Digital Single Market
Digital Economy & Society

ICT 2010

PetaMedia: SpudTV: Affective Content Recommendation

This is a technology or research stand, located in Zone R1 - Smart Systems. Stand number: R1-13

SpudTV research is making it possible for users to find and sample music videos with minimal interaction with their TV or computer.

The focus of this project is on the development of an automatic and affective music-video recommendation system. Such recommendations are based on the user’s physiological responses combined with a collaborative filtering approach based on the user’s listening habits. The system estimates valence, arousal and like/dislike measures, which will steer the collaborative filtering. In other words, if the system detects that the user likes the video which is currently playing, the next video to be recommended will be similar. But if, on the other hand, the user does not like the video being played, the recommendation will be steered towards music with a higher reported valence measure.

Technical description

Coordinator: Ashkan YAZDANI (EPFL, Ecole Polytechnique Federale de Lausanne, Switzerland)

ID: 3243