Privacy Preservation in Affect-Driven Personalization

Knowing exactly how a given audience feels about some content in addition to what mood they are in, can affect the type of user experience that is delivered in the form of personalized advertising and content delivery. As we continue to mine rich data sources and draw insights from emotion analytics, we often leave in our wake, a major privacy gap pertaining to the human subjects involved in our studies. Yes, the goal is to derive insights from a user's emotional state in a bid to select personalized content or predict what the end user is likely to find attractive at a given point in time. Yet, this has to be done in a way that is ubiquitous and above all preserves the privacy of the subjects in question. This literature details the visual privacy gaps in collecting and analyzing emotion data in support of personalized advertising and, consequently, offers a reference framework for preserving end-user privacy throughout the emotion analytics lifecycle.