Voice input authentication device and method

1. A device for authenticating a voice input provided from a user, the device comprising:

a microphone configured to receive the voice input;

a memory configured to store one or more instructions; and

a processor configured to execute the one or more instructions,

wherein the processor is further configured to execute the one or more instructions to:

obtain, from the voice input, signal characteristic data representing signal characteristics of the voice input;

apply the obtained signal characteristic data to a first learning model configured to determine whether an attribute of the voice input corresponds to a voice uttered by a person and a voice output by an apparatus;

obtain context information including at least one of device state information, user's device usage history information, and user schedule information; and

based on a reliability score of determining that the attribute of the voice input corresponds to the voice uttered by the person, determine whether a user authentication is required;

when the user authentication is required, apply the voice input and the context information to a second learning model to authenticate the user; and

based on a result of the user authentication, authenticate the voice input.

2. The device of claim 1 , wherein the signal characteristic data comprises information about a per-frequency cumulative power of the voice input.

3. The device of claim 2 , wherein the first learning model is trained to determine the attribute of the voice input differently according to per-frequency cumulative powers of the voice uttered by the person and the voice output by the apparatus.

4. The device of claim 1 , wherein the processor is further configured to execute the one or more instructions to:

obtain a voice input pattern of the user;

wherein the voice input pattern is determined based on the voice input which is input by the user or a situation in which the voice input is input; and

apply the voice input pattern to the second learning model configured to authenticate the voice input.

5. The device of claim 4 , wherein the voice input pattern comprises a user's usage behavior of inputting voice.

6. The device of claim 1 , wherein the context information comprises the device state information, and the device state information comprises information of at least one of an operation mode of the device, a location of the device, a communication module activation state of the device, and a network connection state of the device.

7. The device of claim 1 , wherein the context information comprises the user's device usage history information, and the user's device usage history information comprises at least one of an application usage history, a user's call history, a user's text history, and a usage frequency of a voice recognition function.

8. The device of claim 1 , wherein the processor is further configured to execute the one or more instructions to:

output a plurality of preset words that are randomly selected by the device;

receive a cipher text including the plurality of preset words from the user; and

additionally authenticate the voice input based on comparison between the received cipher text and a preset cipher text.

9. The device of claim 1 , wherein the context information comprises an operation mode of the device that corresponds to a sound mode, a vibration mode, and a silent mode.

10. The device of claim 1 , wherein the context information comprises a communication module activation state of the device that indicates whether Wi-Fi is turned on or not, and whether Bluetooth is turned on or not.

11. A method of authenticating a voice input provided from a user, the method comprising:

receiving the voice input;

obtaining, from the voice input, signal characteristic data representing signal characteristics of the voice input; and

authenticating the voice input by applying the obtained signal characteristic data to a first learning model configured to determine whether an attribute of the voice input corresponds to a voice uttered by a person or a voice output by an apparatus,

obtaining context information including at least one of device state information, user's device usage history information, and user schedule information,

based on a reliability of determining that the attribute of the voice input corresponds to the voice uttered by the person, determining whether a user authentication is required;

when the user authentication is required, applying the voice input and the context information to a second learning model to authenticate the user; and

based on a result of the user authentication, authenticating the voice input.

12. The method of claim 11 , wherein the signal characteristic data comprises information about a per-frequency cumulative power of the voice input.

13. The method of claim 12 , wherein the first learning model is trained to determine the attribute of the voice input differently according to per-frequency cumulative powers of the voice uttered by the person and the voice output by the apparatus.

14. The method of claim 11 , further comprising authenticating the user by,

obtaining a voice input pattern of the user;

wherein the voice input pattern is determined based on the voice input which is input by the user or a situation in which the voice input is input;

applying the voice input pattern to the second learning model configured to authenticate the voice input.

15. The method of claim 14 , wherein the voice input pattern comprises a user's usage behavior of inputting voice.

16. The method of claim 11 , wherein the context information comprises the device state information, and

the device state information comprises information of at least one of an operation mode of the device, a location of the device, a communication module activation state of the device, and a network connection state of the device.

17. The method of claim 11 , wherein the context information comprises the user's device usage history information, and the user's device usage history information comprises at least one of an application usage history, a user's call history, a user's text history, and a usage frequency of a voice recognition function.

18. A non-transitory computer readable medium configured to store instructions, wherein execution of the instructions by one or more processors of a computer are configured to cause the computer to:

obtain, from a voice input, signal characteristic data representing signal characteristics of the voice input;

apply the obtained signal characteristic data to a first learning model configured to determine whether an attribute of the voice input corresponds to a voice uttered by a person or a voice output by an apparatus;

obtain context information including at least one of device state information, user's device usage history information, and user schedule information;

based on a reliability score of determining that the attribute of the voice input corresponds to the voice uttered by the person, determine whether a user authentication is required;

when the user authentication is required, apply the voice input and the context information to a second learning model to authenticate the user; and

based on a result of the user authentication, authenticate the voice input.