KR20230113942A - A method for performing user authentication according to handwriting input based on artificial intelligence and an apparatus for performing user authentication according to handwriting input based on artificial intelligence - Google Patents

Abstract

A method for performing user authentication according to handwriting input based on artificial intelligence according to an embodiment includes receiving a user authentication request for identifying a user; acquiring a user's handwriting input; obtaining stroke attribute information of the handwriting input; obtaining text corresponding to the handwritten input using an artificial intelligence-based optical character reading model; and processing the user authentication request by using the obtained text and the stroke attribute information together.

Description

A method for performing user authentication according to handwriting input based on artificial intelligence and an apparatus for performing user authentication according to handwriting input based on artificial intelligence}

The present invention relates to a method for performing user authentication according to handwriting input based on artificial intelligence and an apparatus for performing user authentication according to handwriting input based on artificial intelligence. A method for performing user authentication based on handwriting input based on artificial intelligence and an apparatus for performing user authentication based on handwriting input based on artificial intelligence, which provide a method for authenticating a user based on artificial intelligence.

In recent years, as the Internet and the like are widely spread, the number of web sites operated by a membership system is increasing, and the use of electronic payment or online-based banking services is also rapidly increasing.

In general, a general membership-based web site is operated in such a way that user authentication is performed by collecting IDs and passwords of subscribers and checking IDs and passwords of users who want to log in to the web site.

However, in this user authentication method, the user must always remember the ID or password in that the user must input the ID or password whenever logging in.

In addition, efforts have been made to find a method of automatically authenticating and verifying electronic documents. This includes character-based text recognition of images and electronic document recognition processes, which match features or text of an electronic document with equivalent information in a database to authenticate a user.

Therefore, there is a need to improve the precision of user authentication by combining it with a method of automatically authenticating an electronic document while a user easily accesses user authentication by solving the above-mentioned inconvenience of user authentication.

The present invention has been derived from the above-described need, and an object of the present invention is to provide a method in which a user can be accurately identified by an artificial intelligence-based model even if the user simply performs handwriting input using an artificial intelligence model. .

According to the present invention, an object of the present invention is to accurately identify a user by utilizing stroke attribute information of a user's handwriting input together.

A method for performing user authentication according to handwriting input based on artificial intelligence according to an embodiment includes receiving a user authentication request for identifying a user; acquiring a user's handwriting input; obtaining stroke attribute information of the handwriting input; obtaining text corresponding to the handwritten input using an artificial intelligence-based optical character reading model; and processing the user authentication request by using the obtained text and the stroke attribute information together.

A method for performing user authentication according to handwriting input based on artificial intelligence according to an embodiment includes receiving a user authentication request for identifying a user; acquiring a user's handwriting input; obtaining stroke attribute information of the handwriting input;

acquiring text corresponding to stroke attribute information of the handwriting input using an artificial intelligence-based handwriting input analysis model; and processing the user authentication request by using both the acquired text and stroke attribute information of the handwriting input.

The artificial intelligence-based handwriting input analysis model may be an RNN-based model.

The stroke attribute information of the handwriting input may include at least one of stroke weight, number of strokes, font body, size, trajectory direction, generation order, position, pen pressure, and stroke speed of the handwriting input.

The user authentication request processing step may include assigning different weights to the obtained text and the stroke attribute information; and processing the user authentication request based on the weighting result.

The assigning of different weights may include assigning different weights to the obtained text and the stroke attribute information according to the acquisition type of the handwriting input.

The assigning different weights may include assigning different weights to the obtained text and the stroke attribute information based on a predetermined security level of an operation corresponding to the user authentication request.

The user authentication request processing step may include assigning different weights to the stroke attribute information according to the acquisition type of the handwriting input; and processing the user authentication request based on the weighting result.

The processing of the user authentication request may include assigning different weights to the stroke attribute information based on a predetermined security level of an operation corresponding to the user authentication request; and processing the user authentication request based on the weighting result.

The user authentication request processing step may include checking a predetermined security level of an operation corresponding to the user authentication request; selecting stroke attribute information to be considered for user authentication from among stroke attribute information of the handwriting input based on the predetermined security level; and processing the user authentication request based on the selected stroke attribute information.

