A Preliminary Investigation into the Impact of Training for Example-Based Facial Blendshape Creation

Abstract

Our work is a perceptual study into the effects of training poses on the Example-Based Facial Rigging (EBFR) method. We analyse the output of EBFR given a set of training poses to see how well the results reproduced our ground truth actor scans compared to a Deformation Transfer approach. While EBFR produced better results overall, there were cases that did not see any improvement. While some of these results may be explained by lack of sufficient training poses for the area of the face in question, we found that certain lip poses were not improved by training, despite a large number of mouth training poses supplied.