samples to establish an accurate mapping relationship.The me-thod has three steps.Firstly,it collects calibration data.With user’s gaze following the dynamictrajectory,it records the mapping points between the image of user’s eyes and the calibration points in this process.Then,a rationalized outlier removal method is proposed to automatically eliminate the sample noise and select the best point pair to establish a mapping model.The acquisition of eye movement data is delayed,which can reduce the error caused by the dynamic traje-ctory.Furthermore,when the noise data of the sample are removed,a method for eliminating the pupil error data is proposed,and the sample data are further filtered by Random Sample Consensus (RANSAC) algorithm.Finally,the two methods of calibration-free and single-point calibration are combined to simplify the subsequence implicit calibration process.Experiment results show that the average calibration time is 8 s,and the average accuracy is 2.0° of visual angle when visual distance is 60 cm.In the simplified implicit calibration prototype system,for the user who is calibrated,the coordinates of the fixation are obtained by the calibration-free method.The average calibration time is 2 s,and the ave-rage calibration accuracy is 2.47° of visual angle.For the new user who performs eye tracking,the individual difference compensation model is calculated by the single point calibration method to obtain the coordinates of the fixation.The ave-rage calibration time is 3 s,and the average calibration accuracy is 2.49° of visual angle,which further improves the practicality of the implicit calibration method

Key words: Calibration free, Eye tracking, Fixation, Human-computer interaction, Implicit calibration

中图分类号: 

  • TP391
[1]FUHL W,SANTINI T,GEISLER D,et al.Eyes wide open? eyelid location and eye aperture estimation for pervasive eye tracking in real-world scenarios[C]∥Acm International Joint Conference on Pervasive & Ubiquitous Computing:Adjunct.ACM,2016:1656-1665.
[2]MORIMOTO C H,MARCIO R M.Eye gaze tracking techniques for interactive applications.Computer Vision and Image Understanding,2005,98(1):4-24.
[3]SUGANO Y,MATSUSHITA Y,SATO Y.Calibration-free gaze sensing using saliency maps[C]∥Computer Vision and Pattern Recognition.IEEE,2010:2667-2674.
[4]WOOD E,TADAS B,MORENCY L P,et al.A 3D Mor- phable Eye Region Model for Gaze Estimation[C]∥Euro-pean Conference on Computer Vision.Springer InternationalPublishing,2016:297-313.
[5]KHAMIS M,SALTUK O,HANG A,et al.TextPursuits:using text for pursuits-based interaction and calibration on public displays[C]∥Acm International Joint Conference on Pervasive & Ubiquitous Computing.ACM,2016:274-285.
[6]HUANG C W,TSENG S C,JIANG Z S,et al.Projective mapping compensation for the head movement during eye tracking[C]∥IEEE International Conference on Consumer Electronics-taiwan.IEEE,2014:2243-2246.
[7]SUGANO Y,ZHANG X,BULLING A.AggreGaze:Collective Estimation of Audience Attention on Public Displays[C]∥Symposium on User Interface Software and Technology.ACM,2016:821-831.
[8]STEIL J,HUANG M X,BULLING A.Fixation Detection for Head-Mounted Eye Tracking Based on Visual Similarity of Gaze Targets[C]∥Proceedings of the 2018 ACM Symposium on Eye Tracking Research & Applications.New York:ACM,2018:23.
[9]PAIVI M,ANDREAS B.Eye Tracking and Eye-Based Human-Computer Interaction∥Advances in physiological computing.London:Springer,2014: 39-65.
[10]DONGHENG L,DAVID W,DERRICK J P.Starburst:A hybrid algorithm for video-based eye tracking combining feature-based and model-based approaches[C]∥Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.IEEE,2005:2667-2674.
[11]PFEUFFER K,VIDAL M,TURNER J,et al.Pursuit calibration:Making gaze calibration less tedious and more flexible[C]∥Proceedings of the 26th annual ACM symposium on User interface software and technology.ACM,2013:261-270.
[12]SUGANO Y,BULLING A.Self-Calibrating Head-Mounted Eye Trackers Using Egocentric Visual Saliency[C]∥The 28th Annual ACM Symposium.ACM,2015:363-372.
