Forensic verification of speakers: We invite you to a lecture by a leading European expert

Forensic speaker verification is a key procedure for investigators, courts, lawyers – in short, anyone who has ever handled recordings of suspects (police wiretaps, extremist videos, etc.). At our faculty, the Speech@FIT group is intensively engaged in automatic speaker verification. Just a few streets away is the spin-off company Phonexia, whose portfolio also includes this area of research. Brno is simply a recognised centre of excellence in this field of research. This is also why we are able to welcome Daniel Ramos, associate professor at the Universidad Autónoma de Madrid (UAM), to the VGS-IT seminar on Monday 14 September. Ramos is a leading European forensic expert, so this is an exceptional opportunity to hear a key expert speak on Brno soil.

The lecture, Rigorous Forensic Automatic Speaker Recognition: Bayesian Decision Theory, Probabilistic Calibration and Case-Specific Validation, will take place from 13:00 in lecture theatre G202. All interested parties are warmly invited. However, rather than pretty, colourful spectrograms, be prepared for Bayesian decision theory, probabilistic calibration and other mathematical specialities. The topic covers three key elements that are essential for the correct and standardised use of automatic speaker recognition systems in the forensic sciences. Details of the lecture are available HERE.

We should add that the event is part of the VGS-IT (Invited Talks on Vision, Graphics, and Speech) series. This series is dedicated to machine learning, speech processing, computer vision and large language models.