An apparatus for performing user authentication according to handwriting input based on artificial intelligence according to an embodiment includes a processor; and a memory for storing instructions executable by the processor, wherein the processor, by executing the instructions,

Receive a user authentication request for identifying a user, obtain a user's handwriting input, obtain stroke attribute information of the handwriting input, and obtain text corresponding to the handwriting input using an artificial intelligence-based optical character reading model and processing the user authentication request by using the obtained text and the stroke attribute information together.

An apparatus for performing user authentication according to handwriting input based on artificial intelligence according to an embodiment includes a processor; and a memory for storing instructions executable by the processor, wherein the processor, by executing the instructions,

Receiving a user authentication request for identifying a user, obtaining a user's handwriting input, acquiring stroke attribute information of the handwriting input, and responding to the stroke attribute information of the handwriting input using an artificial intelligence-based handwriting input analysis model The user authentication request may be processed using the obtained text and the stroke attribute information of the handwriting input together.

According to the present invention, the user can be accurately identified by the artificial intelligence-based model even if the user simply performs handwriting input using the artificial intelligence model.

In particular, the user can be accurately identified by utilizing the stroke attribute information of the user's handwriting input.

1 is a block diagram of an apparatus 1 for performing user authentication according to handwriting input based on artificial intelligence according to an embodiment.
2 and 3 are flowcharts illustrating a method of performing user authentication according to handwriting input based on artificial intelligence according to an embodiment.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The detailed description of the present invention which follows refers to the accompanying drawings which illustrate, by way of illustration, specific embodiments in which the present invention may be practiced. These embodiments are described in sufficient detail to enable one skilled in the art to practice the present invention. It should be understood that the various embodiments of the present invention are different from each other but are not necessarily mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be implemented in one embodiment in another embodiment without departing from the spirit and scope of the invention. Additionally, it should be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the invention. Accordingly, the detailed description set forth below is not to be taken in a limiting sense, and the scope of the present invention, if properly described, is limited only by the appended claims, along with all equivalents as claimed by those claims. Like reference numbers in the drawings indicate the same or similar function throughout the various aspects.

1 is a block diagram of an apparatus 1 for performing user authentication according to an embodiment.

As shown in FIG. 1 , an apparatus 1 for performing user authentication according to an embodiment may include a processor 10, a memory 20, a touch screen 30, a camera 40, and a sensor 50. there is.

The memory 20 may store various programs and data necessary for the operation of the user authentication device 1 . The memory 20 according to the embodiment may include an optical character reading model 21 and a handwriting input analysis model 22 .

The optical character reading model 21 may be a model implemented to acquire text based on artificial intelligence-based OCR technology from image information such as characters, symbols, and figures included in handwritten input.

The handwriting input analysis model 22 may be implemented to receive stroke attribute information of the handwriting input and obtain corresponding text.

The processor 10 controls the overall operation of the device 1 for performing user authentication.

According to an embodiment, the processor 10 receives a user authentication request for identifying a user, obtains a user's handwriting input, and obtains text corresponding to the handwriting input using an artificial intelligence-based optical character reading model. , The user authentication request may be processed based on the obtained text.

According to an embodiment, the processor 10 receives a user authentication request for identifying a user, obtains a user's handwriting input, obtains stroke attribute information of the handwriting input, and generates an artificial intelligence-based handwriting input analysis model. The text corresponding to the stroke attribute information of the handwriting input may be obtained by using the stroke attribute information of the handwriting input, and the user authentication request may be processed using the obtained text and the stroke attribute information of the handwriting input together.

The touch screen 30 may perform a display function for output information output from the user authentication device 1 . The touch screen 30 may receive and display a handwriting input drawn on the screen with a finger or while holding a pen for electronic signature verification.

The touch screen 30 may receive various handwriting inputs such as single or multi-touch inputs, drag inputs, and drawing inputs by the user using various input tools such as fingers and pens.

The handwriting input through the touch screen 30 may include one of a direct touch input in which an object is directly touched and a hovering input which is an indirect touch input in which an object approaches within a preset recognition distance but does not directly touch the object. That is, the present invention can be equally/similarly applied even in the case of generating a handwriting input by a hovering input.

The touch screen 30 may include a sensor 50 such as a touch sensor and a pressure sensor for sensing a touch and/or touch pressure of an input tool on the screen.

Specifically, the sensor 50 may obtain touch information for identifying the touch position of the input tool on the screen of the touch screen 30 and the width of the touch. The touch information obtained by the sensor 50 is transmitted to the processor 10, and the processor 10 identifies the touch location of the input tool and the width of the touch based on the transmitted touch information.

