Forensic Automatic Speaker Recognition Based on Likelihood Ratio Using Acoustic-phonetic Features Measured Automatically

• Autori principali: Huapeng Wang (Autore), Cuiling Zhang (Autore)
• Natura: Libro
• Pubblicazione: Wolters Kluwer Medknow Publications, 2015-01-01T00:00:00Z.
• Soggetti: article
• Accesso online: Connect to this object online.

Forensic speaker recognition is experiencing a remarkable paradigm shift in terms of the evaluation framework and presentation of voice evidence. This paper proposes a new method of forensic automatic speaker recognition using the likelihood ratio framework to quantify the strength of voice evidence...