Generation of Arm-gesture and Facial Expression for Intelligent Avatar Communications on the Internet

Recentlythesign-languagecommunicationsystemsbetween avatars of different languageshavebeeninvestigatedas a meansof overcomingthe linguistic barrier. In thesystems, an intelligent communicationmethodhas beenemployed, wheresetsof theanimationparameter ssuch asthejoint anglesof the gesture were transmittedinsteadof sendingthe entire-real motion pictures. However, the communication hasbeendonebasedonthegestureonlywithoutconsidering thefacial expression.In thispaperweproposean approach to thecommunicationbasedonthefacial expressionaswell as thearm-gesture, andgenerating themon variousavatar models.To extract the parameter s to be transmitted,three kinds of key-frameeditors are employedusing techniques of inversekinematicsand partial differential equations.In generating facial expressionespecially, the movementsof thecheeksandthe jawsaswell asother facial components are alsoutilized. Thepreliminaryresultsshowa possibility that themethodcouldbeusedasa usefulmeansfor avatar communicationsbetweendifferentlanguagesontheInternet