Suppr超能文献

多模态面对面互动中的面部信号与社交行为

Facial Signals and Social Actions in Multimodal Face-to-Face Interaction.

作者信息

Nota Naomi, Trujillo James P, Holler Judith

机构信息

Donders Institute for Brain, Cognition, and Behaviour, 6525 AJ Nijmegen, The Netherlands.

Max Planck Institute for Psycholinguistics, 6525 XD Nijmegen, The Netherlands.

出版信息

Brain Sci. 2021 Jul 30;11(8):1017. doi: 10.3390/brainsci11081017.

Abstract

In a conversation, recognising the speaker's social action (e.g., a request) early may help the potential following speakers understand the intended message quickly, and plan a timely response. Human language is multimodal, and several studies have demonstrated the contribution of the body to communication. However, comparatively few studies have investigated (non-emotional) conversational facial signals and very little is known about how they contribute to the communication of social actions. Therefore, we investigated how facial signals map onto the expressions of two fundamental social actions in conversations: asking questions and providing responses. We studied the distribution and timing of 12 facial signals across 6778 questions and 4553 responses, annotated holistically in a corpus of 34 dyadic face-to-face Dutch conversations. Moreover, we analysed facial signal clustering to find out whether there are specific combinations of facial signals within questions or responses. Results showed a high proportion of facial signals, with a qualitatively different distribution in questions versus responses. Additionally, clusters of facial signals were identified. Most facial signals occurred early in the utterance, and had earlier onsets in questions. Thus, facial signals may critically contribute to the communication of social actions in conversation by providing social action-specific visual information.

摘要

在对话中,尽早识别说话者的社交行为(如请求)可能有助于后续潜在的说话者快速理解其意图信息,并及时规划回应。人类语言是多模态的,多项研究已经证明身体对交流的作用。然而,相对较少的研究探讨了(非情感性的)对话面部信号,对于它们如何促进社交行为的交流知之甚少。因此,我们研究了面部信号如何映射到对话中两种基本社交行为的表达上:提问和提供回应。我们研究了12种面部信号在6778个问题和4553个回应中的分布和时机,这些信号在34组荷兰语面对面二元对话语料库中进行了整体标注。此外,我们分析了面部信号聚类,以找出问题或回应中是否存在面部信号的特定组合。结果显示面部信号比例很高,在问题和回应中的分布在质上有所不同。此外,还识别出了面部信号簇。大多数面部信号在话语早期出现,在问题中出现的起始时间更早。因此,面部信号可能通过提供特定于社交行为的视觉信息,对对话中社交行为的交流起到关键作用。

https://cdn.ncbi.nlm.nih.gov/pmc/blobs/5680/8392358/ba0ab02faefe/brainsci-11-01017-g0A1.jpg

文献检索

告别复杂PubMed语法,用中文像聊天一样搜索,搜遍4000万医学文献。AI智能推荐,让科研检索更轻松。

立即免费搜索

文件翻译

保留排版,准确专业,支持PDF/Word/PPT等文件格式,支持 12+语言互译。

免费翻译文档

深度研究

AI帮你快速写综述,25分钟生成高质量综述,智能提取关键信息,辅助科研写作。

立即免费体验