Methods for the sensor 50 to sense a touch on the screen of the touch screen 30 include a resistive method, a capacitive method, an ultrasonic (SAW) method, an infrared (IR) method, and the like. The sensor 140 may include a sensing component required for at least one of the aforementioned touch sensing methods.

The camera 40 may acquire an image frame such as a still image or a moving image through an image sensor in a video call mode or a photographing mode. Images captured through the image sensor may be processed through the processor 10 or a separate image processing unit (not shown).

2 and 3 are flowcharts illustrating a method of performing user authentication according to handwriting input based on artificial intelligence according to an embodiment.

For reference, in each flowchart of the present invention, each step is an example, and the present invention can be equally/similarly applied even when each order is differently changed and/or combined.

[Embodiment 1]

Referring to FIG. 2 , the processor 10 may receive a user authentication request for identifying a user (s1). A user authentication request for identifying a user may be input for the user to execute a predetermined operation. That is, it may be an authentication request for determining whether or not there is authority to execute a predetermined operation.

The predetermined operation may include, for example, locking/unlocking the user authentication device 1 or executing a specific application. According to an embodiment, certain operations may be set to certain security levels. For example, an operation such as locking/unlocking may be set to a relatively low security level, whereas an application execution operation such as payment or official document processing may be set to a relatively high security level. However, this is an example, and the present invention may be applied identically/similarly to authentication requests for various purposes for identifying a user.

The processor 10 may obtain a user's handwriting input (s2).

A user's handwriting input according to an embodiment may be input through a pressure-sensitive or capacitive user interface. In addition, the user's handwriting input may be input using not only the user's finger but also a writing tool such as a stylus pen.

According to the embodiment, the handwriting input may be input to the user authentication device 1 by capturing the handwriting input by a camera provided in the user authentication device 1 . For example, the device 1 for performing user authentication may analyze a captured image including the handwriting input and extract the handwriting input from the captured image, so that the handwriting input may be input to the device 1 for performing user authentication. Not limited to the above example, the user may input a handwriting input to the user authentication apparatus 1 through various methods.

A user's handwriting input according to an embodiment may correspond to a handwritten signature, but the present invention is not limited thereto.

The present invention can be equally/similarly applied to the case where the user's handwriting input according to the embodiment is generated through real-time input, or when a pre-generated handwriting input is acquired from the memory 20 . In the latter case, for example, it may be exemplified that a trace of a font body corresponding to a handwritten signature received from a user is digitally imaged and stored.

The processor 10 may obtain stroke attribute information corresponding to each handwriting input (s3).

According to the embodiment, a stroke may refer to a trajectory drawn while maintaining a touch from the time when an input tool (eg, a finger, a stylus pen, etc.) touches the user authentication device 1 .

According to an embodiment, when a handwriting input is obtained from a captured image, the processor 10 may recognize stroke attribute information based on image analysis. On the other hand, when handwriting input is input by an input tool, stroke attribute information may be recognized without image analysis. In the latter case, according to an embodiment, stroke attribute information may be recognized using a combination of the sensor 50 and/or various known algorithms.

According to an embodiment, stroke attribute information may be generated based on locations, motion paths, and intensities associated with contact for any time period.

According to an embodiment, the stroke attribute information may include at least one of stroke weight, number of strokes, font size, size, stroke length, angle, trajectory direction, creation order, position, pen pressure, and stroke speed.

According to the present invention, when it is difficult to identify a user only with text or when more accurate user identification is required, more accurate user identification can be achieved by identifying the user using stroke attribute information in addition to the text corresponding to the handwriting input. .

For example, if the user detected on the text is the same person as the actual authenticated user, it may be difficult to identify the user identification information and the stroke attribute information only with the text. By comparing the text and stroke attribute information obtained from the user identification information stored in the memory 20 and the stroke attribute information mapped thereto, it is possible to more accurately identify whether the user is a valid user.

The processor 10 may obtain text corresponding to the handwriting input using the artificial intelligence-based optical character reading model 21 (s4).

The optical character reading model 21 may be implemented to acquire text based on artificial intelligence-based OCR (Optical Character Recognition) technology from image information such as characters, symbols, and figures included in handwritten input.

The optical character reading model 21 may be generated by labeling text corresponding to a handwriting input for training and learning the neural network model.

