KR102641806B1 - Method and apparatus for authenticating user of wearable device - Google Patents

Abstract

A computing device according to one aspect includes at least one memory; and at least one processor, wherein the at least one processor determines whether the user is wearing the wearable device using a photoplethysmographic signal (PPG signal) of the user, and determines whether the user is wearing the wearable device. In response to being determined to be worn, a processed photoplethysmographic signal is generated by processing the photoplethysmographic signal, inputting the processed photoplethysmographic signal as input data of a deep learning model, and using the processed photoplethysmographic signal as output data. The similarity between the photoplethysmographic signal and the photoplethysmographic signal of a previously registered user is calculated to authenticate the user.

Description

Method and apparatus for authenticating user of wearable device {Method and apparatus for authenticating user of wearable device}

This disclosure relates to a method and device for authenticating a user of a wearable device. More specifically, it relates to a method and device for authenticating a user of a wearable device using a photoplethysmography (PPG) signal.

Recently, various types of wearable devices have been developed and used. Examples include a wrist-worn smart watch that has a communication function and provides various user convenience functions, or smart glasses that allow the user to view a specific screen and process information accordingly.

Since such wearable devices store and process personal information and can also be used as a financial transaction or payment method, technology is needed to allow only authorized users to use the wearable device. In addition, due to the nature of wearable devices, not only authenticated users but also outsiders can easily access them, so a simple and accurate user authentication method is needed.

Conventionally, users of wearable devices were authenticated through fingerprint recognition, iris recognition, vein recognition, etc. However, the above authentication methods required additional actions from the user to authenticate the user. In addition, the conventional technology for authenticating users of wearable devices using photoplethysmography had the problem that it took a long time to extract features for user registration.

Accordingly, in order to solve the above problem, the present invention seeks to provide a method for authenticating the user of a wearable device within a short period of time without requiring additional actions from the user of the wearable device using photoplethysmography.

The object is to provide a method and device for authenticating a user of a wearable device. Additionally, the object is to provide a computer-readable recording medium on which a program for executing the method on a computer is recorded. The technical challenges to be solved are not limited to those described above, and other technical challenges may exist.

According to one aspect of the present disclosure, a method of authenticating a user of a wearable device includes: determining whether the user is wearing the wearable device using a photoplethysmographic signal (PPG signal) of the user; generating a processed photoplethysmographic signal by processing the photoplethysmographic signal in response to determining that the user is wearing the wearable device; And inputting the processed photoplethysmographic signal as input data of a deep learning model, and calculating the similarity between the processed photoplethysmographic signal and the photoplethysmographic signal of a pre-registered user as output data to authenticate the user. A method including ; can be provided.

A computing device according to another aspect of the present disclosure includes a memory storing at least one program; and at least one processor executing the at least one program, wherein the at least one processor determines whether the user is wearing the wearable device using a photoplethysmographic signal (PPG signal) of the user. , In response to determining that the user is wearing the wearable device, generating a processed photoplethysmographic signal by processing the photoplethysmographic signal, and inputting the processed photoplethysmographic signal as input data of a deep learning model. And, as output data, the similarity between the processed photoplethysmographic signal and the photoplethysmographic signal of a pre-registered user can be calculated to authenticate the user.

A computer-readable recording medium according to another aspect of the present disclosure includes a recording medium recording a program for executing the above-described method on a computer.

Security can be improved by continuously and non-consciously extracting photoplethysmographic signals while the wearable device user is wearing the wearable device, and using the extracted signals as data required for user authentication.

Additionally, convenience can be provided to the user by using a remote photoplethysmographic signal without requiring any additional action from the user.

Additionally, since only the photoplethysmographic sensor is used, a separate sensor may not be required for user authentication.

The effects of the examples are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the present invention.

