Speaker recognition using temporal contours in linguistic units: the case of formant and formant-bandwidth trajectories | Publicación