The neural network model NM may be composed of a plurality of neural network layers, each layer has a plurality of weights, and an operation of a current layer may be performed through an operation result of a previous layer and a plurality of weights. For example, it may be learned using a learning algorithm including error back-propagation or gradient descent. According to the embodiment, the neural network model may be trained through supervised learning using handwriting input for learning as input data.

The processor 10 may perform user authentication based on the obtained text (s5).

The processor 10 may compare the text obtained from the handwritten signature with user identification information (eg, ID, name, etc.) stored in the memory 20 and perform authentication as a valid user when the text is matched.

According to an embodiment, the processor 10 may perform user authentication by using both the acquired text and the stroke attribute information of the handwriting input.

According to an embodiment, the processor 10 may assign different weights to text and stroke attribute information, and perform user authentication based on weighting results. That is, a higher weight may be assigned to a more important element among text and stroke attribute information, and user authentication may be performed based on a weighted result. According to the embodiment, each weighting criterion and weight value may be preset and stored in the memory 20 .

According to an embodiment, the processor 10 may assign different weights to text and stroke attribute information according to the acquisition type of handwriting input, and perform user authentication based on weighting results.

For example, when a pre-generated handwriting input is obtained from the memory 20, since the pre-generated handwriting input is a standardized image and does not change according to circumstances, a relatively high weight is given to the text and the stroke attribute User authentication may be performed using a result of assigning a relatively low weight to information. On the other hand, when real-time handwriting input is obtained, since the handwriting input obtained in real time is highly likely to change depending on the situation beyond the standardized image, a relatively low weight is given to text and a relatively high weight is given to stroke attribute information. User authentication can be performed using the assigned result.

According to another embodiment, the processor 10 may assign different weights to the stroke attribute information according to the acquisition type of the handwriting input and perform user authentication based on the weighting result.

For example, a pre-generated handwriting input is a standardized image and does not change according to circumstances, whereas a handwriting input acquired in real time is highly likely to change according to situations outside of a standardized image. In this case, authentication of the user may be performed by assigning a relatively higher weight to the stroke attribute information than in the case of a pre-generated handwriting input. According to the embodiment, at this time, the above process may be performed while maintaining the same weight for the text.

According to another embodiment, the processor 10 may assign a different weight to each of at least one of a plurality of stroke attribute information according to the acquisition type of the handwriting input, and perform user authentication based on the weighting result. . That is, the processor 10 may assign different weights to the stroke attribute information in consideration of the acquisition type of handwriting input and the type of stroke attribute information. For example, if the handwriting input is obtained in real time, a high weight is given to the order of creation, the speed of writing, and the direction of the trajectory. On the other hand, if the handwriting input is pre-generated, a low weight is assigned to the order of creation, the speed of writing, and the direction of the trajectory. can

According to another embodiment, the processor 10 assigns different weights to text and stroke attribute information based on a predetermined security level of an operation corresponding to a user authentication request, and performs user authentication based on the weighting result. can do.

For example, if the security level of the operation corresponding to the user authentication request corresponds to a relatively low security level (e.g.> lock/unlock), the text-oriented information of the handwriting input itself rather than the detailed form of the handwriting input is used by the user. Since it is necessary for authentication, user authentication can be performed using a result of assigning a relatively high weight to text and a relatively low weight to stroke attribute information. On the other hand, if the security level of the operation corresponding to the user authentication request corresponds to a relatively high security level (e.g. payment/official document processing, etc.), the detailed form of the handwriting input may have a significant impact on user identification, User authentication may be performed using a result of assigning a relatively low weight to text and a relatively high weight to stroke attribute information.

According to another embodiment, the processor 10 may assign different weights to stroke attribute information according to a predetermined security level of an operation corresponding to a user authentication request, and perform user authentication based on the result of weighting. .

For example, if the security level of the operation corresponding to the user authentication request corresponds to a relatively low security level, since the character-oriented information of the handwriting input itself is required to perform user authentication rather than the detailed form of the handwriting input, the stroke attribute User authentication may be performed using a result of assigning a relatively low weight to information. On the other hand, when the security level of the operation corresponding to the user authentication request corresponds to a relatively high security level, the detailed shape of the handwriting input can have a significant impact on user identification, so a relatively high weight is given to stroke attribute information. User authentication can be performed using the assigned result. According to the embodiment, at this time, the above process may be performed while maintaining the same weight for the text.

