Method and system for redirecting movement styles of virtual human animation

Abstract

The invention discloses a method and a system for redirecting movement styles of virtual human animation. The method includes representing a plurality of movement data with different individual features, different pieces of stylistic information and identical movement contents as a plurality of low-dimension coordinates which are irrelevant to the individual features and only have different pieces of stylistic information and identical movement contents; building relative time-space transformation models among the different pieces of stylistic information in the multiple low-dimension coordinates; redirecting the movement styles of movement data with other movement contents by the aid of the built relative time-space transformation models. The method and the system have the advantages that the method for synthesizing the virtual human animation is visual and is high in efficiency, and the stylistic information is reused among the different movement data, so that stylized virtual human animation movement data which meet requirements of users can be synthesized.

Description

technical field [0001] The invention relates to the field of computer graphics, in particular to a motion style redirection method and system for virtual human animation synthesis. Background technique [0002] Virtual human animation technology has always been a research hotspot in the international graphics community, and has been widely used in many fields, such as film and television production, game entertainment, training simulation, safety rehearsal, etc. However, with the continuous improvement of the quality requirements for virtual human animation in various fields, people hope that on the basis of "displaying" the virtual human, they can endow the virtual human with a "soul". However, due to the limitation of production costs and other conditions, the previous avatar animations often used a set of fixed action templates repeatedly, which led to the behavior of different characters being exactly the same except for their different images. This kind of "cookie-cutt...

AI Technical Summary

Problems solved by technology

Spatio-temporal feature transformation methods can generate stylized human motion extremely fast, but these transformations are usually related to motion content and cannot reuse style information among heterogeneous motions

Implicit mapping methods can reuse style information between some motion content with similar behaviors, but this type of method is not suitable for more complex motion behaviors

The multi-linear analysis method is also not suitable for more complex human motion, because it requires the motion capture database to be composed of a large number of motion data with similar structures but semantically distinguishable from each other.

Images