Content retargeting using facial layers

摘要

Techniques are disclosed for retargeting facial expressions. Input is received that represents a facial expression of a first character. Facial layers are generated based on the received input. The facial layers include one or more parameters extracted from the received input. A facial expression for a second character and corresponding to the facial expression of the first character is generated, based on the facial layers and without defining any spatial correspondence between the first character and the second character.

主权项

1. A computer-implemented method to retarget facial expressions by providing a parameter-parallel retargeting space between an input facial expression and an output facial expression and via a plurality of facial layers, the computer-implemented method comprising:
receiving input representing a facial expression of a first character; generating a plurality of facial layers for the first character, based on the received input and a composition function and by operation of one or more computer processors, wherein the plurality of facial layers includes a plurality of parameters extracted from the received input, the plurality of parameters including a simplex coefficient common to the plurality of facial layers and weighted by a respective measure of influence for each of the plurality of facial layers; wherein the simplex coefficient is extracted by an optimization operation that includes determining a set of inputs for which a given function attains a minimum value, wherein the optimization operation is subject to a set of constraints and is based on a sequential quadratic programming algorithm; wherein each facial layer encodes, in a simplicial basis, one or more semantically significant aspects of the facial expression of the first character; and wherein a simplex is formed from components of each simplicial basis; and generating, for a second character different from the first character in appearance, a facial expression corresponding to the facial expression of the first character, based on the plurality of facial layers.