Motion migration method driven by 3D human body model

Abstract

The invention discloses a motion migration method driven by a 3D human body model. The motion migration method comprises the following steps: converting training data into a UV space, and constructing and optimizing the 3D human body model according to complementary information between adjacent video frames; then projecting the optimized 3D human body model to a 2D plane to retain 3D information of original motion, and driving the optimized 3D human body model with a target attitude; taking the attitude of the 2D projection and the training data as the input of a pre-training model, and storing the trained model; normalizing the posture of the target person; and finally, taking the 2D projection of the optimized 3D human body model driven by the posture of the target person and the normalized posture of the target person as the input of a trained motion image generation model to carry out final motion migration, so that the problems of blurring, shape distortion and the like existing in the generation of a 2D plane image are solved; and the generated moving image is ensured to have reliable depth information, accurate shape and clear face.

Description

technical field [0001] The invention belongs to the technical field of motion migration, and in particular relates to a motion migration method driven by a 3D human body model. Background technique [0002] Human motion transfer aims to synthesize human motion images with human textures and target poses from training images. It has been used in filmmaking, game design, and medical rehabilitation. Based on human motion transfer technology, the characters of the training images can be freely animated to perform user-defined imitation actions. Traditional computer graphics-based motion transfer methods require complex rendering operations to generate appearance textures, which are time-consuming and computationally complex, but ordinary users or small institutions cannot afford the extremely high computational and time costs. [0003] Human motion is a complex natural phenomenon, all real motion occurs in 3D space, the reason why real motion images look natural is that they a...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the deficiencies in the prior art, the present invention provides a motion migration method driven by a 3D human body model, which not only overcomes the problems of blurring and shape distortion in the generation of 2D plane images, but also ensures that the generated motion images have reliable depth information , accurate shape and clear human face

Images