Feature-based likelihood ratios for speaker recognition from linguistically-constrained formant-based i-vectors | Publicación