Multispace behavioral model for face-based affective social agents
Journal on Image and Video Processing
Motion in Games
Hi-index | 0.00 |
This paper describes an animated conversationalagent called Kare1 which integrates a talkinghead interface with a linguistically motivatedhuman-machine dialogue system. The agent hasa range of nonverbal behaviors, which involvea mixture of machine vision, computer animationand natural language processing techniques.The system's architecture couples the agent's nonverbalcommunicative processes very tightly toits model of verbal interaction. We discuss severalconsequences of this architecture, in particularthe ability to use different non-verbal dialoguemanagement signals when speaking different languages.