Abstract: In accordance with some embodiments, an apparatus for privacy protection is provided. The apparatus includes a housing arranged to hold a second device; one or more sensors, at least partially supported by the housing, operable to continuously collect biometric data of a user; an authentication neural network, operable to extract from the biometric data a plurality feature vectors associated with a plurality of identifiable scores; and a decision unit, coupled to the authentication neural network, operable to generate an authentication score as a function of the plurality of identifiability scores and the plurality of feature vectors, determine whether or not the authentication score satisfies an authentication threshold, and gate electronic access to the second device base on whether or not the authentication score satisfies the authentication threshold.

Abstract: In accordance with some embodiments, a training method for biometric identity and authentication is provided. The method includes obtaining biometric data from a plurality of sources. The method further includes establishing a candidate set of neural network parameters. The method additional includes extracting a plurality of feature vectors from the biometric data using the candidate set of neural network parameters. The method also includes determining whether or not the plurality of feature vectors match a training vector set within an error threshold. The method further includes updating the candidate set of neural network parameters in response to determining that the plurality of feature vectors do not match the training vector set within the error threshold.

Abstract: In accordance with some embodiments, a training method for biometric identity and authentication is provided. The method includes obtaining biometric data from a plurality of sources. The method further includes extracting a plurality of feature vectors from the biometric data. The method also includes determining a plurality of identifiability scores correspondingly associated with the plurality of feature vectors, where each of the plurality of identifiability scores provides a quantitative characterization of a relative uniqueness of a corresponding one of the plurality of feature vectors.

Abstract: In accordance with some embodiments, an apparatus for privacy protection is provided. The apparatus includes a housing arranged to hold a second device; one or more sensors, at least partially supported by the housing, operable to continuously collect biometric data of a user; an authentication neural network, operable to extract from the biometric data a plurality feature vectors associated with a plurality of identifiable scores; and a decision unit, coupled to the authentication neural network, operable to generate an authentication score as a function of the plurality of identifiability scores and the plurality of feature vectors, determine whether or not the authentication score satisfies an authentication threshold, and gate electronic access to the second device base on whether or not the authentication score satisfies the authentication threshold.

Abstract: In accordance with some embodiments, a training method for biometric identity and authentication is provided. The method includes obtaining biometric data from a plurality of sources. The method further includes extracting a plurality of feature vectors from the biometric data. The method also includes determining a plurality of identifiability scores correspondingly associated with the plurality of feature vectors, where each of the plurality of identifiability scores provides a quantitative characterization of a relative uniqueness of a corresponding one of the plurality of feature vectors.

Abstract: In accordance with some embodiments, a training method for biometric identity and authentication is provided. The method includes obtaining biometric data from a plurality of sources. The method further includes establishing a candidate set of neural network parameters. The method additional includes extracting a plurality of feature vectors from the biometric data using the candidate set of neural network parameters. The method also includes determining whether or not the plurality of feature vectors match a training vector set within an error threshold. The method further includes updating the candidate set of neural network parameters in response to determining that the plurality of feature vectors do not match the training vector set within the error threshold.