An Ecological Approach to Measuring User Experience (UX) from Facial Expressions

Abstract

: A system for assessing UX issues automatically is proposed in this paper. The facial behavior of an individual performing a specific activity is tracked in real-time with software that tracks facial motion features. Evaluated with the conventional studies, this approach has several advantages: ease of deployment in the user's natural setting; avoidance of invasive devices; and severe cost minimization. An evaluation of the user experience of the system was conducted using 144 videos that showed 12 users executing three tasks on four commercial media players. To predict the presence/absence of UX issues based on the tracker's features, we used different machine learning algorithms. We show promising outcomes that open up opportunities for automated real-time UX estimation in an environmental context