Expression redirection training method and device, electronic equipment and readable storage medium

Abstract

The invention provides an expression redirection training method and device, electronic equipment and a readable storage medium, and the method comprises the steps: collecting a plurality of video frames of a training user; calculating the expression base coefficients of a plurality of preset expression base templates relative to the video frames for each video frame; adjusting a control parameter for controlling the expression of a virtual image, so that the expression of the virtual image is consistent with the expression of the training user in the video frame; and finally, training the constructed network model based on the expression basis coefficients and the control parameter corresponding to the video frames to obtain a redirection model. In the scheme, firstly, different expressions of the training user are characterized by the expression base template and the expression base coefficient, so that various expressions can be characterized by normative information, and a redirection model capable of realizing expression migration and keeping personalized characteristics of a virtual image can be obtained through training by adjusting the virtual image to obtain the control parameter, and good-applicability and natural expression redirection is realized.

Description

technical field [0001] The present application relates to the technical field of model control, in particular, to an expression redirection training method, device, electronic equipment and readable storage medium. Background technique [0002] Expression redirection refers to transferring the user's expression to another image. For example, in a live broadcast scene, the anchor's expression can be transferred to an avatar. When transferring the user's expression to the avatar, a commonly used processing method generally hopes to completely and accurately transfer the user's expression to the avatar. [0003] However, there are various styles of avatars, such as loli images, animal images, hero images, etc. The facial features of different types of avatars have personalized features, the proportion relationship between the various parts of the face, the The size of the features are all different. If the user's facial expression information is simply copied to the avatar, d...

AI Technical Summary

Problems solved by technology

If the user's facial expression information is simply copied to the avatar, due to the differences in facial features between the avatar and the user and the differences in facial features between different types of avatars, this complete copying method will The transferred expressions are not very suitable for the avatar, making the avatar's expression look unnatural

Images