人体运动的非线性降维及新运动生成

Nonlinear Dimensional Reduction and Generation of Human Motion

为了提高现有运动数据的可重用性,生成更为丰富的新运动,提出快速自适应比例高斯过程隐变量模型,以及基于该模型的人体运动数据降维及运动生成方法.通过对运动数据进行统计学习,获得运动数据在隐空间的一个低维映射来实现非线性降维,同时获得了该运动的姿态空间的概率分布,其大小反映了该姿态的自然逼真程度;在给定末端约束条件下求取满足约束的、同时概率最大的姿态,并将其作为逆向运动学的解,克服了传统逆向运动学算法计算烦琐、效果不逼真的缺点.实验结果表明,该模型具有更快的收敛速度和更高的收敛精度,同时能够自适应运动编辑的方向,有效地扩大运动的可编辑幅度.

To reuse existing motion data to generate new motions,a nonlinear dimensional reduction and generation of human motion is presented,which is based on fast adaptive scaled Gaussian process latent variable models.Through statistical learning,the motion data are mapped from high-dimensional observation space to low-dimensional latent space to implement nonlinear dimensional reduction,and probability distributing of posture space which measures the nature of posture is obtained.The posture which meets constraints and has maximal probability can be computed as the solution of inverse kinematics.This method can avoid cockamamie computation and posture distortion existing in traditional inverse kinematics.The experiments show that our method has higher convergence speed and precision and extends editing range of motion by adapting motion editing direction.