KR20200092779A - Voice-based identity authentication and digital signature - Google Patents

Abstract

The present invention relates to a method for authenticating a talker (speaking person) who instructed a correct command to perform sensitive commands such as ordering goods without errors, and further to authenticate the person by voice, and to a device thereof. The method of the present invention processes the primary identity authentication through the caller identity verification using a voice password, and secondly performs identification using speech recognition equipment. The method of the present invention uses voiceprint analysis as an aid, and identifies the person as a clear person by using a key built into a smartphone or device. And, by measuring the internal sound or vibration, the fact that the person said the same was certified, uniqueness and stability are secured at the same time. Through the same, problems such as order and payment errors are confirmed, problems such as an incorrect order situation caused by an external voice can be solved by verifying the identity and the speaker.

Description

Voice-based identity authentication and digital signature

The present invention is intended to present a method for authenticating a speaker who instructed an exact command, performing sensitive commands such as ordering an object without error, and further authenticating himself by voice.

In addition to global companies such as Google and Amazon, Naver in Korea is also showing off voice AI. As interest in home IoT becomes very high, voice AI is becoming more and more common, such as operating devices with voice and ordering items. However, problems such as listening to TV and Amazon's voice AI delivering pizza to American homes are occurring. to be. To supplement this, the analysis of the gates was added, but there is a problem that the gates cannot be a value representing uniqueness.

In the case of NEC Acoustic Authentication, both earphones are used to check the degree of resonance in the ear and use it for the purpose of personal authentication. NEC's Acoustic Authentication method measures resonance in both earphones after sending sound of a specific frequency from both earphones. Through this, information about the space in the ear is checked to identify the person. There is not yet enough data available on the newly proposed method, the possibility of being mistaken with others, or the factors that influence it. Therefore, it is considered that it still takes time to be in full use.

The problem to be solved by the present invention is to reduce the possibility of errors in voice-based identity authentication and voice payment.

In order to solve the above-described problem, the identity of the talker is authenticated based on the voice message through the following steps.

-When the talker speaks, check whether the sound comes from the talker or from outside

Do it. As a method for this, use a device that senses the vibration of the speaker's neck or face or a microphone that measures the sound inside. The shape of the equipment is

Phones, necklaces, headsets, glasses, etc. are possible.

-Receive sound through microphone equipment.

I can receive input through other microphone equipment.

-Recognize the speaker's sound pattern. Using deep learning technology, the speaker's sound is converted into text information.

-Confirm that the speaker is the person. After confirming that the speaker's sound is a keyword that has been input in advance, analyze the gates using deep learning technology to confirm the person's identity once more.

-Sends an encryption message from the device to a communication device such as a smartphone. It uses the key information contained in the device to transmit the identification information of the talker to authenticate itself. example

It can be used for electronic signature through authentication, or for payment approval.

At this time, the method of authenticating the user by using the key of the device can be divided into two methods according to the performance of the device or the required security level.

-Option 1: Store the user authentication secret key on the device

- Send random numbers to equipment where authentication is requested

- Equipment that receives a random number sends a signature using an internal secret key

- Verify the registered device by checking with the public key on the outside to complete the user authentication.

-Option 2: Store the authentication key in the smartphone\

- Send a random number to the smartphone where authentication is requested.

- Ask whether to allow authentication from the smartphone to the device\

- The device sends an authentication signal to the smartphone

- When authentication is confirmed, sign the random number given by the smartphone

Send information

- Check the authenticity of the signature against the public key.

First, the present invention can be used for payment solutions for the visually impaired using voice. The visually impaired person can be used to perform banking tasks such as remittance freely by authenticating himself or herself through voice.

Second, a convenient shopping mall payment method is presented. For example, it can be used for in-game payments. In a situation where real-time payment is required, such as when an item needs to be purchased urgently during the game, real-time payment is possible using voice.

1 is an example of speech recognition equipment according to a preferred embodiment of the present invention.
2 is an example of a method for inputting a speaker's sound according to a preferred embodiment of the present invention.
3 is a flowchart illustrating a process of recognizing a voice pattern according to a preferred embodiment of the present invention.
4 is a flow chart illustrating a process of confirming a speaker's identity using a voice password.
5 is a flow chart illustrating the process of the identification and payment process using the speech recognition equipment.
6 is a flowchart for explaining method 1 of the user authentication method using the secret key of the speech recognition device.
7 is a flowchart for explaining method 2 of the user authentication method using a secret key of a smartphone.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 illustrates an example of a speech recognition equipment.

Speech equipment includes (1-1) headsets, (1-2) earphones, (1-3) glasses, and (1-4) necklaces/neckbands.

