Computing the Face Syntax of Social Communication

Lead Research Organisation: University of Glasgow
Department Name: School of Psychology

Abstract

One of the most important tools for social communication is the face. With its complex structure comprising variations of movements (expressions), morphology (shape/structure) and complexion (colour/texture) it can elicit myriad social judgements and be used manipulate the sender's message. Yet, little is known about how these face signals are composed and perceived because the face is a highly complex visual stimulus. Here, we will address this knowledge gap by exploring the syntactical basis of face signalling using innovative methods that combine social/cultural psychology, 3D dynamic computer graphics, vision science psychophysical methods, and mathematical psychology. Our aim is to deliver new fundamental knowledge of face signalling and a new theoretical framework of face signalling. In addition, we aim to transfer this syntax to social robots because virtual agents are it becoming increasingly important part of human communication in today's society. By transferring psychological knowledge of face signals to social robots we anticipate that our results will improve human-robot interacts, with tremendous opportunities to impact real-life scenarios such as in schools or for the elderly.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
ES/P000681/1 01/10/2017 30/09/2027
1944569 Studentship ES/P000681/1 01/10/2017 31/10/2021 Laura Hensel
 
Description Through mapping the specific facial features (e.g., changes in different areas of the face) associated with each of four trait dimensions (dominant-submissive, competent-incompetent, trustworthy-untrustworthy, warm-cold), we built a library of common and unique features between these traits, showing that while some of them may overlap, they cannot be equated. Not only do these findings have implications for social trait-perception theories but they can have wider reaching impact, for example, in social robotics where faces need to convey specific traits to engage the human user. In addition, this work has honed my skills in data-driven methods and handling high-dimensional data-sets (each face's features are described by 30,000+ 3D shape vertices). Our findings also open up the question about the minimal amount of trait-specific facial information necessary for a given trait to be perceived. In addition, we can now add to the literature on the overgeneralisation of emotional expressions to social trait perception (e.g., happy looking faces are perceived as trustworthy/warm) through statistically modelling the overlap between different trait features and emotional expression features in the same 3D face space.
Exploitation Route The most direct and significant impact of this work is that on the field of social robotics. Social robots interact with human users on a daily basis and need to be equipped with facial features that helps engage the user and allows for natural interaction. The next step towards this, is to implement our social trait features on the social robot head Furhat and investigate how this changes interactions between the social robot and the user. We will do this in collaboration with Furhat Robotics, Sweden.
In the research domain, these results inform western theories of social trait perception and may now be investigated in a broader cultural context as well as different social settings.
Sectors Communities and Social Services/Policy,Digital/Communication/Information Technologies (including Software),Electronics

 
Description Nuffield Student Placement 
Form Of Engagement Activity Participation in an open day or visit at my research institution
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Schools
Results and Impact We hosted 2 pupils during their Nuffield Student placement in our lab. They both were actively involved in our ongoing research, gaining insight into the research process from conception to execution and analysis.
Year(s) Of Engagement Activity 2018,2019
 
Description Pint of Science 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Public/other audiences
Results and Impact Pint of Science is a nation wide event hosted every year in Glasgow. Between 100 and 200 audience members of the general public attended the specific event which I helped organise. This included talks by scientists from around Glasgow, knowledge games, and discussions with the audience.
Year(s) Of Engagement Activity 2019