Facial expression recognition approach for performance animation

We attempt a performance animation approach that makes use of a facial expression recognition technique. Our approach consists of two steps. First, we identify the performer´s facial expressions from the input image sequence, using a method that is modified from the well-known eigenface method but here not for recognizing a face but for recognizing expression. Second, we use the recognition result to initiate the change of facial expressions in the animated character. Since the eigen method-based system requires only the intensity maps not distinct features on the images to work, the intrusive makeup of the human performer and the feature extraction process needed in the current performance animation approaches are no longer needed. The simplicity in matching the input images with the database images for various facial expressions allows the system to be suitable for near-real-time application. In addition, in contrast with the current approaches that map facial dynamics directly from the human performer to the animated character, our approach could be used to drive any animated character, human-like or not