Search for: All records

We present “Double Doodles” to make full use of two sequential inputs of a VR controller with 9 DOFs in total, 3 DOFs of the first input sequence for the generation of motion paths and 6 DOFs of the second input sequence for motion gestures. While engineering our system, we take ergonomics into consideration and design a set of user-defined motion gestures to describe character motions. We employ a real-time deep learning-based approach for highly accurate motion gesture classification. We then integrate our approach into a prototype system, and it allows users to directly create character animations in VR environments using motion gestures with a VR controller, followed by animation preview and animation inter- active editing. Finally, we evaluate the feasibility and effectiveness of our system through a user study, demonstrating the usefulness of our system for visual storytelling dedicated to amateurs, as well as for providing fast drafting tools for artists.