Gesture Synthesis for Human-Computer Interaction

Gesture recognizers require a large pool of training data to achieve good accuracy. However, recruiting participants, data collection and labeling, etc. necessary for achieving this goal are usually time-consuming and expensive. Fortunately, the Kinematic Theory allows to easily bootstrap gesture generation. In this paper, we show that the synthesized gestures not only perform equally similar to gestures generated by human users but also they look and feel the same. Ultimately, this work benefits researchers and designers who wish to prototype gesture-driven applications.