首页> 外文会议>IEEE/SICE International Symposium on System Integration >Look at you, look at me: Detection and analysis of mutual gaze convergence in face-to-face interaction
【24h】

Look at you, look at me: Detection and analysis of mutual gaze convergence in face-to-face interaction

机译:看着你,看着我:对面对面相互作用的相互凝视会聚的检测和分析

获取原文

摘要

In human communication, mutual attentiveness between individuals is essential for social interaction. The most significant and reliable indicator of mutual attentiveness is gaze. Despite the availability of diverse exorbitant eye-tracking devices, all claim highly accurate gaze positioning, which exceeds the necessity for average researches and applications. We propose an alternative model for mutual gaze convergence detection using economical video capturing devices and straightforward techniques that can investigate natural gaze behaviors in human-human interaction. We conducted a task supposing lecture in face-to-face interaction and applied our model to investigate (1) average percentage of total mutual gaze convergence, (2) duration of mutual gaze convergence, (3) the domination relationship between lectures and students during short lectures, including (4) the effects of attention span, which claimed to last for 10???20 minutes, toward short lecture task (5 minutes). Our results showed that the average total mutual gaze convergence between the lecturer and the student is 52.83%. It reveals the fact that in both long and short lecture, humans have momentary gaze fixation at a particular direction. We also found out that attention span has effect on different durations of lecture task, resulting in different dominance between lecturer and student, be it student for long lecture and lecturer for short lecture. We believe that our developed model implementation and findings exhibit practical applicability and worthy of contribution to the communities of related research areas.
机译:在人际交往,个人之间的相互注意力是社会互动是必不可少的。相互注意力的最显著和可靠的指标是目光。尽管不同的过高眼睛跟踪设备的可用性,所有权利要求高度准确的视线定位,这超过了平均研究和应用的必要性。我们建议使用经济的视频采集设备和简单的技术,可以探讨人与人的互动自然的凝视行为的相互凝视会聚检测的替代模式。我们进行了一项任务,脸对脸的互动假设演讲并应用于我们的模型来研究(1)总的相互凝视收敛的平均百分比,(2)的相互凝视收敛的时间,(3)在讲课和学生之间的支配关系短讲座,包括:(4)注意广度的效果,它声称要持续10 ??? 20分钟,朝向短讲座任务(5分钟)。我们的研究结果表明,讲师和学生之间的平均总的相互凝视融合是52.83%。它揭示了一个事实,即在多空双方的讲座,人类必须在一个特定的方向瞬时目光注视。我们还发现,注意力对演讲任务的不同持续时间,导致讲师和学生之间不同的显性效应,无论是学生长期讲座和讲师简短的讲座。我们相信,我们建立的模型的实施和结果表现出实用性,值得相关研究领域的社区的贡献。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号