According to another embodiment, the processor 10 selects the type and/or number (or ratio) of stroke attribute information to be considered for user authentication according to a predetermined security level of an operation corresponding to a user authentication request, and selects the stroke attribute information. User authentication may be performed based on attribute information.

For example, when the security level of the operation corresponding to the user authentication request corresponds to a relatively low security level, stroke attribute information of less than 30% of the total stroke attribute information is selected, and user authentication is performed based on this can do. On the other hand, if the security level of the operation corresponding to the user authentication request corresponds to a relatively high security level, stroke attribute information of 80% or more of the total stroke attribute information may be selected, and user authentication may be performed based on this.

In addition, when the security level of the operation corresponding to the user authentication request corresponds to a relatively low security level, the stroke thickness, number of strokes, and font font considering the character-oriented information of the handwriting input itself are selected as stroke attribute information, while user authentication When the security level of the operation corresponding to the request corresponds to a relatively high security level, the generation order, stroke stroke, and trajectory direction considering the detailed shape of the handwriting input may be selected as stroke attribute information.

[Embodiment 2]

In the case of Embodiment 2, the content of Embodiment 1 may be applied in the same/similar manner.

Referring to FIG. 3 , the processor 10 receives a user authentication request for identifying a user (s10), obtains a user's handwriting input (s20), and obtains stroke attribute information of the handwriting input. (s30).

s10 to s30 of FIG. 3 may be equally/similarly applied to s1 to s3 of FIG. 2 .

The processor 10 may obtain text corresponding to the stroke attribute information of the handwriting input using the artificial intelligence-based handwriting input analysis model 22 (S40).

Specifically, the processor 10 may recognize a character sequence corresponding to stroke attribute information of the handwriting input using the handwriting input analysis model 22 and obtain text based on the character sequence.

According to an embodiment, stroke attribute information of handwriting input may be input to the handwriting input analysis model 22 . Specifically, the stroke attribute information of the handwriting input can be obtained as information that can be input into the handwriting input analysis model 22, which is obtained through a preprocessing process for the handwriting input (eg, baseline extraction, size correction, slant correction, etc.) can be obtained through execution).

The processor 10 may sequentially input stroke attribute information into the handwriting input analysis model 22 using the handwriting input analysis model 22, thereby recognizing a character sequence corresponding to the input stroke attribute information. Also, text may be obtained based on the recognized character sequence.

The handwriting input analysis model 22 according to the embodiment may be a model based on a recurrent neural network (RNN). That is, the handwriting input analysis model 22 may be a model obtained through learning in which stroke attribute information of the handwriting input for learning is input to the RNN.

RNN has a structure in which the hidden layer forms a chain structure, and there is a loop connected so that the result of the hidden layer goes back to the input of the same hidden layer, and the past data affects the future. That is, in a recurrent neural network (RNN), hidden nodes are connected to directed edges to form a circular structure.

According to the embodiment, the handwriting input analysis model 22 receives the stroke attribute information of the handwriting input for time-sequential learning separated in consideration of the temporal order and determines whether or not to reflect the previous memory cell state in the current memory cell so that a corresponding character sequence is output. It may be composed of a Recurrent Neural Network (RNN)-based layer composed of Long Short Term Memory (LSTM) cells including a forget gate that determines whether

The processor 10 may process the user authentication request by using both the acquired text and stroke attribute information of the handwriting input (S50).

In s50 of FIG. 3 , all user authentication processes using stroke attribute information of text and handwriting input described in s5 of FIG. 2 may be applied in the same/similar manner.

The above-described embodiments may be implemented in the form of program instructions that can be executed through various computer components and recorded on a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination.

Program instructions recorded on the computer-readable recording medium may be specially designed and configured for the present invention, or may be known and usable to those skilled in the art of computer software.

Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magneto-optical media such as floptical disks. optical media), and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter or the like as well as machine language codes such as those produced by a compiler. The hardware device may be configured to act as one or more software modules for executing processes according to the present invention and vice versa.

Aspects herein may take the form of entirely hardware, entirely software (including firmware, resident software, microcode, etc.) or a computer program product embodied entirely in one or more computer readable media having computer readable program code embodied thereon. .

