| 摘要 |
Systems for generating a substantially life-like, individualized three- dimensional (3D) head model from at least one input image. In one example, a blend fit head model is created from a weighted combination of multiple basis head models after comparing a face image of the individual with the basis head models. The blend fit model can be deformed to approximate the individual's facial and/or head geometry. The method may include associating an animation rig with the 3D head model. In certain examples, head models are used to create media content having at least a first layer with at least one background character and a second layer with a head model, alignment information, and lighting information such that a head of the background character is replaced and/or overlaid by the head model. A head model can also be automatically selected for inclusion in an interactive advertisement based on the viewer's profile or preferences. |
