Soraia Raupp Musse,  Fernando Santos Osório,  Fabien Garat,  Michael Gomez,  Daniel Thalmann. 

Interaction between real and virtual humans spans several research topics from creation and animation of virtual actors to user-interfaces and virtual agents' control. In this paper we present a system to provide the user interaction with virtual human crowds using hands postures capture. An artificial neural network process recognizes the hand postures and a Client/Server system handles the communication between the several modules. In addition, ViCrowd, a system to manage crowds is responsible for dealing with crowd behaviors.

http://www.lbd.dcc.ufmg.br/colecoes/wrv/2000/008.pdf

Processando consulta.