Output details
13 - Electrical and Electronic Engineering, Metallurgy and Materials
University of Surrey
Model-Based Synthesis of Visual Speech Movements from 3D Video
This paper explains how to synthesise visual-speech movements using a lip-shape-and-velocity parameterisation of 3D face dynamics. Although faces are vital in communication, human sensitivity to facial perception makes such synthesis challenging. In this paper, visual speech animations are constructed from audio input by mapping onto lip movements and selecting stored phonetic units to the target utterance. This combination of representation with unit selection shows visible improvement, facilitating facial animation solutions for film products of actors' digital doubles (Framestore - Nico Scapel, Head of Animation <n.scapel@framestore.com>) and for visual communication.