An image based non-verbal behaviour analysis of HRI

Acceptance and engagement of users towards social robots is the yardstick by which the success and efficacy of Human-Robot Interaction (HRI) is measured. Metrics such as acceptance and engagement are normally measured through subjective (such as interviews) or objective measures (such as non-verbal behaviour). In this study, we report on a methodologically novel process of monitoring non-verbal behaviour of humans in visual images with a social robot. We qualitatively code pictures of humans with the Nao robot as found on Instagram. We specifically coded the emotions of the users as depicted by their facial expressions as well as their social distance maintained towards the robot. Our results show that in general humans persist with an overall positive, enthusiastic and engaging behaviour towards the Nao. Moreover, females were found to be much more expressive in their responses towards the Nao as compared to males - a gender effect. We conclude with the implications of our results and possible avenues of future research on the topic of measuring human engagement with social robots through visual stimuli.