Netmarble, a major Korean mobile game developer, said on Aug. 5 that it has disclosed a technology to automatically generate facial animations based on voice commands in a global computer graphics event held in Los Angeles, the United States.
Netmarble has published a paper on multi-task audio-driven facial animation (MTADFA) technology at SIGGRAPH 2019.
"Netmarble has developed MTADFA technology for the first time in the industry,” said Ahn Su-nam, head of the Netmarble AI Center who published the paper.
MTADFA is a technology that can simultaneously learn the voices and facial animations of 14 different characters and create more facial animations of the characters. In this paper, Netmarble focused on solving problems of the existing technology that requires learning of the voices and facial animations of individual characters separately. The company will apply the new technology to future games.