Automatic behavior generation techniques for conversational characters in computer games

Kurzinhalt:

"Character animation is widely used in 3D games extensively e.g. in crowd simulations or dialog situations. When it comes to social dialog such characters are envisioned to have the same properties as humans in face-to-face conversation, including the ability to generate simultaneous verbal and nonverbal behaviors. The challenge in providing conversational characters is to equip them with believable behaviors in dialogue situations. Most character engines use manual tagging processes to align specific animations to spoken sentences. To be able to handle large dialogue content, automatic behavior generation techniques could significantly enhance the character animation workflow. 

In particular the talk focuses on the factors that promote and maintain believable behaviors in virtual characters during social conversations. Its major contribution is the transformation of important aspects in face-to-face conversation competence e.g. from cognitive psychology into mathematical models using automatic animation generation techniques. This includes the automatic generation of important nonverbal behavior which may remarkably improve the experience of a dialogue."