[13]HUANG M X,KWOK T C K,NGAI G,et al.Building a Personalized,Auto-Calibrating Eye Tracker from User Interactions[C]∥Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems.ACM, 2016:5169-5179.
[14]BERNET S,CUDEL C,LEFLOCH D,et al.Autocalibration- based partioning relationship and parallax relation for head-mounted eye trackers.Machine Vision and Applications,2013,24(2):393-406.
[15]SANTINI T,FUHL W,KASNECI E.CalibMe:Fast and Unsupervised Eye Tracker Calibration for Gaze-Based Pervasive Human-Computer Interaction[C]∥Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems.ACM,2017:2594-2605.
[16]SUN F R,LIU J R.Fast Hough Transform Algorithm.Chinese Journal of Computers,2001,24(10):1102-1109.(in Chinese) 孙丰荣,刘积仁.快速霍夫变换算法.计算机学报,2001,24(10):1102-1109.
[17]YANG H.A study of Calibration Techniques on Eye Gaze Tracking System with Head Fixed.Xi’an:Xidian University,2015.(in Chinese) 杨慧.头部静止的视线跟踪系统中标定技术的研究.西安:西安电子科技大学,2015.
[18]KOBAYASHI T,TOYAMA T,SHAFAIT F,et al.Recognizing Words in Scenes with a Head-Mounted Eye-Tracker[C]∥Iapr International Workshop on Document Analysis Systems.IEEE Computer Society,2012:2142-2149.
[19]LIU K,GE J F,LUO Y P.Probability guided random sample consensus.Journal of Computer:Aided Design & Computer Graphics,2009,21(5):657-662.(in Chinese) 刘坤,葛俊锋,罗予频.概率引导的随机采样一致性算法.计算机辅助设计与图形学学报,2009,21(5):657-662.
[20]LAWRENCE H Y,EIZENMAN M.A new methodology for determining point-of-gaze in head-mounted eye tracking systems.IEEE Trans.Biomed.Engineering,2004,51(10):1765-1773.
[21]CERROLAZA J J,VILLANUEVA A,CABEZA R.Study of Polynomial Mapping Functions in Video-Oculography Eye Trackers.ACM Transactions on Computer-Human Interaction (TOCHI),2012,19(2):1-25.
[22]HUANG G B,ZHU Q Y,SIEW C K.Extreme learning ma- chine.Theory and applications .Neurocomputing,2006,70:489-501.
[23]JIANG G Y.Research on data fusion of head movement and gaze tracking.Xi’an:Xi’an Technological University,2017.(in Chinese) 蒋光毅.头部运动与视线追踪数据融合技术的研究.西安:西安工业大学,2017.
[24]ESSIG K,POMPLUN M,RITTER H.A neural network for 3D gaze recording with binocular eye trackers.International Journal of Parallel,Emergent and Distributed Systems,2006,21(2):79-95.
[25]HOU S W.The Research of the Calibration Method in the head-mounted Eye Tracking Systems.Hefei:University of Scien-ce and Technology of China,2013.(in Chinese) 侯树卫.头戴式视线跟踪系统的标定方法研究.合肥:中国科学技术大学,2013.
[26]MORIMOTO C H,MIMICA M R M.Eye gaze tracking techniques for interactive applications.Computer Vision & Ima-ge Understanding,2005,98(1):4-24.
[27]CHENG S W,SUN Z Q,LU Y H.An eye tracking approach to cross-device interaction .Journal of Computor-Aided Design &Computor Graphics,2016,28(7):1094-1104.(in Chinese) 程时伟,孙志强,陆煜华.面向多设备交互的眼动跟踪方法.计算机辅助设计与图形学学报,2016,28(7):1094-1104.
[28]POLA J,WYATT H J.Target position and velocity:The stimuli for smooth pursuit eye movements.Vision Research,1980,20(6):523-534.
[29]ALNAJAR F,GEVERS T,VALENTI R,et al.Auto-Calibrated Gaze Estimation Using Human Gaze Patterns.International Journal of Computer Vision,2017,124(2):223-236.
[1] 李荪, 曹峰.
智能语音技术端到端框架模型分析和趋势研究
Analysis and Trend Research of End-to-End Framework Model of Intelligent Speech Technology
计算机科学, 2022, 49(6A): 331-336. https://doi.org/10.11896/jsjkx.210500180
[2] 张继凯, 李琦, 王月明, 吕晓琪.
基于单目RGB图像的三维手势跟踪算法综述
Survey of 3D Gesture Tracking Algorithms Based on Monocular RGB Images
计算机科学, 2022, 49(4): 174-187. https://doi.org/10.11896/jsjkx.210700084
[3] 朱晨爽, 程时伟.
虚拟现实环境下基于眼动跟踪的导航需求预测与辅助
Prediction and Assistance of Navigation Demand Based on Eye Tracking in Virtual Reality Environment
计算机科学, 2021, 48(8): 315-321. https://doi.org/10.11896/jsjkx.200500031
[4] 刘亮, 蒲浩洋.
基于LSTM的多维度特征手势实时识别
Real-time LSTM-based Multi-dimensional Features Gesture Recognition
计算机科学, 2021, 48(8): 328-333. https://doi.org/10.11896/jsjkx.210300079
[5] 钱基德, 熊仁和, 王乾垒, 杜冬, 王在俊, 钱基业.
边缘计算在飞行训练中的应用
Application of Edge Computing in Flight Training
计算机科学, 2021, 48(6A): 603-607. https://doi.org/10.11896/jsjkx.201000035
[6] 程时伟, 陈一健, 徐静如, 张柳新, 吴剑锋, 孙凌云.
一种基于脑电信号的眼动方向分类方法
Approach to Classification of Eye Movement Directions Based on EEG Signal
计算机科学, 2020, 47(4): 112-118. https://doi.org/10.11896/jsjkx.190200342
[7] 陈甜甜, 姚璜, 左明章, 田元, 杨梦婷.
基于深度信息的动态手势识别综述
Review of Dynamic Gesture Recognition Based on Depth Information
计算机科学, 2018, 45(12): 42-51. https://doi.org/10.11896/j.issn.1002-137X.2018.12.006
[8] 张宁, 刘迎春, 沈智鹏, 郭晨.
虚拟现实技术在专门用途英语教学中的应用研究综述
Review of Virtual Reality Technology Based Application Research in English Teaching for Special Purposes
计算机科学, 2017, 44(Z6): 43-47. https://doi.org/10.11896/j.issn.1002-137X.2017.6A.009
[9] 刘喆,李智.
基于多通道交互技术的计算机辅助需求分析系统的研发
Research and Development of Computer-aided Requirements Engineering Tool Based on Multi-modal Interaction Technologies
计算机科学, 2017, 44(4): 177-181. https://doi.org/10.11896/j.issn.1002-137X.2017.04.039
[10] 陈勇,徐超.
基于符号执行和人机交互的自动向量化方法
Symbolic Execution and Human-Machine Interaction Based Auto Vectorization Method
计算机科学, 2016, 43(Z6): 461-466. https://doi.org/10.11896/j.issn.1002-137X.2016.6A.109
[11] 易靖国,程江华,库锡树.
视觉手势识别综述
Review of Gestures Recognition Based on Vision
计算机科学, 2016, 43(Z6): 103-108. https://doi.org/10.11896/j.issn.1002-137X.2016.6A.025
[12] 辛义忠,马莹,于霞.
直接和间接的笔倾斜输入性能对比
Comparison of Direct and Indirect Pen Tilt Input Performance
计算机科学, 2015, 42(9): 50-55. https://doi.org/10.11896/j.issn.1002-137X.2015.09.011
[13] 高燕飞,陈俊杰,强彦.
Hadoop平台下的动态调度算法
Dynamic Scheduling Algorithm in Hadoop Platform
计算机科学, 2015, 42(9): 45-49. https://doi.org/10.11896/j.issn.1002-137X.2015.09.010
[14] 王玉,任福继,全昌勤.
口语对话系统中对话管理方法研究综述
Review of Dialogue Management Methods in Spoken Dialogue System
计算机科学, 2015, 42(6): 1-7. https://doi.org/10.11896/j.issn.1002-137X.2015.06.001
[15] 何正海,李智.
基于人机交互的计算机辅助软件需求分析工具的研发
Research and Development of Computer-aided Requirements Analysis Tool Based on Human-computer Interaction
计算机科学, 2015, 42(12): 181-183.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!