GestureCanvas: A Programming by Demonstration System for Prototyping Compound Freehand Interaction in VR

As the use of hand gestures becomes increasingly prevalent in virtual reality (VR) applications, prototyping Compound Freehand Interactions (CFIs) effectively and efficiently has become a critical need in the design process. Compound Freehand Interaction (CFI) is a sequence of freehand interactions where each sub-interaction in the sequence conditions the next. Despite the need for interactive prototypes of CFI in the early design stage, creating them is effortful and remains a challenge for designers since it requires a highly technical workflow that involves programming the recognizers, system responses and conditionals for each sub-interaction. To bridge this gap, we present GestureCanvas, a freehand interaction-based immersive prototyping system that enables a rapid, end-to-end, and code-free workflow for designing, testing, refining, and subsequently deploying CFI by leveraging the three pillars of interaction models: event-driven state machine, trigger-action authoring, and programming by demonstration. The design of GestureCanvas includes three novel design elements — (i) appropriating the multimodal recording of freehand interaction into a CFI authoring workspace called Design Canvas, (ii) semi-automatic identification of the input trigger logic from demonstration to reduce the manual effort of setting up triggers for each sub-interaction, (iii) on the fly testing for independently validating the input conditionals in-situ. We validate the workflow enabled by GestureCanvas through an interview study with professional designers and evaluate its usability through a user study with non-experts. Our work lays the foundation for advancing research on immersive prototyping systems allowing even highly complex gestures to be easily prototyped and tested within VR environments.