US Patent 7388586 Method and apparatus for animation of a human speaker

Methods and apparatus for representing speech in an animated image. In one embodiment, key point on the object to be animated are defined, and a table of trajectories is generated to map positions of the key points over time as the object performs defined actions accompanied by corresponding sounds. In another embodiment, the table of trajectories and a sound rate of the video are used to generate a frame list that includes information to render an animated image of the object in real time at a rate determined by the sound rate. In still another embodiment, a 2D animation of a human speaker is produced. Key points are selected from the Motion Picture Expert Group 4 (MPEG4) defined points for human lips and teeth.