基于粒子群优化的虚拟表情建模的应用研究

Applying PSO to virtual actor's facial expression modeling

探讨了如何训练虚拟人物表情这一新课题,提出了将TSK模糊神经网络应用于虚拟人物表情的建模研究,并用粒子群优化(PSO)算法训练TSK模糊神经网络。实验结果表明,当该算法应用于训练虚拟人物表情这一问题时,能在保证精度的前提下快速收敛,并能避免陷入局部最优,从而使得不同的动画导演利用这个系统,能够根据自己的知识产生出符合自己要求的虚拟表情输出。

Virtual actor＇s facial expression modeling is an important and practical topic.In this paper,TSK fuzzy neural network is used to accommodate director＇s expert knowledge such that appropriate virtual actor＇s facial expression can be produced.In order to speed up the convergence of the learning algorithm,PSO algorithm is incorporated into the TSK fuzzy neural network.The experimental results indicate the success of the approach here.