Collaborative Aculab and University of York voice biometrics research project receives funding

Aculab has announced the full WRoCAH PhD funding has been awarded for a collaborative voice biometric project with the University of York’s Department of Language and Linguistic Science to examine how information used by automatic speaker recognition (ASR) systems for characterizing speakers can be compared to the linguistic properties of the voice.

The project is titled “Towards linguistically-informed automatic speaker recognition,” and the Collaborative Doctoral Award (CDA) studentship has been awarded to Elliot Holmes, who will begin working on the project in October. The collaborative project was originally announced in late-2018.

Dr. Vincent Hughes of the Department of Language and Linguistic Science at the University of York says the institution is home to a world-leading research center for forensic speech science, and that it is “pleased to work with a leading commercial provider of voice biometric systems.”

ASR is increasingly used for both commercial and forensic purposes, according to Aculab, but although experiments show very low error rates, little is known about what data advanced ASR systems actually capture. The project seeks to ascertain more about what data is used, to aid in the development and improvement of ASR technology, including voice analysis accuracy in forensic cases, and enhance the public, legal, and judicial understanding of speech recognition.

“(T)he project represents a large-scale investigation into how tangible linguistic information can be used to improve the understanding and performance of speaker recognition systems,” explains Hughes.

“The time he spends at Aculab will give Elliot a unique opportunity to familiarise himself with advanced speech technology, while broadening his skills in a commercial R&D environment,” explains Aculab Speech Technology Director Dr. Ladan Baghai-Ravary. “We are looking forward to embarking on this project, which has been designed to explain and enhance forensic applications of new and exciting technologies. We hope it will ultimately lead to increased uptake of automatic speech and voice analysis within the forensics industry, by providing a legally acceptable interpretation and justification of the evidence that such systems can give.”

Aculab’s VoiSentry system was recently confirmed compatible with Avaya IX Contact Center solutions in compliance testing.

Article Topics

Aculab  |  biometrics  |  forensics  |  speaker recognition  |  voice