FIG. 1 is a diagram illustrating an example of a method for authenticating a user of a wearable device according to an embodiment.
FIG. 2 is a configuration diagram illustrating an example of a device for authenticating a user of a wearable device according to an embodiment.
FIG. 3 is a flowchart illustrating an example of a method for authenticating a user of a wearable device according to an embodiment.
FIG. 4 is a diagram illustrating an example of removing the tendency of a photoplethysmographic signal according to an embodiment.
FIG. 5 is a diagram illustrating an example of separating a photoplethysmographic signal into a plurality of first sub-signals according to an embodiment.
FIG. 6 is a diagram illustrating an example of generating a plurality of second sub-signals having the same period by interpolating a plurality of first sub-signals according to an embodiment.
FIG. 7 is a diagram illustrating an example of generating a processed photoplethysmographic signal using second sub-signals according to an embodiment.
FIG. 8 is a diagram illustrating an example of calculating the similarity between a photoplethysmography signal processed using a deep learning model according to an embodiment and a photoplethysmography signal of a pre-registered user.

The present disclosure can be implemented in various ways and can have various embodiments, and specific embodiments are illustrated in the drawings and described in the detailed description. The effects and features of the present disclosure and methods for achieving them will become clear with reference to the embodiments described in detail below along with the drawings.

The terms used in the embodiments are general terms that are currently widely used as much as possible, but may vary depending on the intention or precedent of a technician working in the art, the emergence of new technology, etc. In addition, in certain cases, there are terms arbitrarily selected by the applicant, and in this case, the meaning will be described in detail in the relevant description. Therefore, terms used in the specification should be defined based on the meaning of the term and the overall content of the specification, not just the name of the term.

When it is said that a part "includes" a certain element throughout the specification, this means that, unless specifically stated to the contrary, it does not exclude other elements but may further include other elements. In addition, terms such as "~ unit" and "~ module" described in the specification refer to a unit that processes at least one function or operation, which may be implemented as hardware or software, or as a combination of hardware and software.

Additionally, terms including ordinal numbers such as “first” or “second” used in the specification may be used to describe various components, but the components should not be limited by the terms. The above terms may be used for the purpose of distinguishing one component from another component.

Additionally, in the drawings, the sizes of components may be exaggerated or reduced for convenience of explanation. For example, the size and thickness of each component shown in the drawings are arbitrarily shown for convenience, and the present disclosure is not necessarily limited to what is shown.

Hereinafter, the present disclosure will be described in detail with reference to the attached drawings. However, the embodiments may be implemented in various different forms and are not limited to the examples described herein.

FIG. 1 is a diagram illustrating an example of a method for authenticating a user of a wearable device according to an embodiment.

Referring to FIG. 1, an example of a method of extracting a photoplethysmography signal 130 using the wearable device 120 of the user 110 and authenticating the user using the extracted photoplethysmography signal 130 is shown. It is done.

For example, the photoplethysmographic signal 130 may be extracted from a specific body part of the user 110, and the user 110 may be authenticated using the extracted photoplethysmographic signal 130. For example, the photoplethysmographic signal 130 may be extracted from the face of the user 110, but the body part from which the photoplethysmographic signal 130 is extracted may be other body parts such as the wrist, arm, and leg as well as the face. You can.

In addition, since the condition of the skin (e.g., light intensity, temperature, movement, etc.) may vary depending on the body part from which the photoplethysmographic signal 130 is extracted, the photoplethysmographic signal 130 may be generated depending on the different body part. The method of extracting may also vary.

In addition, Figure 1 shows only a watch-shaped wearable device worn on the wrist of the user 110, but this is for convenience of explanation, and wearable devices include not only watches, but also rings, bracelets, anklets, necklaces, earrings, glasses, etc. may include.

FIG. 2 is a configuration diagram illustrating an example of a device for authenticating a user of a wearable device according to an embodiment.

Referring to FIG. 2 , the device 200 may include a communication unit 210, a processor 220, a memory 230, and a sensor unit 240. In the device 200 of FIG. 2, only components related to the embodiment are shown. Therefore, it is obvious to those skilled in the art that other general-purpose components may be included in addition to the components shown in FIG. 2.

Device 200 may be a smartphone, a home appliance, a device equipped with a camera, and other mobile or non-mobile computing devices. Additionally, the device 200 may be a wearable device such as a watch, bracelet, anklet, glasses, hair band, or ring equipped with a data processing function. However, it is not limited to this.

The communication unit 210 may include one or more components that enable wired/wireless communication with an external server or external device. For example, the communication unit 210 may include a short-range communication unit (not shown) and a mobile communication unit (not shown) for communication with an external server or external device.

