Skip to main content

Voice biometric systems can be defeated by "wolves"

posted onMarch 15, 2012
by l33tdawg

 The founder of Australian voice biometrics outfit Armorvox, Dr. Clive Summerfield, claims that between 10% and 20% of utterances collected by voice biometric systems are not strong enough identifiers on the individual that spoke them.

Most voice biometric implementations require users to utter a phrase as part of the authentication process. However, a small fraction of the population Dr Summerfield labels as “wolves”, have voice prints that can match many others, while "goats" are those with generic voice prints. 

A "less effective" voice print can be caused by ambient noise or signal clipping when the original signal was collected. Dr. Summerfiled also believes that certain voice biometric systems build in tolerances for these less distinct voice prints, therefore applying a lower authentication standard when processing them. 

Source

Tags

Security Technology

You May Also Like

Recent News

Friday, November 8th

Friday, November 1st

Tuesday, July 9th

Wednesday, July 3rd

Friday, June 28th

Thursday, June 27th

Thursday, June 13th

Wednesday, June 12th

Tuesday, June 11th

Friday, June 7th