When users of computer systems are given the opportunity to provide feedback on their preferences or interests in most cases they are presented with a questionnaire to be filled. However, this means of interaction requires an additional cognitive step, since the user is required to rationalize feelings and attitudes by answering predetermined questions; as a result, users often skip this part or answer hastily, depriving systems of valuable criticism ¡from the field¢. In this paper, we present a system which relates data extracted from the posture and movement of the user¢s head and eye gaze with states related to interest and engagement. This system is deployed in the context of Human Computer Interaction, where users read documents from a computer screen and provides non-verbal input which, in turn, can be further processed.