Artificial intelligence-based generation of anthropomorphic signatures and use thereof
The technology disclosed relates to authenticating users using a plurality of non-deterministic registration biometric inputs. During registration, a plurality of non-deterministic biometric inputs are given as input to a trained machine learning model to generate sets of feature vectors. The non-deterministic biometric inputs can include a plurality of face images and a plurality of voice samples of a user. A characteristic identity vector for the user can be determined by averaging feature vectors. During authentication, a plurality of non-deterministic biometric inputs are given as input to a trained machine learning model to generate a set of authentication feature vectors. The sets of feature vectors are projected onto a surface of a hyper-sphere. The system can authenticate the user when a cosine distance between the authentication feature vector and a characteristic identity vector for the user is less than a pre-determined threshold.
1. A computer-implemented method of establishing authentication credentials using a plurality of non-deterministic registration biometric inputs and at least one deterministic biometric input, the method including:
feeding, during registration, the plurality of non-deterministic biometric inputs and the deterministic biometric input to a trained machine learning model and generating sets of feature vectors, wherein the non-deterministic biometric inputs include a plurality of face images and a plurality of voice samples of a user, and wherein the deterministic biometric input includes genomic data of the user;
projecting the sets of feature vectors onto a surface of a hyper-sphere;
computing a characteristic identity vector representing the user based on a user's set of the projected feature vectors; and
saving the characteristic identity vector for use during authentication of the user.
2. The method of claim 1 , further including registering a cryptographic signature of the characteristic identity vector with an identity server, accompanied by a photograph of the user selected from the plurality of face images.
3. The method of claim 1 , further including registering the characteristic identity vector with an identity server, accompanied by a photograph of the user that becomes one of the plurality of face images used to compute the characteristic identity vector.
4. The method of claim 1 , wherein the sets of feature vectors are projected onto a surface of a unit hyper-sphere.
5. The method of claim 1 , further including:
feeding the genomic data to a hash; and
using the hashed genomic data as at least one dimension of the characteristic identity vector.
6. The method of claim 1 , further including preprocessing the face image data to separate the face from background before generating the feature vectors.
7. The method of claim 1 , further including preprocessing the face image data by selecting individual frames from a video sequencing, wherein the individual frames are selected based on an image sharpness metric.
8. The method of claim 1 , further including preprocessing the voice data by selecting an audio segment corresponding to a phrase specified for the user to read.
9. A system including one or more processors coupled to memory, the memory loaded with computer instructions to establish authentication credentials using a plurality of non-deterministic registration biometric inputs and at least one deterministic biometric input, the instructions, when executed on the processors, implement actions comprising:
feeding, during registration, the plurality of non-deterministic biometric inputs and the deterministic biometric input to a trained machine learning model and generating sets of feature vectors, wherein the non-deterministic biometric inputs include a plurality of face images and a plurality of voice samples of a user, and wherein the deterministic biometric input includes genomic data of the user;
projecting the sets of feature vectors onto a surface of a hyper-sphere;
computing a characteristic identity vector representing the user based on a user's set of the projected feature vectors; and
saving the characteristic identity vector for use during authentication of the user.
10. The system of claim 9 , further implementing actions comprising:
registering a cryptographic signature of the characteristic identity vector with an identity server, accompanied by a photograph of the user selected from the plurality of face images.
11. The system of claim 9 , further implementing actions comprising:
registering the characteristic identity vector with an identity server, accompanied by a photograph of the user that becomes one of the plurality of face images used to compute the characteristic identity vector.
12. The system of claim 9 , wherein the sets of feature vectors are projected onto a surface of a unit hyper-sphere.
13. The system of claim 9 , further implementing actions comprising:
feeding the genomic data to a hash; and
using the hashed genomic data as at least one dimension of the characteristic identity vector.
14. The system of claim 9 , further implementing actions comprising:
preprocessing the face image data to separate the face from background before generating the feature vectors.
15. The system of claim 9 , further implementing actions comprising:
preprocessing the face image data by selecting individual frames from a video sequencing, wherein the individual frames are selected based on an image sharpness metric.
16. The system of claim 9 , further implementing actions comprising:
preprocessing the voice data by selecting an audio segment corresponding to a phrase specified for the user to read.
17. A non-transitory computer readable storage medium impressed with computer program instructions to establish authentication credentials using a plurality of non-deterministic registration biometric inputs, the instructions, when executed on a processor, implement a method comprising:
feeding, during registration, the plurality of non-deterministic biometric inputs and the deterministic biometric input to a trained machine learning model and generating sets of feature vectors, wherein the non-deterministic biometric inputs include a plurality of face images and a plurality of voice samples of a user, and wherein the deterministic biometric input includes genomic data of the user;
projecting the sets of feature vectors onto a surface of a hyper-sphere;
computing a characteristic identity vector representing the user based on a user's set of the projected feature vectors; and
saving the characteristic identity vector for use during authentication of the user.
18. The non-transitory computer readable storage medium of claim 17 , implementing the method further comprising:
registering a cryptographic signature of the characteristic identity vector with an identity server, accompanied by a photograph of the user selected from the plurality of face images.
19. The non-transitory computer readable storage medium of claim 17 , implementing the method further comprising:
registering the characteristic identity vector with an identity server, accompanied by a photograph of the user that becomes one of the plurality of face images used to compute the characteristic identity vector.
20. The non-transitory computer readable storage medium of claim 17 , wherein the sets of feature vectors are projected onto a surface of a unit hyper-sphere.