电子学报 ›› 2018, Vol. 46 ›› Issue (12): 2993-3001.DOI: 10.3969/j.issn.0372-2112.2018.12.024

• 学术论文 • 上一篇    下一篇

基于眼动特征的人机交互行为意图预测模型

梁永强1, 王崴1, 瞿珏1,2, 杨洁1, 刘晓卫1   

  1. 1. 空军工程大学防空反导学院, 陕西西安 710051;
    2. 西北工业大学航空学院, 陕西西安 710072
  • 收稿日期:2017-10-23 修回日期:2018-05-08 出版日期:2018-12-25
    • 通讯作者:
    • 王崴
    • 作者简介:
    • 梁永强 男,1994年生于山西石楼.空军工程大学硕士研究生.研究方向为武器装备人机交互设计、自适应人机界面.E-mail:1102268250@qq.com
    • 基金资助:
    • 国家自然科学基金 (No.51675530)

Human-Computer Interaction Behavior and Intention Prediction Model Based on Eye Movement Characteristics

LIANG Yong-qiang1, WANG Wei1, QU Jue1,2, YANG Jie1, LIU Xiao-wei1   

  1. 1.Air and Missile Defense College, Air Force Engineering University. Xi'an, Shaanxi 710051, China;
    2.School of Aeronautics, Northwestern Polytechnical University. Xi'an, Shaanxi 710072, China
  • Received:2017-10-23 Revised:2018-05-08 Online:2018-12-25 Published:2018-12-25
    • Corresponding author:
    • WANG Wei

摘要: 针对自适应人机界面对用户行为意图预测的需求,提出一种基于眼动特征的人机交互行为分类及意图预测方法.通过建立简化的界面模型,将用户的行为意图分为5类,设计视觉交互实验收集相关行为意图状态下的眼动特征数据,利用SVM(Support Vector Machine)算法建立分类预测模型,结合差异性分析方法选取眼动特征分量,最终确定连续3个采样注视点的位置X坐标、Y坐标、注视时间、眼跳幅度以及瞳孔直径共15个分量作为特征参数可以获得较好的预测效果,其预测精度可达90%以上.

关键词: 自适应界面, 眼动特征, 交互意图, 支持向量机

Abstract: Aiming at the demand of predicting adaptive interface user's intention, This paper presents a method of human-computer interaction behavior classification and intention prediction based on eye movement characteristics. By establishing a simplified interface model, the user's operating behavior is divided into 5 categories, design visual interaction experiment to collect the relevant states' eye movement data. The SVM (Support Vector Machine) algorithm is used to establish the classification prediction model, combined with the difference analysis method to select the eye movement feature component. Finally, the position X coordinate, the position Y coordinate, the gaze time, the eye jump amplitude and the pupil diameter of the 3 consecutive sampling fixation points can be used as the characteristic parameters to obtain the better prediction effect, and the prediction accuracy can reach more than 90%.

Key words: adaptive interface, eye movement characteristics, intention prediction, SVM

中图分类号: