KR20230094062A - Face recognition system and method for controlling the same - Google Patents

Abstract

The present invention relates to a system for recognizing a user's face and a method for controlling the same in order to perform user authentication. The present invention includes receiving a face image and determining authenticity of the image based on a temperature value corresponding to each of a plurality of pixels included in the face image, and determining authenticity of the image. The step of performing the comparison may include a comparison result of temperature values of each of pixels disposed along a preset direction based on a point of the face image and a temperature value of pixels constituting one region of the face image and other values of the face image. A face recognition method characterized in that it is performed based on at least one of comparison results between temperature values of pixels constituting one area may be provided.

Description

Face recognition system and control method for face recognition system {FACE RECOGNITION SYSTEM AND METHOD FOR CONTROLLING THE SAME}

The present invention relates to a system for recognizing a user's face and a method for controlling the same in order to perform user authentication.

As technology develops, utilization of terminals is increasing, and in particular, terminals are used as a means of authenticating users in various situations.

For example, user identification at an access control device installed in an access control space (e.g. access control gate or speed gate), attendee verification at a conference, patient visit verification at a hospital, user verification for payment at a cafe As such, there is a need to authenticate genuine users in various situations. In this way, the use of the terminal in a situation requiring user authentication has the effect of reducing unnecessary labor and, furthermore, reducing the time consumed for user authentication.

There are various user authentication methods using terminals. For example, identification information (eg, phone number, e-mail address, etc.) or biometric information (eg, fingerprint information, etc.) There may be a method in which authentication is performed by inputting into a terminal.

Recently, active research is being conducted on a method for quickly and accurately recognizing a user's face in order to perform user authentication without the user directly applying information to the terminal.

An object of the present invention is to provide a face recognition system that more quickly and accurately recognizes the face of a user attempting user authentication, and a method for controlling the face recognition system.

Furthermore, the present invention is to provide a face recognition system and a method for controlling the face recognition system that can quickly determine the authenticity of a user's face attempting user authentication.

Furthermore, the present invention is to provide a face recognition system and a control method capable of quickly determining the authenticity of a user's face attempting user authentication only with computational resources built into a face recognition device without separate computational resources. .

In order to solve the above problems, the present invention includes receiving a face image and determining whether the image is authentic based on a temperature value corresponding to each of a plurality of pixels included in the face image, Determining the authenticity of the image may include a comparison result of temperature values of each of pixels disposed along a preset direction with respect to a point of the face image and a temperature value of pixels constituting one region of the face image. The face recognition method may be performed based on at least one of a comparison result between a temperature value and a temperature value of a pixel constituting another area of the face image.

In addition, the present invention includes a camera unit that acquires a face image and a controller that determines whether or not the image is authentic based on a temperature value corresponding to each of a plurality of pixels included in the face image, wherein the controller determines the authenticity of the face image. The comparison result of the temperature values of each of the pixels arranged along a predetermined direction based on a point of the image and the temperature values of the pixels constituting one area of the face image and the pixels constituting another area of the face image It is possible to provide a face recognition system characterized in that it determines the authenticity of the image based on at least one of the comparison results between the temperature values of the.

In addition, the present invention includes a server including a communication unit that receives a face image and a controller that determines whether or not the image is authentic based on a temperature value corresponding to each of a plurality of pixels included in the face image, wherein the controller is the comparison result of the temperature values of each of the pixels arranged along a predetermined direction with respect to one point of the face image, and the temperature values of pixels constituting one region of the face image and another region of the face image. It is possible to provide a face recognition system characterized in that the authenticity of the image is determined based on at least one of the comparison results between the temperature values of the constituent pixels.

In addition, the present invention is a program executed by one or more processes in an electronic device and stored in a computer-readable recording medium, comprising steps of receiving a face image and temperatures corresponding to each of a plurality of pixels included in the face image. It includes instructions for performing a step of determining authenticity of the image based on a value, wherein the step of determining whether the image is authentic is arranged along a predetermined direction based on a point of the face image. performed based on at least one of a comparison result of the temperature values of each of the selected pixels and a comparison result between the temperature values of pixels constituting one area of the face image and the temperature values of pixels constituting another area of the face image. It is possible to provide a program stored in a computer-readable recording medium, characterized in that.

As described above, since the face recognition system and its control method according to the present invention determine the authenticity of a face image only with the pixel values included in the face image, fast authenticity determination can be made without separate computational resources and artificial intelligence models. it could be possible

In addition, the face recognition system and method for controlling the face recognition system according to the present invention refer to both the comparison result between adjacent pixels included in the face image and the comparison result between areas where the face is located and areas where the face is not located to determine authenticity of the face image. Therefore, it is possible to accurately distinguish between a spoofing image and a normal image obtained by photographing a face wearing an accessory that partially covers the face.

1 is a conceptual diagram for explaining a face recognition system according to the present invention.
2 is a conceptual diagram illustrating a face image for face recognition according to the present invention.
3 is a conceptual diagram showing a device to which a face recognition system according to the present invention is applied.
4 is a flowchart for explaining a control method of a face recognition system according to the present invention.
5 is a conceptual diagram illustrating a method of determining whether a face image satisfies a first condition.
6 is a flowchart illustrating a method of determining whether a face image satisfies a first condition.
7 is a conceptual diagram illustrating a method of determining whether a face image satisfies a second condition.
8 is a flowchart illustrating a method of determining whether a face image satisfies a second condition.

Hereinafter, the embodiments disclosed in this specification will be described in detail with reference to the accompanying drawings, but the same reference numerals will be assigned to the same or similar components regardless of reference numerals, and overlapping descriptions thereof will be omitted. The suffixes "module" and "unit" for components used in the following description are given or used together in consideration of ease of writing the specification, and do not have meanings or roles that are distinct from each other by themselves. In addition, in describing the embodiments disclosed in this specification, if it is determined that a detailed description of a related known technology may obscure the gist of the embodiment disclosed in this specification, the detailed description thereof will be omitted. In addition, the accompanying drawings are only for easy understanding of the embodiments disclosed in this specification, the technical idea disclosed in this specification is not limited by the accompanying drawings, and all changes included in the spirit and technical scope of the present invention , it should be understood to include equivalents or substitutes.

Terms including ordinal numbers, such as first and second, may be used to describe various components, but the components are not limited by the terms. These terms are only used for the purpose of distinguishing one component from another.

It is understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle. It should be. On the other hand, when an element is referred to as “directly connected” or “directly connected” to another element, it should be understood that no other element exists in the middle.

Singular expressions include plural expressions unless the context clearly dictates otherwise.

In this application, terms such as "comprise" or "have" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

The present invention is to provide a face recognition system and method for accurately and quickly determining the authenticity of a user's face image without using separate computational resources and artificial intelligence models. The invention will be described in detail. 1 is a conceptual diagram for explaining a face recognition system according to the present invention, FIG. 2 is a conceptual diagram for explaining a face image for face recognition according to the present invention, and FIG. 3 is a device to which a face recognition system according to the present invention is applied. It is a conceptual diagram that represents

As shown in FIG. 1 , user authentication in the present invention may be performed through interworking with the face recognition system 100 and the server 200 . In the present invention, the server 200 may be located in the same place as the face recognition system 100 or in a separate place, and may be configured to process user authentication for a face recognized by the face recognition system 100.

Meanwhile, the server 200 may be implemented as one component of the face recognition system 100, and in this case, the face recognition system 100 may perform functions including the role of the server 200.

In addition, as shown in FIG. 1 , the face recognition system 100 and the server 200 may be configured separately, and in the present invention, for convenience of description, the face recognition system 100 and the server 200 separate and explain.

Meanwhile, regardless of whether the face recognition system 100 and the server 200 are physically separated, the functions and controls performed by the face recognition system 100 and the server 200 described in the present invention are "face recognition system" Of course, it can be described as being performed in .

In the present invention, a user's face may be recognized, user authentication may be performed, and an event or function related to user authentication (hereinafter, a target function) may be performed according to a user authentication result.

Here, the type of “target function” may be very diverse, and may mean a function that requires user authentication, such as a specific space access control function, payment function, attendance confirmation function, visit confirmation function in a hospital, and the like. For example, a specific space access control function is performed by controlling the access control gates (10 and 20, or access control device or speed gate) shown in FIG. 1 to operate in either an open state or a closed state. can be performed

As such, the face recognition system 100 in the present invention may have various hardware configurations depending on the functions to be achieved through face recognition and user authentication (eg, access control, attendance check, payment acceptance, etc.), In the present invention, as shown in FIG. 1 as a representative example, user authentication in the access control gate 10 will be described. However, it goes without saying that the face recognition and user authentication methods described below can be applied to face recognition systems having various functions.

As shown in FIG. 1, the face recognition system 100 (hereinafter referred to as “system 100”) according to the present invention includes a sensor unit 110, a storage unit 120, a communication unit 130, It may be configured to include at least one of the display unit 140, the sound output unit 150, and the control unit 160. At this time, the system 100 according to the present invention is not limited to the above-described components, and may further include components performing the same or similar roles as the functions described herein.

Furthermore, the face recognition system 100 may be interlocked with the access control gate 10 or may include the access control gate 10 as a component. Furthermore, it goes without saying that the face recognition system 100 can be a component of the access control gate 10 . The sensor unit 110 senses (or senses) a user (or object) using a configuration of a sensing module (eg, a camera 111, a thermal imaging camera, a proximity sensor, an infrared sensor, etc.) provided in the system 100. or filming).

In the present invention, “sensing” can be interpreted in various ways depending on the sensing module. For example, when using the camera 111, sensing may be understood as capturing an image. As another example, when using a proximity sensor, it may be understood as sensing the presence of a user (or object). Hereinafter, “sensing” may be used interchangeably with “photographing (or obtaining) an image”.

The sensor unit 110 may capture an image (eg, at least one of an RGB image frame and a thermal image frame). At this time, the sensor unit 110 may capture an image in the form of at least one of a video and an image.

The sensor unit 110 may capture an image based on a preset time period or a user's request.

The storage unit 120 may be configured to store data related to face recognition and/or user authentication according to the present invention using a memory provided (or inserted) in the system 100 or a cloud server. For example, the storage unit 120 may store information sensed by the sensor unit 110 or acquired images or information obtained from them.

Furthermore, the communication unit 130 uses a configuration of communication modules (eg, a mobile communication module, a short-distance communication module, a wireless Internet module, a location information module, a broadcast reception module, etc.) provided in the system 100 to use a server ( 200), an external server (not shown), or a cloud server (not shown).

Furthermore, the display unit 140 is configured to output visual information, and under the control of the controller 160, it may be configured to output information related to at least one of face recognition and user authentication.

For example, the display unit 140 outputs information related to user authentication success (or failure) based on the user authentication result information for the user to be authenticated received from the server 200, or the sensor unit 110 may be controlled to output guide information for guiding the user's location so that the user can be photographed.

For example, based on the authentication result information on the user to be authenticated, the display unit 140 may display an authentication result or guide information related to performing a target function.

Furthermore, the audio output unit 150 may audibly output information related to the present invention.

For example, under the control of the control unit 160, based on authentication result information on the user to be authenticated, the audio output unit 150 may output an authentication result or guide information related to performing a target function in voice.

Furthermore, the controller 160 may utilize at least one of a CPU and a cloud server provided in the system 100 to perform control of each configuration and overall control related to the present invention.

The controller 160 may have control authority for at least one of the components of the system 100 in order to carry out the present invention.

The controller 160 may recognize a user's face and perform a process for user authentication based on information sensed by the sensor unit 110 or received through the communication unit 130 .

The controller 160 may include a face recognition unit 161, and the face recognition unit 161 may recognize a face in an image captured by the sensor unit 110 based on an algorithm having a face recognition function. . Furthermore, the face recognition unit 161 may perform a series of controls related to the recognized face, which will be discussed in more detail later. In the present invention, for convenience of description, the configuration of the face recognition unit 161 and the controller 160 is not distinguished, and the controller 160 is unified and described.

Meanwhile, the system 100 and the server 200 according to the present invention operate in conjunction with each other to perform user authentication for the target function and control the execution of the target function based on the user authentication result for the target function. can

The server 200 may serve to perform user authentication based on a user authentication request from the system 100 and transmit a user authentication result to the system 100 .

Meanwhile, in the present invention, when the function performed as a result of user authentication is the operation control of the access control gate 10, the door of the access control gate 10 is moved based on the user authentication result for the user to be authenticated. It can be controlled to open or not to open. The access control gate 10 may be controlled based on a control command transmitted from at least one of the server 200 and the system 100 .

Meanwhile, as shown in FIG. 1 , the server 200 according to the present invention may be configured to include at least one of a communication unit 210, a storage unit 220, and a control unit 230. At this time, the server 200 according to the present invention is not limited to the above-described components, and may further include components performing the same or similar functions as those described herein.

The communication unit 210 may transmit/receive data with at least one of the system 100, the access control gate 20, an external server (not shown), or a cloud server (not shown).

For example, the communication unit 210 receives image information including the face of the user to be authenticated or face information on the user to be authenticated from the system 100, and the user authentication result for the user to be authenticated is displayed in the system 100. ) can be transmitted.

The storage unit 220 may be configured to store various information related to user authentication.

The storage unit 220 may include at least one of a user DB 221 containing reference facial feature information related to a pre-registered face and a terminal DB 222 containing information related to the system 100 .

Here, the reference facial feature information may include at least one of a pre-registered face image and pre-registered feature information (eg, feature vector information).

User information for each person may be stored in the user DB 221 in units of individual persons (or persons). That is, in the user DB, user information corresponding to each different person may be matched and stored.

Here, the user information includes the name, date of birth, address, phone number, e-mail address, account information, employee number, ID, biometric information (fingerprint information, iris information, etc.), reference facial feature information, It may include at least one of schedule (schedule) related information.

The reference face image included in the reference face feature information is a face image of a specific person, and may be one or more. In the present invention, the user DB may also be referred to as "user related information", "user personal information" or "user identification information".

Meanwhile, information about the system 100 performing the user authentication process may be stored in the terminal DB 222 .

Here, the information on the terminal may include one of identification information (serial number) of each system 100 or access control gate 20, location, and information related to a target function linked with user authentication.

Meanwhile, the controller 230 may perform user authentication on the user to be authenticated using reference facial feature information stored in the user DB 221 .

At this time, user authentication may be performed through the user authentication unit 231 of the control unit 230. That is, it can be understood that the user authentication unit 231 performs a role of the control unit 230 . In the present invention, the control unit 230 and the user authentication unit 231 are used interchangeably without being separately distinguished.

Hereinafter, based on the configuration of the system 100 and the server 200 described above, a basic process in which authenticity of a face image is determined in the present invention will be described. The present invention provides a face recognition system capable of distinguishing between a normal face image and a spoofing image (hereinafter referred to as 'determining authenticity of a face image').

Conventionally, normal images and spoof images are collected, and the collected images are used for training of deep learning models. A deep learning model trained with normal and spoofed images can distinguish between normal and spoofed images. However, this method is difficult to utilize when determining authenticity of a face image by itself within the face recognition system 100 having limited computational resources, and even if utilized, it is difficult to expect fast face recognition due to a large amount of computation.

In particular, referring to FIG. 2 , as the population wearing accessories on the face increases, an image captured during face authentication is an image 211 in which the accessory is not worn on the face, or an accessory (eg, a mask) covering the face. ) may be any one of the images 212 worn. When face recognition is performed through the image 212 wearing an accessory covering the face, it may be difficult to determine authenticity of the face image.

Also, as shown in FIG. 3 , when the face recognition system 100 is used for the access control gate 10 where users are concentrated, delay in face recognition time may cause inconvenience to many users 321 to 333 .

The present invention can provide a face recognition method capable of quickly and accurately detecting spoofing using computational resources built into the face recognition system 100 . Hereinafter, the face recognition method according to the present invention will be described in detail.

4 is a flowchart illustrating a control method of a face recognition system according to the present invention, FIG. 5 is a conceptual diagram illustrating a method of determining whether a face image satisfies a first condition, and FIG. 7 is a conceptual diagram illustrating a method of determining whether a face image satisfies a second condition, and FIG. 8 is a flowchart showing a method of determining whether a face image satisfies a second condition.

First, a step of receiving a face image may proceed (S110).

As described above, the face recognition system 100 according to the present invention may capture a user's face through a camera. A controller included in the face recognition system 100 or a controller 230 included in the server 200 may receive a face image in order to recognize a face and determine whether the face image is authentic. In the present specification, it is described that face recognition and determination of authenticity of a face image are performed by the controller 160 built into the face recognition system 100, but this will be performed by the controller 230 built into the server 200. It is obvious to those skilled in the art.