For example, the communication unit 210 may receive the user's photoplethysmographic signal or other biometric information acquired by the sensor unit 240.

The memory 230 is hardware that stores various data processed within the device 200, and can store programs for processing and control of the processor 220.

For example, the memory 230 may store various data, such as a photoplethysmographic wave signal or other biometric information obtained from the sensor unit 240, or data generated according to the operation of the processor 220. Additionally, the memory 230 may store an operating system (OS) and at least one program (eg, a program necessary for the processor 220 to operate, etc.).

The memory 230 includes random access memory (RAM) such as dynamic random access memory (DRAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), CD- It may include ROM, Blu-ray or other optical disk storage, a hard disk drive (HDD), a solid state drive (SSD), or flash memory.

The sensor unit 240 may include at least one sensor. For example, the sensor unit 240 may include a PPG sensor, an infrared sensor, an LED sensor, an electrostatic sensor, etc., but is not limited thereto.

The sensor unit 240 can be used to obtain various information. For example, the sensor unit 240 may be used to obtain information for authenticating a user. As an example, the sensor unit 240 may acquire the user's photoplethysmographic signal. As another example, the sensor unit 240 may acquire fingerprint information, face information, voice information, iris information, electrocardiogram information, etc., but is not limited to this and may acquire various biometric information.

The processor 220 controls the overall operation of the device 200. For example, the processor 220 executes programs stored in the memory 230 to overall manage the input unit (not shown), display (not shown), communication unit 210, memory 230, sensor unit 240, etc. It can be controlled with .

The processor 220 includes application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, and microcontrollers. It may be implemented using at least one of micro-controllers, microprocessors, and other electrical units for performing functions.

The processor 220 may control the operation of the device 200 by executing programs stored in the memory 230. As an example, the processor 220 may perform at least part of the method for authenticating a user of a wearable device described with reference to FIGS. 3 to 8 .

In other words, the processor 220 uses the user's photoplethysmographic signal (PPG signal) to determine whether the user is wearing the wearable device, and processes the photoplethysmographic signal in response to determining that the user is wearing the wearable device. By doing so, a processed photoplethysmographic signal is generated, the processed photoplethysmographic signal is input as input data of the deep learning model, and the similarity between the processed photoplethysmographic signal and the already registered user's photoplethysmographic signal is calculated as output data. You can calculate and authenticate the user.

First, the processor 220 can determine whether the user is wearing a wearable device using the user's photoplethysmographic signal (PPG signal).

Additionally, the processor 220 may generate a processed photoplethysmography signal by processing the photoplethysmographic signal in response to determining that the user is wearing a wearable device.

For example, the processor 220 calculates the SNR (Signal to Noise Ratio) value of the photoplethysmographic signal, extracts the photoplethysmographic signal whose SNR value is greater than or equal to a preset threshold, and trends the extracted photoplethysmographic signal. can be removed. Additionally, the processor 220 separates the photoplethysmography signal from which the tendency has been removed into a plurality of first sub-signals, interpolates the first sub-signals to generate a plurality of second sub-signals having the same period, and generates at least two sub-signals. At least one processed photoplethysmographic signal can be generated using the above second sub-signals.

Here, the deep learning model may include a first model and a second model. For example, the first model may extract the first feature vector from the processed photoplethysmographic signal. In addition, the second model calculates the difference between the first feature vector and the second feature vector, generates a third feature vector representing the difference, and uses the third feature vector to match the user's photoplethysmographic signal and the previously registered The similarity between the user's photoplethysmographic signals can be calculated. Here, the second feature vector may be a feature vector extracted from the photoplethysmography signal of a previously registered user. For example, the processor 220 generates a third feature vector indicating the difference between the first feature vector and the second feature vector, and the second model uses the third feature vector to register the user's photoplethysmography signal and the second model. The similarity between the user's photoplethysmographic signals can be calculated.

Additionally, the processor 220 may determine whether the similarity is within a preset range and authenticate the user in response to determining that the similarity is within the preset range.

Here, the deep learning model can be retrained using the updated first feature vector of the authenticated user as learning data. Additionally, the second feature vector may be updated based on the first feature vector of the authenticated user. Additionally, the second model may calculate the difference between the first feature vector extracted from the photoplethysmographic signal continuously received by the wearable device and the updated second feature vector, and generate a third feature vector representing the difference. .

FIG. 3 is a flowchart illustrating an example of a method for authenticating a user of a wearable device according to an embodiment.

Referring to FIG. 3, a method for authenticating a user of a wearable device may include steps 310 to 330. However, it is not limited to this, and other general steps in addition to the steps shown in FIG. 3 may be further included in the method of authenticating a user of a wearable device. Additionally, as described above with reference to FIGS. 1 and 2, at least one of the steps in the flowchart shown in FIG. 3 may be processed by the processor 220.

In step 310, the processor 220 may determine whether the user is wearing a wearable device using the user's photoplethysmographic signal.

For example, the processor 220 may determine whether the user is wearing a wearable device based on the presence or absence of the user's photoplethysmographic signal detected by the sensor unit 240. As an example, when the sensor unit 240 detects the user's photoplethysmographic signal, the processor 220 may determine that the user is wearing a wearable device. As another example, if the sensor unit 240 does not detect the user's photoplethysmographic signal, the processor 220 may determine that the user is not wearing the wearable device. When the processor 220 determines that the user is not wearing the wearable device, the sensor unit 240 can detect the user's photoplethysmographic signal again.

Additionally, the processor 220 may determine whether the user is wearing a wearable device according to the user's movement detected by the sensor unit 240. For example, if the sensor unit 240 detects the user's movement and also detects the user's photoplethysmography signal, the processor 220 may determine that the user is wearing a wearable device.

Referring again to FIG. 3 , in step 320, the processor 220 may generate a processed photoplethysmography signal by processing the photoplethysmographic signal in response to determining that the user is wearing a wearable device.

First, the processor 220 may calculate the Signal to Noise Ratio (SNR) value of the photoplethysmographic signal and extract the photoplethysmographic signal whose SNR value is greater than or equal to a preset threshold.

For example, the processor 220 may calculate the SNR value of the photoplethysmographic signal using Equation 1. Equation 1 is as follows:

From here, means the power value when the photoplethysmographic signal is converted to a frequency band using Fourier transform. also, is the maximum power value of the photoplethysmographic signal at 0.8 to 3 Hz (about 50 to 180 bpm) corresponding to the heart rate band, means the sum of the power values of the photoplethysmographic signal in the entire frequency band.

The processor 220 can extract only photoplethysmographic signals whose SNR value calculated using Equation 1 is equal to or greater than a preset threshold. Here, the preset threshold may be 0.8, but is not limited thereto. Therefore, the processor 220 can extract only photoplethysmographic signals whose SNR value calculated using Equation 1 is 0.8 or more.

Additionally, the processor 220 may remove the tendency of the extracted photoplethysmographic signal.

Hereinafter, with reference to FIG. 4, an example in which the processor 220 removes the tendency of the extracted photoplethysmography signal will be described.

FIG. 4 is a diagram illustrating an example in which the processor 220 removes the tendency of a photoplethysmographic wave signal according to an embodiment.

Referring to FIG. 4, the processor 220 may process the extracted photoplethysmography signal 410 into a photoplethysmography signal 420 from which the tendency has been removed.

Here, trend removal may be one of the data processing methods that removes the linear trend of time series data. The overall change of a signal from which the trend has been removed is reduced, making it possible to accurately and easily find the maximum and minimum values of the signal. In addition, the horizontal axis of the graphs of the extracted photoplethysmographic signal 410 and the detrended photoplethysmographic signal 420 shown in FIG. 4 may represent time (seconds), and the vertical axis may represent amplitude (millimeter).

The processor 220 may use a detrending operation to remove the tendency of the extracted photoplethysmographic signal 410. The processor 220 normalizes the photoplethysmography signal 420 from which the trend has been removed using a detrending operation to generate a photoplethysmographic signal with an average of 0. Accordingly, the processor 220 can accurately and easily find the maximum and minimum values of the normalized photoplethysmographic signal.

Additionally, the processor 220 may separate the photoplethysmography signal from which the tendency has been removed into a plurality of first sub-signals.

Hereinafter, with reference to FIG. 5 , an example in which the processor 220 separates the photoplethysmography signal from which the tendency has been removed into a plurality of first sub-signals will be described.

