Identifying features for multi-modality coding

Identifying feature areas in head and shoulder image is an important stage in solving the problems present in current telecommunication systems. For example, in the case of videophone systems, to gain lip synchronisation, it is believed that integrating the speech and image data would help reduce the loss of synchronisation between the voice and the lip movements. To do so, the mouth area should be known/detected and the corresponding speech information analysed. We discuss how the area of the mouth is detected and is used to help create a basic system for gaining lip synchronisation