Stylized body movement generating and editing method based on sub-space technology

Abstract

This invention discloses a generation and editing method, which bases on space technology style of human motion. Its Principal component analysis is used to establish a contact high-dimensional data space and movement style annihilator low-dimensional space mapping. Low-dimensional subspace has maintained a campaign by the inherent style, as well as because of significantly reducing dimension, lower cost, easy to quantify the degree of style. In low-dimensional subspace can conduct real-time calculation, and the results generated or edit reconstruction through mapping data movement in the high-dimensional space. By extrapolation, can be generated or editing samples than the style of training more exaggerated movement. Because of the diversity of human motion style, this method also proposed a new style of dealing with multiple sports programs, the program was effective in resolving the multiple styles between the weight of the problems brought about by orthogonal.

Description

technical field [0001] The invention relates to the field of multimedia human three-dimensional animation, in particular to a method for generating and editing stylized human motion based on subspace technology. Background technique [0002] Stylized human motion generation and editing is an important topic in the field of computer animation. In a large virtual reality scene, a large number of human motions with different styles need to be rendered. Now the application of Optical Motion Capture (optical human body motion capture equipment) enables us to capture motions with certain typical styles in a targeted manner. How to start from these captured motions and generate or edit a large number of motions with different styles according to needs , is an important question. In recent years, there are also some methods to try to solve this kind of problem. For example, "MotionGraphs" published on Siggraph in 2002 proposed a method that can be used to generate real and control...

AI Technical Summary

Problems solved by technology

In recent years, there are also some methods to try to solve this kind of problem. For example, "MotionGraphs" published on Siggraph in 2002 proposed a method that can be used to generate real and controllable motion. The article proposed the data structure of motion graph; 00 "Style Machines" published on Siggraph in 2011 learns motion patterns through some motion capture data samples, and then interpolates and extrapolates the learned patterns to generate different styles of motion, but in that method, users cannot Quantify the motion style, and at the same time, this method can only complete motion generation, and cannot edit an existing piece of motion data according to the user's style requirements; "Automated Extraction and Parameterization of Motions in Large Data Sets" published on Siggraph in 2004 proposed An automated method is proposed, which can search for logically similar motions in a large motion database, and then use these logically similar motions to build a continuous stylized parameter space, and then the user can control the parameters by The parameter values ​​in the optimized space are used to control the style of human motion; "PCA-based WalkingEngine using Motion Capture Data" published on Computer Graphics International in 2004 uses principal component analysis to construct a walking motion generation engine, but in The style at a higher level of semantics in that method is not taken into account

[0003] In short, the current methods are often difficult to solve some of the following problems:

[0004] 1. Motion data are very high-dimensional data. Since we consider continuous motion, frames cannot be viewed separately. Usually, the human skeleton model has more than a dozen joint points and dozens of degrees of freedom. In this way, a Just 30 frames of motion is a thousand-dimensional vector

[0005] 2. The motion style is on the high-level semantics. The connection between the low-level motion vector and the high-level style semantics is very complicated. It is difficult to directly obtain the style directly from the data, or generate data from the style.

[0007] 4. In the field involving multiple sports styles, since the effects of different sports styles are not orthogonal, the independent adjustment of each style cannot achieve the required effect

Images