Gaze and Peripheral Vision Analysis for Human-Environment Interaction: Applications in Automotive and Mixed-Reality Scenarios
M. Moniri. Fachrichtung Informatik, Universität des Saarlandes, Saarbrücken, Dissertation, (February 2018)
This thesis studies eye-based user interfaces which integrate information about the user’s perceptual focus-of-attention into multimodal systems to enrich the interaction with the surrounding environment. We examine two new modalities: gaze input and output in the peripheral field of view. All modalities are considered in the whole spectrum of the mixed-reality continuum. We show the added value of these new forms of multimodal interaction in two important application domains: Automotive User Interfaces and Human-Robot Collaboration. We present experiments that analyze gaze under various conditions and help to design a 3D model for peripheral vision. Furthermore, this work presents several new algorithms for eye-based interaction, like deictic reference in mobile scenarios, for non-intrusive user identification, or exploiting the peripheral field view for advanced multimodal presentations. These algorithms have been integrated into a number of software tools for eye-based interaction, which are used to implement 15 use cases for intelligent environment applications. These use cases cover a wide spectrum of applications, from spatial interactions with a rapidly changing environment from within a moving vehicle, to mixed-reality interaction between teams of human and robots.