Abstract: A computer-implemented method for generating a machine-learned inverse rig model. The machine-learned inverse rig model outputs rig parameters for a facial animation rig used to generate animations of facial expressions in video games. The method comprises receiving one or more training examples. Each training example comprises target mesh data for a face portraying a particular facial expression. For each training example, predicted rig parameters are generated using an inverse rig model, comprising processing the target mesh data of the training example. Predicted mesh data is generated, comprising processing the predicted rig parameters using a forward rig model. A loss is determined, comprising performing a comparison between the predicted mesh data and the target mesh data. The inverse rig model is trained, comprising updating parameters of the inverse rig model based on the losses of one or more training examples.

Abstract: A computer-implemented method for generating a machine-learned inverse rig model. The machine-learned inverse rig model outputs rig parameters for a facial animation rig used to generate animations of facial expressions in video games. The method comprises receiving one or more training examples. Each training example comprises target mesh data for a face portraying a particular facial expression. For each training example, predicted rig parameters are generated using an inverse rig model, comprising processing the target mesh data of the training example. Predicted mesh data is generated, comprising processing the predicted rig parameters using a forward rig model. A loss is determined, comprising performing a comparison between the predicted mesh data and the target mesh data. The inverse rig model is trained, comprising updating parameters of the inverse rig model based on the losses of one or more training examples.