Computer Science MSc student Xinyi Zhang and Professor Michiel van de Panne won the best paper award at the 11th annual conference on Motion, Interaction and Games (MIG 2018) held in Limassol, Cyprus.
Their paper, titled "Data-driven Autocompletion for Keyframe Animation", explores the potential of using machine learning to automatically produce motions that match a user-specified set of keyframes. The learned model is first trained on the motion characteristics of example motions and sampled keyframes from those motions. The method is based on an autoregressive recurrent neural network (ARNN) that is conditioned on the target keyframes, and a loss function that seeks to match motion predictions from the training data, as well as matching the specified keyframes.
Congratulations, Xinyi and Michiel!