MagicMoves: A Gesture Creation Framework for Virtual Reality Applications
Philipp Michael Markus Peter Asteriou, Jeremiah Diephuis, Philipp Wintersberger · 2023
This paper presents MagicMoves, a novel, non-machine learning-based approach to gesture recognition for Unity, a widely used platform for developing virtual Reality applications. The need for an accessible and customizable gesture recognition system in Unity is emphasized, given the challenges associated with machine learning approaches, such as the requirement for extensive datasets and potential alienation of developers without deep machine learning knowledge. MagicMoves is designed to prioritize minimal overhead and high customization capabilities, allowing developers to record and incorporate their own gestures into their applications. The system is capable of recognizing a wide range of gestures, even those that are overly similar. The paper also discusses the technical requirements for developing such a framework, including gesture editing, gesture recognition, and user interface and integration.