Intuitive human-device interaction for video control and feedback
Toon De Pessemier, Luc Martens, Wout Joseph · 2017
Current video services are still controlled in an old-fashioned way using keyboard and mouse for computers and remote control for TV sets. This paper presents more attractive and intuitive interaction methods using commercially available motion sensing input devices. These devices are typically based on a webcam-style add-on peripheral, thereby enabling facial recognition, gesture control, and speech recognition. These technologies are applied in this research to automatically authenticate a user, enable video control (play, pause, seeking), and browsing, selecting, and rating content by hand gestures or voice commands. By monitoring the user's gaze and using emotion recognition techniques, the user's interests and engagement with the content can be estimated. This is interpreted as implicit user feedback for the video content, and establishes an automatic feedback channel, which can be used for content personalization and recommendation. User tests showed an accurate recognition of voice and gestures and confirmed the attractiveness and intuitiveness of these techniques for end-users.