基于视线追踪交互式突显技术对视觉搜索的影响研究

Research on Interactive Highlighting based on Eye Tracking Technology

本研究通过将视线追踪技术与突显技术相结合,在传统突显技术基础上增加动态交互特性,并通过实验考察这种基于视线追踪交互式突显系统对提高海量信息下的视觉目标定位绩效的有效性。首先通过结合低通滤波算法和眼动事件自适应技术,实现基于视线追踪的交互式突显技术,然后让24名大学生分别在有和没有采用基于视线追踪的交互式突显方式下,完成从数字及图标材料的海量刺激中搜索并定位目标刺激的任务,并记录任务完成时间、错误数以及被试的主观疲劳评分。结果发现,自适应低通滤波算法能有效过滤漂移、震颤和微小的不随意眼跳,实现交互式突显系统;无论是在数字材料还是图标材料下,使用交互式突显技术的任务完成时间显著快于未采用该技术,完成错误数上两者无显著差异。在数字材料下,采用交互式突显技术的疲劳度主观评分显著低于没有采用该技术,而在图标材料下,这种差异并不显著。本研究表明基于视线追踪的交互式突显系统能够有效提高海量信息下的目标定位绩效,这种交互式突显技术具有一定的应用前景。

Because of the limit of the physical size of displayer and the bandwidth of human visual system, the resulting bottleneck in raising the efficiency of displayer presents huge amount of information. The traditional Highlighting technology, the Effective Layout technology and the Focus＋ Context technology cannot be adjusted by individual difference and the interactive characteristic, thereby limiting further improvement of the efficiency of information transmission. Although the display mode of the Focus＋ Context technology is interactive, the interactive mode of the traditional mouse and keyboard also limits further efficiency improvement. In this research, by combining highlighting technology with interactive technology to give dynamic characteristics to the highlighting technology, we tried to apply the eye tracking technology to overcome the limit of the interactive mode of mouse and keyboard and make the human-computer interaction more natttral and efficient. The interactive highlighting system based on eye tracking can improve the performance of the visual target localization task when processing large amount of information. In this research, by combining the Low-Pass Filter with the self-adapting technology of the eye movement event, we implemented the interactive highlighting system based on eye tracking. 24 college students completed the visual target localization task involving the mass numeric materials and icon materials with and without using interactive highlighting system; the task time, the number of failures and the fatigue evaluation were recorded. The results showed that the adaptive low-pass filtering algorithm could effectively filter the drift, glissadic and tiny involuntary saccadesand that the interactive highlighting system had been implemented successfully; when the stimuli materials were numeric, the task time using the interactive highlight system was significantly faster than that not using the system（p〈.05） and the fatigue evaluation mark was significantly lower than that not using the system（p〈.05）. However there wasno significant difference in the number of task failures（p〉.05）. Regarding the icon stimuli materials used in the sutdy, similarly, the task time using the interactive highlight system wassignificantly faster than that not using the system（p〈.05）, oHwever, there wasno significant difference in the number of task failures（p〉.05） and the fatigue evaluation mark difference was not significant either（p〉.05）. The results suggest that the interactive highlighting system based on eye tracking can effectively improve the performance of massive visual target localization task.