Method and apparatus for biometric authentication using face radar signal

1. An electronic device, comprising:

a memory configured to store preregistered user data;

a radar transceiver; and

a processor operably connected to the memory and the radar transceiver, the processor configured to:

transmit, via the radar transceiver, radar signals,

identify signals of interest that represent biometric information of a user based on reflections of the radar signals received by the radar transceiver,

generate an input based on the signals of interest that include the biometric information,

determine, from the input, whether the signals of interest represent a living object,

after a determination that the signals of interest represent the living object, extract, using a neural network, a feature vector from the input, and

authenticate the user based on comparison of the feature vector to a threshold of similarity with the preregistered user data.

2. The electronic device of claim 1 , wherein:

the electronic device further comprises a camera configured to capture an image of the user; and

the processor is further configured to extract the feature vector based on both the image and the input, wherein the input is based on the radar signals.

3. The electronic device of claim 1 , wherein the neural network is configured to generate the feature vector from the input and a photographic image of the user.

4. The electronic device of claim 1 , wherein:

the feature vector is one of a plurality of feature vectors;

the preregistered user data includes a set of pre-generated feature vectors; and

to authenticate the user, the processor is configured to:

identify feature vector pairs that include one of the plurality of feature vectors and one of the set of pre-generated feature vectors, respectively,

generate similarity scores for the feature vector pairs, respectively,

determine whether each of the similarity scores are above the threshold of similarity,

assign first decisions accepting one or more of the feature vector pairs having similarity scores above the threshold of similarity, respectively,

assign second decisions rejecting one or more of the feature vector pairs having similarity scores below the threshold of similarity, respectively, and

determine to authenticate the user when a quantity of the first decisions is larger than a quantity of the second decisions.

5. The electronic device of claim 1 , wherein the processor is configured to:

receive, via the radar transceiver, raw radar signals associated with different antenna configurations from the reflections of the radar signals;

combine the raw radar signals; and

select the signals of interest from the combined raw radar signals.

6. The electronic device of claim 1 , wherein:

the preregistered user data includes preregistered radar signals that are reflected off of the user;

the processor is further configured to:

receive a set of received radar signals that are reflected off of an object;

identify whether a material that the set of received radar signals and the preregistered radar signals are reflected off of is different, and

determine whether to classify the object as alive or fake based on the material that the set of received radar signals and the preregistered radar signals are reflected off of is different; and

to authenticate the user, the processor further configured to authenticate the user when a score based on the feature vector and the preregistered user data is above the threshold of similarity and the object is classified as alive.

7. The electronic device of claim 1 , wherein the processor is further configured to:

identify a distance between the electronic device and the user, and

select a model based on the distance between the electronic device and the user, wherein the feature vector is extracted based on the model.

8. A method comprising:

transmitting, via a radar transceiver, radar signals;

identifying signals of interest that represent biometric information of a user based on reflections of the radar signals received by the radar transceiver;

generating an input based on the signals of interest that include the biometric information;

determining, from the input, whether the signals of interest represent a living object;

after a determination that the signals of interest represent the living object, extracting, using a neural network, a feature vector from the input; and

authenticating the user based on comparison of the feature vector to a threshold of similarity with preregistered user data.

9. The method of claim 8 , further comprising

capturing, via a camera, an image of the user; and

extracting the feature vector based on both the image and the input, wherein the input is based on the radar signals.

10. The method of claim 8 wherein the neural network generates the feature vector from the input and a photographic image of the user.

11. The method of claim 8 , wherein:

the feature vector is one of a plurality of feature vectors;

the preregistered user data includes a set of pre-generated feature vectors; and

authenticating the user comprises:

identifying feature vector pairs that include one of the plurality of feature vectors and one of the set of pre-generated feature vectors, respectively,

generating similarity scores for the feature vector pairs, respectively,

determining whether each of the similarity scores are above the threshold of similarity,

assigning first decisions accepting one or more of the feature vector pairs having similarity scores are above the threshold of similarity, respectively,

assigning second decisions rejecting one or more of the feature vector pairs having similarity scores are below the threshold of similarity, respectively, and

determining to authenticate the user when a quantity of the first decisions is larger than a quantity of the second decisions.

