摘要
为了合成应对不确定数目对手和不确定攻击类型动作的多角色交互性动画,提出基于蜂窝结构和咏春拳的四门原理方法.该方法将复杂网络和人群动画相结合,分别提出了基于蜂窝结构的随机图和均匀网络来合成多角色交互性动画;采用将运动重建和动作识别相分离的策略,并使用一个双级结构来重建用户运动,使得用户可以控制虚拟角色参与交互性动画合成.实验结果表明,文中方法可以实时地合成逼真的多角色交互性动画.
This paper introduces a method based on honeycomb structure and four gates principle to synthesize interactive animation scenes of multi-characters, where the number of opponents is unknown nor the types of actions. Our method uses theory from complex network area to synthesize crowd animation. Random graphs and a honeycomb structure-based uniform network are used to synthesize animations of multi-characters. In order to create interactive scenes, a three-dimension human motion reconstruction method with two-layer is implemented, which separates the motion reconstruction from the motion recognition. Experimental results show that our method is capable of synthesizing realistic interactive animations of multi-characters.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2013年第6期906-917,共12页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(61070043)
浙江省自然科学基金(Y1110882
Y1110688
R1110679)
教育部高等学校博士学科点专项科研基金(20113317110001)
关键词
蜂窝
多角色交互
四门原理
运动合成
复杂网络
honeycomb
multi-character interaction
four gates principle
motion synthesis
complexnetwork