Output details

A Constraint-Based Approach to Visual Speech for a Mexican-Spanish Talking Head

<26> Synthetic talking heads are increasingly being employed in interfaces for the public, yet the movement of their mouths often appears unrealistic. Our constraint-based optimisation technique is a new approach to the problem of coarticulation, producing more controllable realistic visual speech than that produced using the leading paradigm. Our work has been extended at Ohio State University to produce emotional visual speech; it has also informed an ESRC/EPSRC research project led by Maddock, a New Scientist interview, and a data capture collaboration (MRI, electromagnetic articulography, video) with the GIPSA-Lab, Grenoble, France, supported by a Royal Academy of Engineering travel grant.