Skeletal animation vertex correction method and model learning method, device and apparatus

Abstract

The invention provides a skeleton animation vertex correction method and a model learning method, device and apparatus. The method comprises the steps of collecting the whole body images and posture data of various motions including T-shaped postures of different people, and obtaining the model training data, wherein the model training data comprises the posture data P and Pm of a standard 3D mesh model, the coordinates Gr of N vertexes, N vertex coordinates T and n joint point coordinates J of a T-shaped posture 3D mesh model; constructing and training a joint position prediction model, constructing a postures regression network model and a skeleton animation vertex correction model comprising a standard skin function, the trained joint position prediction model and the postures regression network model, and training. The skeleton animation vertex correction model provided by the invention can compensate and correct the unreasonable deformation generated on the original skin surface in the movement process of a character role, so that the character role looks vivid and natural, the animation quality is improved, the workload of an animator is reduced, and the time spent on making animations with the same quality is shorter.

Description

technical field [0001] The present application relates to the technical field of character skin production, and in particular to a learning method for a vertex correction model of skeletal animation, a learning device for a vertex correction model of skeletal animation, a method for vertex correction of skeletal animation, and computer equipment. Background technique [0002] At present and AR smart technology is developing rapidly, Internet users have more and more content intake requirements, and the demand for animation characters, AR and other content is becoming stronger and stronger. The brush weight of the mapping of the vertices of the mesh makes the surface of the animated character form a reasonable deformation during the movement process, which is the character skinning. The production of traditional skeletal animation skins is completely dependent on manual work. Different parts of different characters need to be manually assigned different weights. This process ...

AI Technical Summary

Problems solved by technology

[0003] In the process of character skinning, for different characters and different parts, the animator needs to give different weights by repeatedly observing the deformation of the character's surface during the movement process. This workload is not only very complicated, but also It is difficult to ensure that the final output animation is consistent with the expected performance, such as collapse in the linear hybrid skin model, dry skin surface at joint joints, etc.

This caused the animation engineer to repeatedly reset the weight of the character's skin, which greatly reduced the quality and output efficiency of the animation, making the cost of the animation high

At present, there are existing methods to solve the problem of skin collapse through IK constraints, and to solve the weight through human tissue dynamics, but these methods have problems such as heavy workload, long calculation time, inaccurate results, and the final effect depends on the experience of animators and cannot be obtained. wide application

Images