Back to results
US 11,699,449 B2Granted
In-ear liveness detection for voice user interfaces
AssigneeJVCKENWOOD Corporation
InventorJan Jasper van den Berg
Life of this patent
Estimating the term…
Abstract
Introduced here are approaches to authenticating the identity of speakers based on the “liveness” of the input. To prevent spoofing, an authentication platform may establish the likelihood that a voice sample represents a recording of word(s) uttered by a speaker whose identity is to be authenticated and then, based on the likelihood, determine whether to authenticate the speaker.
Claims16
A non-transitory computer-readable medium with instructions stored thereon that, when executed by a processor, cause the processor to perform operations comprising:
- determining that a phrase has been uttered by a speaker over an interval of time as a biometric credential for authentication;
- acquiring data that is generated by an electronic device worn by an individual in an auricle of an ear,
- wherein the electronic device is configured to emit a series of probing signals into the ear canal, and
- wherein the data is representative of a series of return signals generated by reflection of the series of probing signals against a surface of the ear canal;
- establishing, based on the data, a shape of the ear canal over the interval of time;
- producing
- a first score indicative of a likelihood that the individual is the speaker based on the shape of the ear canal, and
- a second score indicative of correspondence between actual deformation of the ear canal over the interval of time as determined from the data and expected deformation of the ear canal over the interval of time given the phrase;
- and
- determining, based on the first and second scores, whether to authenticate the speaker as the individual.
7 more elements
A method comprising:
- causing, by an authentication platform, a notification to be generated that prompts an individual to utter a phrase while wearing an electronic device in an auricle of an ear;
- acquiring, by the authentication platform, data that is representative of a series of return signals generated by reflection of a series of probing signals emitted into an ear canal by the electronic device;
- producing, by the authentication platform based on the series of return signals, a model indicative of deformation of the ear canal as the phrase is uttered by the individual, wherein the model is representative of a series of discrete positions indicating how the shape of the ear canal varies over time as the individual utters the phrase;
- and
- storing, by the authentication platform, the model in a profile associated with the individual in a biometric database.
2 more elements
A method comprising:
- causing, by an authentication platform, a notification to be generated that prompts an individual to utter a phrase while wearing an electronic device in an auricle of an ear;
- acquiring, by the authentication platform, data that is representative of a series of return signals generated by reflection of a series of probing signals emitted into an ear canal by the electronic device;
- producing, by the authentication platform based on the series of return signals, a model indicative of deformation of the ear canal as the phrase is uttered by the individual;
- and
- storing, by the authentication platform, the model in a profile associated with the individual in a biometric database, wherein the model is one of multiple models stored in the profile associated with the individual, and wherein each model is associated with a different phrase.
2 more elements
Dependencies are inferred from the claim text.