FIG. 5 is a diagram illustrating an example in which the processor 220 separates a photoplethysmographic wave signal into a plurality of first sub-signals according to an embodiment.

Referring to FIG. 5 , the processor 220 may separate the photoplethysmography signal 510 from which the tendency has been removed into first sub-signals 520 .

Here, the first sub-signal 520 refers to a single period signal, that is, a signal with one period. In addition, the horizontal axis of the graph of the photoplethysmographic signal 510 from which the trend shown in FIG. 5 has been removed represents time (seconds), the vertical axis represents amplitude (millimeter), and the horizontal axis of the graph of the first sub-signal 520 represents the sample. The number (units) and the vertical axis may indicate amplitude (millimeter).

For example, the processor 220 detects the inflection points that exist below the normalized photoplethysmographic signal from which the tendency has been removed, that is, the peak values of the photoplethysmographic signal and separates them into single period signals. . In other words, the processor 220 detects peak values of the photoplethysmographic signal 510 from which the tendency has been removed, and separates the signal between two adjacent peak values into the first sub-signal 520.

Here, the processor 220 can set parameters and detect peak values. For example, the processor 220 may detect peak values using the minimum distance and minimum value as parameters.

The processor 220 may generate a plurality of second sub-signals having the same period by interpolating the first sub-signals.

Hereinafter, with reference to FIG. 6, an example of a method in which the processor 220 interpolates first sub-signals to generate a plurality of second sub-signals having the same period will be described.

FIG. 6 is a diagram illustrating an example in which the processor 220 according to an embodiment generates a plurality of second sub-signals having the same period by interpolating a plurality of first sub-signals.

Referring to FIG. 6, the processor 220 may generate second sub-signals 610 by interpolating the first sub-signals 620 into interpolation signals 630.

Here, the upper horizontal axis of the graph shown in FIG. 6 is the number of samples of the second sub-signal 610, the lower horizontal axis is the number of samples of the first sub-signal 620, and the vertical axis is the amplitude (millimeter). It can mean.

Since the user's heart rate is different for each heartbeat, there is also a difference for each cycle of the user's photoplethysmographic signal. If there is a difference for each period, the average value cannot be obtained at the same point in each period, so a signal with the same period is needed.

Accordingly, the processor 220 uses interpolation to generate a signal of the same length, that is, a second sub-signal 610 of the same period. Here, the processor 220 may generate the second sub-signal 610 using quadratic spline interpolation. Quadratic spline interpolation is an interpolation method that forms a curve through a plurality of data points and is a method of estimating values interpolated between each data point. However, examples of the interpolation method used by the processor 220 are not limited to this.

Additionally, the processor 220 may generate at least one processed photoplethysmography signal using at least two or more second sub-signals.

Hereinafter, with reference to FIG. 7 , an example in which the processor 220 generates a processed photoplethysmographic signal using the second sub-signals will be described.

FIG. 7 is a diagram illustrating an example of generating a processed photoplethysmographic signal using second sub-signals according to an embodiment.

Referring to FIG. 7 , the processor 220 may generate one processed photoplethysmography signal 720 using a plurality of second sub-signals 710.

Here, the horizontal axis of the graph shown in FIG. 7 may indicate the number of samples of the processed photoplethysmography signal 720, and the vertical axis may indicate the amplitude (cm).

For example, the processor 220 may generate one processed photoplethysmography signal 720 by overlapping the plurality of second sub-signals 710 and calculating an average. Here, the processor 220 may generate one processed photoplethysmographic signal 720 by overlapping five second sub-signals 710, but the number of second sub-signals 710 is not limited thereto. .

For example, the processor 220 may calculate the average of the plurality of second sub-signals 710 using ensemble average. Here, the ensemble average means calculating a plurality of sample averages while fixing the time axis of the photoplethysmographic signal. Additionally, the processor 220 may calculate the ensemble average of the plurality of second sub-signals 710 using Equation 2. Equation 2 is as follows:

From here, refers to the generated processed photoplethysmographic signal 720, means the number of overlapping second sub-signals 710, means the period of the second sub-signal 710, of the processed photoplethysmographic signal (720) refers to the second sample.

Accordingly, the processor 220 may calculate the average of the plurality of second sub-signals 710 using Equation 2 and generate one processed photoplethysmographic signal 720 representing the calculated average.

Referring again to FIG. 3, in step 330, the processor 220 inputs the processed photoplethysmographic signal as input data of the deep learning model, and inputs the processed photoplethysmographic signal and the pre-registered user's photoplethysmographic signal as output data. The user can be authenticated by calculating the similarity of the pulse wave signal.

For example, the processor 220 may determine whether the similarity is within a preset range and authenticate the user in response to determining that the similarity is within the preset range.

Here, the deep learning model may include a first model and a second model. For example, the first model may extract the first feature vector from the processed photoplethysmographic signal. For example, the second model may calculate the difference between the first feature vector and the second feature vector and generate a third feature vector representing the difference. Additionally, the second model may calculate the similarity between the user's photoplethysmographic signal and the photoplethysmographic signal of a pre-registered user using the third feature vector. Here, the second feature vector may be a feature vector extracted from the photoplethysmography signal of a previously registered user. For example, the processor 220 generates a third feature vector indicating the difference between the first feature vector and the second feature vector, and the second model uses the third feature vector to register the user's photoplethysmography signal and the second model. The similarity between the user's photoplethysmographic signals can be calculated.

Hereinafter, with reference to FIG. 8, an example of a method by which the processor 220 authenticates a user using a deep learning model will be described.

FIG. 8 is a diagram illustrating an example of calculating the similarity between a photoplethysmographic wave signal processed using a deep learning model and a photoplethysmographic signal of a pre-registered user, according to an embodiment.

Referring to FIG. 8, the processor 220 inputs the processed photoplethysmographic signal 810 as input data of the deep learning model 820 and obtains the similarity 830 as output data.

First, the deep learning model 820 may include a first model and a second model. Here, the first model and the second model do not each mean one deep learning model, but may each perform different operations as partial components or parts of one deep learning model 820. For example, the first model may be a stack of several one-dimensional convolutional layers, and the second model may be a stack of several linear layers.

For example, the first model may extract the first feature vector from the processed photoplethysmographic signal. In other words, when the processor 220 inputs the processed photoplethysmographic signal 810 described above with reference to FIGS. 4 to 7 as input data of the deep learning model 820, the first model is the processed photoplethysmographic pulse wave. A first feature vector expressing the characteristics of the photoplethysmographic signal 810 processed from the signal 810 may be generated.

For example, the second model may calculate the difference between the first feature vector and the second feature vector. Here, the second feature vector may be a feature vector extracted from the photoplethysmographic signal of a pre-registered user, or may be a feature vector previously stored in the second model.

As an example, the second model may generate a third feature vector representing the difference between the first feature vector and the second feature vector, and use the third feature vector to match the user's photoplethysmography signal and the previously registered user's photoplethysmographic signal. Similarity between photoplethysmographic signals can be calculated. As another example, the processor 220 generates a third feature vector indicating the difference between the first feature vector and the second feature vector, and the second model is previously registered with the user's photoplethysmographic signal using the third feature vector. The similarity between the user's photoplethysmographic signals can be calculated. Here, the second model can output the similarity 830 as a value between 0 and 1 using a predetermined function.

For example, the processor 220 may determine whether the output similarity 830 is within a preset range and authenticate the user in response to determining that the output similarity 830 is within the preset range. there is. Here, the preset range may mean a range between 0.7 and 1, but is not limited thereto.

For example, a deep learning model can be retrained using the updated first feature vector of an authenticated user as training data. In other words, the deep learning model can be retrained by updating the first feature vector on which the processor 220 authenticated the user with learning data.

For example, the second feature vector may be updated based on the first feature vector of the authenticated user. In other words, the processor 220 can use the first feature vector that served as the basis for authenticating the user as the second feature vector in the subsequent user authentication process.

As an example, the second model may calculate the difference between the first feature vector extracted from the photoplethysmographic signal continuously received by the wearable device and the updated second feature vector, and generate a third feature vector representing the difference. there is. In other words, even when there is no user authentication request, the processor 220 continuously extracts the user's photoplethysmographic signal and generates the first feature vector and the third feature vector to optimize the user authentication process for the authenticated user. You can.

