U.S. Patents 6,307,576 (“the ’576 patent”) and 6,611,278 (“the ’278 patent”) related to automating part of a preexisting 3-D animation method. As explained in the background of the patents, the admitted prior art method used multiple 3-D models of a character’s face to depict various facial expressions made during speech. In the prior art system, the animator used a keyframe approach, where the artist set the appropriate weights for the facial expression at important times in the sequence, i.e., keyframes, instead of at every frame. An animator would have to subjectively identify the problematic sequence and manually fix it by adding an appropriate keyframe.
Read the article here.