Utilizing these utterance equipment, when the talker speaks, it checks whether the sound comes from the talker or from outside. As a method for this, use a device that senses the vibration of the speaker's neck or face or a microphone that measures the sound inside.

2 illustrates an example of a voice input method.

(2-1) Headset, (2-2) ear microphone, and (2-3) general microphone can receive voice input.

3 illustrates a voice pattern recognition method.

(3-1) When the talker delivers the voice, the waveform of the (3-2) sound of the delivered voice is recognized by using (3-3) deep learning. This is converted into (3-4) character information.

4 illustrates a method of authenticating a speaker using a voice password.

(4-1) When the speaker pronounced "Crow" as the voice password, (4-2) the pre-entered voice password "Crow" and (4-3) the speaker's voice password "Crow" (4) -4) Check whether the voice passwords match, and if they match, it is judged that (4-6) authentication was successful. Otherwise, it is considered as (4-5) authentication failure. In addition, the analysis of the gates is conducted to confirm the identity of the person once more.

5 is a description of the identity verification and payment process using the speech recognition equipment.

Authenticate by sending the identification information of the utterance to the (5-6) personal authentication equipment by using the information in the ignition recognition equipment (5-1,2,3,4,5). (5-6) Through personal authentication, it is used for (5-7) electronic signature, or (5-8) payment approval.

At this time, the method of authenticating the user by using the key of the device can be divided into two methods depending on the performance of the device or the required security level. 6 is a description of Method 1, and FIG. 7 is Method 2;

6 is a description of a user authentication method using a secret key of a speech recognition device.

(6-1,2) In the state where the secret key for authentication is stored in the speech recognition device, the (6-3) authentication requesting authority sends a random number to the (6-1) speech recognition device for authentication request. (6-1,2) The speech recognition equipment sends an electronic signature to the (6-3) certification requesting institution using an internal secret key. (6-3) The certification requesting institution checks whether it is a registered device by comparing it with the (6-5) public key on the outside of (6-4) to complete authentication.

7 is a description of a user authentication method using a secret key of a smartphone.

(7-4) The authentication request agency sends a random number to the (7-2) smartphone for authentication request. (7-2) The smartphone (7-1) asks whether or not to allow authentication to the ignition recognition device. (7-1) The authentication signal is sent from the speech recognition device to the (7-2) smartphone. When the authentication is confirmed, (7-2) digitally sign the random number given by the smartphone and send it to the (7-4) authentication requesting institution. (7-4) The certification requesting institution verifies the authenticity of the signature against the (7-5) public key. Through this, the user is checked for authentication.

1-1: Headset
1-2: earphone
1-3: glasses
1-4: necklace/neckband
2-1: Headset
2-2: Ear microphone
2-3: normal microphone
3-1: Talker (Speaker)
3-2: Waveform of sound
3-3: Speech Pattern Recognition Using Deep Learning
3-4: Character information conversion
4-1: Talker
4-2: Pre-entered voice password
4-3: speaker pronunciation password
4-4: Voice password matching confirmation unit
4-5: Authentication failed
4-6: Successful authentication
5-1: Headset
5-2: Earphone
5-3: glasses
5-4: necklace/neckband
5-5: Secret key for user authentication
5-6: identity verification
5-7: Electronic signature
5-8: Payment
6-1: Speech recognition equipment
6-2: Secret key
6-3: Certification request authority
6-4: Public key
6-5: Registered Equipment Confirmation Unit
7-1: Speech recognition equipment
7-2: Smartphone
7-3: Secret key for user authentication
7-4: Certification request organization
7-5: Public key
7-6: Registration equipment existence confirmation

Claims (3)

-(4-1) Confirm (4-4) match through voice password to confirm the caller
Throat (4-5,6) identifying the person;
-(5-1,2,3,4,5) authenticating yourself using the speech recognition equipment (5-6,7,8);
-(6-1,2) Step of authenticating the user by using the secret key of the speech recognition device (6-3,4,5);
-Authenticate yourself using (7-1) speech recognition equipment and (7-2) smartphone (7-3) secret key
Steps (7-4,5,6);
(e) The method according to claim 1, wherein the authentication in steps (a) and (c) is satisfied at the same time, it is recognized as full authentication. Therefore, the step of passing through the self-identification procedure of step (a) and the identity verification procedure of step (c) in double
(f) The method of claim 1, wherein the authentication in steps (a) and (d) is satisfied at the same time, and thus, it is recognized as full authentication. Therefore, the step of passing through the self-identification procedure of step (a) and the identity verification procedure of step (d) in double

Images