Features, structures, effects, etc. described in the embodiments above are included in one embodiment of the present invention, and are not necessarily limited to only one embodiment. Furthermore, the features, structures, and effects illustrated in each embodiment can be combined or modified with respect to other embodiments by those skilled in the art in the field to which the embodiments belong. Therefore, contents related to these combinations and variations should be construed as being included in the scope of the present invention.

In addition, although the above has been described with a focus on the embodiments, these are only examples and do not limit the present invention, and those skilled in the art to which the present invention belongs can exemplify the above to the extent that does not deviate from the essential characteristics of the present embodiment. It will be seen that various variations and applications that have not been made are possible. For example, each component specifically shown in the embodiment can be modified and implemented. And differences related to these modifications and applications should be construed as being included in the scope of the present invention as defined in the appended claims.

Claims (12)

receiving a user authentication request for identifying a user;
acquiring a user's handwriting input;
obtaining stroke attribute information of the handwriting input;
obtaining text corresponding to the handwritten input using an artificial intelligence-based optical character reading model; and
Processing the user authentication request by using the obtained text and the stroke attribute information together;
A method for performing user authentication according to handwriting input based on artificial intelligence.
receiving a user authentication request for identifying a user;
acquiring a user's handwriting input;
obtaining stroke attribute information of the handwriting input;
obtaining text corresponding to stroke attribute information of the handwriting input using an artificial intelligence-based handwriting input analysis model; and
Processing the user authentication request by using the obtained text and the stroke attribute information of the handwriting input together;
A method for performing user authentication according to handwriting input based on artificial intelligence.
According to claim 2,
The artificial intelligence-based handwriting input analysis model is a model based on RNN,
A method for performing user authentication according to handwriting input based on artificial intelligence.
According to claim 1 or 2,
The stroke attribute information of the handwriting input includes at least one of stroke weight, number of strokes, font body, size, trajectory direction, generation order, position, pen pressure, and stroke speed of the handwriting input.
A method for performing user authentication according to handwriting input based on artificial intelligence.
According to claim 1 or 2,
The user authentication request processing step,
assigning different weights to the obtained text and the stroke attribute information; and
Processing the user authentication request based on the weighting result; including,
A method for performing user authentication according to handwriting input based on artificial intelligence.
According to claim 5,
The different weighting step,
Including, assigning different weights to the acquired text and the stroke attribute information according to the acquisition type of the handwriting input.
A method for performing user authentication according to handwriting input based on artificial intelligence.
According to claim 5,
The different weighting step,
assigning different weights to the obtained text and the stroke attribute information based on a predetermined security level of an operation corresponding to the user authentication request;
A method for performing user authentication according to handwriting input based on artificial intelligence.
According to claim 1 or 2,
The user authentication request processing step,
assigning different weights to the stroke attribute information according to an acquisition type of the handwriting input; and
Processing the user authentication request based on the weighting result; including,
A method for performing user authentication according to handwriting input based on artificial intelligence.
According to claim 1 or 2,
The user authentication request processing step,
assigning different weights to the stroke attribute information based on a predetermined security level of an operation corresponding to the user authentication request; and
Processing the user authentication request based on the weighting result; including,
A method for performing user authentication according to handwriting input based on artificial intelligence.
According to claim 1 or 2,
The user authentication request processing step,
checking a predetermined security level of an operation corresponding to the user authentication request;
selecting stroke attribute information to be considered for user authentication from among stroke attribute information of the handwriting input based on the predetermined security level; and
Processing the user authentication request based on the selected stroke attribute information;
A method for performing user authentication according to handwriting input based on artificial intelligence.
processor; and
A memory for storing instructions executable by the processor; includes,
the processor,
By executing the above commands,
Receive a user authentication request for identifying a user, obtain a user's handwriting input, obtain stroke attribute information of the handwriting input, and obtain text corresponding to the handwriting input using an artificial intelligence-based optical character reading model and processing the user authentication request by using the obtained text and the stroke attribute information together.
A device for performing user authentication according to handwriting input based on artificial intelligence.
processor; and
A memory for storing instructions executable by the processor; includes,
the processor,
By executing the above commands,
Receiving a user authentication request for identifying a user, obtaining a user's handwriting input, acquiring stroke attribute information of the handwriting input, and responding to the stroke attribute information of the handwriting input using an artificial intelligence-based handwriting input analysis model Obtaining text that is, and processing the user authentication request by using the obtained text and stroke attribute information of the handwriting input together.
A device for performing user authentication according to handwriting input based on artificial intelligence.