12. The method of claim 8 , further comprising:

receiving, via the radar transceiver, raw radar signals associated with different antenna configurations from the reflections of the radar signals;

combining the raw radar signals; and

selecting the signals of interest from the combined raw radar signals.

13. The method of claim 8 , wherein:

the preregistered user data includes preregistered radar signals that are reflected off of the user;

the method further comprising:

receiving a set of received radar signals that are reflected off of an object;

identifying whether a material that the set of received radar signals and the preregistered radar signals are reflected off of is different, and

determining whether to classify the object as alive or fake based on the material that the set of received radar signals and the preregistered radar signals are reflected off of is different; and

authenticating the user, comprises authenticating the user when a score based on the feature vector and the preregistered user data is above the threshold of similarity and the object is classified as alive.

14. The method of claim 8 , further comprising:

identifying a distance between the radar transceiver and the user; and

selecting a model based on the distance between the radar transceiver and the user, wherein the feature vector is extracted based on the model.

15. A non-transitory computer readable medium embodying a computer program, the computer program comprising computer readable program code that, when executed by a processor of an electronic device, causes the processor to:

transmit, via a radar transceiver, radar signals;

identify signals of interest that represent biometric information of a user based on reflections of the radar signals received by the radar transceiver;

generate an input based on the signals of interest that include the biometric information;

determine, from the input, whether the signals of interest represent a living object;

after a determination that the signals of interest represent the living object, extract, using a neural network, a feature vector from the input; and

authenticate the user based on comparison of the feature vector to a threshold of similarity with preregistered user data.

16. The non-transitory computer readable medium of claim 15 , wherein the computer readable program code, when executed by the processor, further causes the processor to:

capture an image of the user; and

extract the feature vector based on both the image and the input, wherein the input is based on the radar signals.

17. The non-transitory computer readable medium of claim 15 , wherein the computer readable program code, when executed by the processor, further causes the neural network to generate the feature vector from the input and a photographic image of the user.

18. The non-transitory computer readable medium of claim 15 , wherein:

the feature vector is one of a plurality of feature vectors;

the preregistered user data includes a set of pre-generated feature vectors; and

the computer readable program code, when executed by the processor, further causes the processor to:

identify feature vector pairs that include one of the plurality of feature vectors and one of the set of pre-generated feature vectors, respectively,

generate similarity scores for the feature vector pairs, respectively,

determine whether each of the similarity scores are above the threshold of similarity,

assign first decisions accepting one or more of the feature vector pairs having similarity scores are above the threshold of similarity, respectively,

assign second decisions rejecting one or more of the feature vector pairs having similarity scores are below the threshold of similarity, respectively, and

determine to authenticate the user when a quantity of the first decisions is larger than a quantity of the second decisions.

19. The non-transitory computer readable medium of claim 15 , wherein the computer readable program code, when executed by the processor, further causes the processor to:

receive, via the radar transceiver, raw radar signals associated with different antenna configurations from the reflections of the radar signals;

combine the raw radar signals; and

select the signals of interest from the combined raw radar signals.

20. The non-transitory computer readable medium of claim 15 , wherein:

the preregistered user data includes preregistered radar signals that are reflected off of the user; and

the computer readable program code, when executed by the processor, further causes the processor to:

receive a set of received radar signals that are reflected off of an object;

identify whether a material that the set of received radar signals and the preregistered radar signals are reflected off of is different, and

determine whether to classify the object as alive or fake based on the material that the set of received radar signals and the preregistered radar signals are reflected off of is different; and

to authenticate the user, the computer readable program code, when executed by the processor, further causes the processor to authenticate the user when a score based on the feature vector and the preregistered user data is above the threshold of similarity and the object is classified as alive.