Computational modeling of gaze behaviour in real-world settings
MetadataShow full item record
Understanding Human gaze behavior is of value from basic science to a myriad of applications. Current technology places constraints on possible use cases in requiring that much of this analysis is restricted to a lab setting, with constraints imposed by limitations of the hardware setup required. The objective of this thesis is to establish methods that allow for collection and analysis of gaze data in real-world settings, and allow for development of computational models of human gaze strategies that are more representative of real-world behaviour. To achieve these objectives, the thesis presents methods for deriving rich models of the real world in the form of 3D models that provide a base representation that gaze direction may be measured against, and strategies for associating experimental eye-tracking data with these models derived from unconstrained real-world settings. The thesis also presents techniques that address qualitative and quantitative analysis, and visualization of gaze data associated with 3D models. As a whole, the body of techniques presented provides a foundation for future research efforts in real-world gaze analysis, presenting many new opportunities for experimental studies, and computational modeling efforts.