A concept of unobtrusive method for complementary emotive user profiling and personalization for IPTV platforms

Abstract

Mobile technologies, new interactive applications and the service providers' customer-centric approach are influencing the way of assessing QoE nowadays. Traditional QoE assessment methods proved to be effective when dealing with legacy audio/video services; however, current IPTV services provide features beyond traditional TV and are not limited to delivering audiovisual content but may also rely on auxiliary services (e.g. content recommendation). Personalization mechanisms that learn instantaneous user-context relation are interesting extension of the QoE parameters enabling improved experience customization. This paper is focused on the QoE-context relation for context-aware IPTV platforms offering personalized TV experience. The latter systems are in the scope of the UP-TO-US project which is treated in this paper as a reference project dealing with user experience and IPTV. Authors define a QoE architecture for validating traditional subjective assessment methodologies (e.g. based on human visual system modeling, or standardized methodologies like ITU-T BT.500-11) by adopting additional context characteristics - user emotions. Moreover the proposed QoE module is aligned with the architecture defined in the UP-TO-US. In the proposed approach to affective QoE authors foresee important role for learning algorithms that can be applied in order to build a user model (an agent reasoning on QoE based on the gathered knowledge about user-content relation). © 2013 Springer-Verlag.