As another example, the processor 220 generates a third feature vector representing the difference between the first feature vector extracted from the photoplethysmographic signal continuously received by the wearable device and the updated second feature vector, and the second model is The user authentication process can be optimized for authenticated users by using the third feature vector.

According to the above description, the present invention can authenticate a user of a wearable device without using a contact device.

Additionally, the present invention can provide an optimized user authentication process to authenticated users by continuously extracting and updating photoplethysmography.

Meanwhile, the above-described method can be written as a program that can be executed on a computer, and can be implemented in a general-purpose digital computer that operates the program using a computer-readable recording medium. Additionally, the data structure used in the above-described method can be recorded on a computer-readable recording medium through various means. The computer-readable recording media includes storage media such as magnetic storage media (e.g., ROM, RAM, USB, floppy disk, hard disk, etc.) and optical read media (e.g., CD-ROM, DVD, etc.) do.

Those skilled in the art related to the present embodiment will understand that the above-described substrate can be implemented in a modified form without departing from the essential characteristics. Therefore, the disclosed methods should be considered from an explanatory rather than a limiting perspective, and the scope of rights is indicated in the claims, not the foregoing description, and should be interpreted to include all differences within the equivalent scope.

110: user
120: Wearable device
130: Photoplethysmographic signal

Claims (7)

In a method for authenticating a user of a wearable device,
determining whether the user is wearing the wearable device using the user's photoplethysmographic signal (PPG signal);
generating a processed photoplethysmographic signal by processing the photoplethysmographic signal in response to determining that the user is wearing the wearable device; and
Inputting the processed photoplethysmographic signal as input data of a deep learning model, and calculating a similarity between the processed photoplethysmographic signal and the photoplethysmographic signal of a pre-registered user as output data to authenticate the user; Including,
The deep learning model includes a first model and a second model,
The first model extracts a first feature vector from the processed photoplethysmographic signal,
The second model calculates a difference between the first feature vector and a second feature vector extracted from the photoplethysmographic signal of the registered user, generates a third feature vector representing the difference, and the third feature vector A method of calculating the similarity using . According to paragraph 1,
The step of generating the processed photoplethysmographic signal is,
Calculating a signal to noise ratio (SNR) value of the photoplethysmographic signal and extracting the photoplethysmographic signal whose SNR value is equal to or greater than a preset threshold;
removing the tendency of the extracted photoplethysmographic signal;
separating the photoplethysmographic signal from which the tendency has been removed into a plurality of first sub-signals;
generating a plurality of second sub-signals having the same period by interpolating the first sub-signals; and
generating at least one of the processed photoplethysmographic signals using at least two of the second sub-signals;
Method, including. According to paragraph 1,
The authentication step is,
determining whether the similarity is within a preset range and authenticating the user in response to determining that the similarity is within the preset range;
Method, including. According to paragraph 1,
The method wherein the deep learning model is retrained using the updated first feature vector of the authenticated user as learning data. According to paragraph 1,
The second feature vector is updated based on the first feature vector of the authenticated user,
The second model is,
A method for calculating a difference between a first feature vector extracted from a photoplethysmographic signal continuously received by the wearable device and the updated second feature vector, and generating a third feature vector representing the difference. In a device for authenticating a user of a wearable device,
a memory in which at least one program is stored; and
At least one processor executing the at least one program,
The at least one processor,
Determine whether the user is wearing the wearable device using the user's photoplethysmographic signal (PPG signal),
In response to determining that the user is wearing the wearable device, generating a processed photoplethysmography signal by processing the photoplethysmographic signal,
Inputting the processed photoplethysmographic signal as input data of a deep learning model, and calculating the similarity between the processed photoplethysmographic signal and the photoplethysmographic signal of a pre-registered user as output data to authenticate the user;
The deep learning model includes a first model and a second model,
The first model extracts a first feature vector from the processed photoplethysmographic signal,
The second model calculates a difference between the first feature vector and a second feature vector extracted from the photoplethysmographic signal of the registered user, generates a third feature vector representing the difference, and the third feature vector A device for calculating the similarity using . A computer-readable recording medium recording a program for executing the method of claim 1 on a computer.