Multi-Characters Interactive Animation Synthesis Based On Honeycomb

Scenes where multiple characters densely interact with each other frequently appear in movies and computer games. For examples, high quality interactions between multiple characters in the movie "300" are very attractive. Then, in the "NBA" series and "The Lord of the Rings", players need to control avatars with computer-controlled avatar to attack or defend together and so on, which provides a better gameplay. Based on this strong demand, interactive character animation synthesis has become a very popular research direction. In paper, we complete three contents for current states and problems.Firstly, we propose a honeycomb structure and four principles of the Wing Chun based method to synthesize uncertain number of opponents and uncertain attack type action interactive animation scene. Based this, we propose interactive locations expression and design finite-state machine based on honeycomb to synthesize the role reactive motion. This method reaches our goal.Then, We propose a method of combining complex networks and crowd animation. Using a random graph based on the honeycomb and the uniform network to synthesize crowd animation. From the experimental results, it is feasible to combine complex networks and crowd animation.At last, the paper let users control avatars to interact. Low degrees of freedom based on Kinect to achieve real-time interactive animation more action details. And we employee action motion reconstruction and action recognition phase separation method, and based on honeycomb simulated animation.