Control of movement of a synthesised mouth from a text stream

Moving towards broadband communications considerable amount of research is dedicated to finding ways to compress images. An object oriented codec will compress an image by analysing it into the objects of which it consists and send their shape colour and motion. The paper describes how the lips (from a face and shoulders image) were modelled and how the model was used to synthesise a mouth using sample data. As a means of further compression a text stream was used to drive the mouth and make it `talk´