In-World Interface for Generated Avatar Movements

点击翻译按钮获取中文摘要

Emily Carr University of Art + Design is partnering with Anomotion Interactive Inc. to design an interaction method for the creation and chaining of movement sequences in a virtual reality (VR) environment.Anomotion has created a system for chaining movement sequences in a game. These sequences, recorded manually or using motion capture technology, can be chained using a machine learning algorithm which bridges movements between sequences -- for example when writing on one part of the white board to another, then sitting down. When several motion sequences are chained, a game scene can be acted by an avatar, in real-time, with fewer movement artifacts that might bring a game user out of immersion.Emily Carr University will work with Anomotion to create a movement and voice interface for the recording of movement sequences and the creation of movement sequence chains. This interface will be used in a virtual reality (VR) environment and will make use of intuitive gestures and voice commands to trigger movements ("write on the white board, here") and chaining (", then tie your shoe"). Other commands will edit the sequences (insert, delete, lengthen) and create new movements through mimicry or puppeteering. This new interface will serve multiple functions, including quick recording of game cut scenes and testing of potential in-game scenarios.The interface will be designed for the Unreal Engine and the HTC Vive virtual reality headset. As much as possible, modular coding practices will be used to ensure future porting of the interface to other platforms.