In the
framework of the BioSecure Network of Excellence (http://www.biosecure.info), a
talking-face identity verification reference system was developed: it is
open-source and made of replaceable modules. This is an extension of the BECARS
speaker verification toolkit, implementing theGMM approach. In this paper, the
audio and visual features extraction front-ends are presented. The performance
of the system on the Pooled protocol of the BANCA database are described.