Meanwhile, the face image may be a thermal image captured by a thermal imaging camera as well as an RGB image captured by an RGB camera.

In an embodiment, the controller 160 may receive a face image from a camera, detect a face region from the face image using a pre-stored algorithm, and generate a separate image using only the detected region. That is, in the present invention, only a part of the entire image captured by the camera can be used to determine authenticity of a face image. In this specification, the face image may be an original image captured by a camera or a separate image generated by detecting a face region from the original image and using the detected region.

Next, a step of determining whether the face image satisfies the first condition may be performed (S120).

Spoofing detection using a conventional thermal image has been performed based on an average value or a variance value of temperature values of pixels constituting a face image. In this case, when the user wears the accessory, the input value by the spoofing attack and the value calculated through the normal face image become similar, making it difficult to distinguish the spoof image from the normal image. The spoofing detection method using the first condition can detect spoofing with high accuracy even when a user wears an accessory based on pattern analysis between adjacent pixels.

Here, the first condition may be a condition related to a comparison result of temperature values of each of pixels disposed along a predetermined direction with respect to a point of the face image.

The first condition may be related to a pixel value of each of a plurality of pixels included in the face image. Specifically, the face image may include a plurality of pixels. For each of the plurality of pixels, coordinate information of the pixel and pixel value of the pixel may be matched. The controller may determine whether the face image satisfies the first condition based on coordinate information and pixel values corresponding to each of the plurality of pixels.

In one embodiment, the coordinate information included in each of the plurality of pixels may include an x-axis coordinate corresponding to the x-axis and a y-axis coordinate corresponding to the y-axis.

In an embodiment, the pixel values included in each of the plurality of pixels may include pixel values corresponding to each of R, G, and B.

Meanwhile, a temperature value may be matched to each of the pixels constituting the face image. Here, a temperature value matched to each of the pixels may correspond to a pixel value of each of the pixels.

For example, when the face image is captured by a thermal imaging camera, each pixel value included in the face image corresponds to a temperature. For example, when a thermal imaging camera captures an object having a high temperature, pixels constituting an image corresponding to the object include pixel values close to red. In contrast, when a thermal imaging camera captures an object having a low temperature, pixels constituting an image corresponding to the object may include pixel values close to blue.

The controller 160 may calculate a temperature value corresponding to a specific pixel by using a pixel value of the specific pixel and a pre-stored pixel value-temperature value table.

In this case, each pixel value may correspond to a specific temperature value. The controller 160 may calculate a temperature value corresponding to a pixel value of a reference pixel and a temperature value corresponding to a pixel value of another pixel, and calculate a difference between the two temperature values. In this specification, when the controller 160 “compares two pixels” or “compares information related to each of the two pixels”, it may mean comparing temperature values matched to the pixels.

In an embodiment, in order to determine whether the face image satisfies the first condition, the controller 160 sets a reference pixel and selects at least one comparison target pixel from among the plurality of pixels based on the position of the reference pixel. Afterwards, the temperature values of the reference pixel and the comparison target pixel may be compared.

Here, the reference pixel and the comparison target pixel may be any one of pixels located along a predetermined direction based on a point of the face image.

In one embodiment, the one point may be set to at least one of a center, a corner, or an arbitrary point of the face image, but is not limited thereto. In one embodiment, the controller 160 may set one point of the face image. One of a plurality of pixels located along a predetermined direction based on is set as a reference pixel, and among a plurality of pixels located along the predetermined direction, a pixel located at a predetermined distance from the reference pixel is set as a comparison target pixel. can Thereafter, the controller 160 may determine whether a difference between the temperature values of the reference pixel and the comparison target pixel exceeds a first reference value.

Meanwhile, the control unit 160 may reset the reference pixel so that the reference pixel is farther away from a point of the face image along the predetermined direction, and repeat the above-described temperature comparison operation. For example, the controller 160 sets a pixel located in the center of the face image as a reference pixel (hereinafter, an initially set reference pixel is referred to as a start pixel), and the reference pixel is far from the start pixel in the upper left diagonal direction. To achieve this, the reference pixel may be changed.

In the present specification, as described above, in the temperature comparison operation in a specific direction, the control unit 160 resets the reference pixel so that the reference pixel is farther away from a point of the face image along the specific direction, It may mean calculating a difference between a pixel and a temperature value of a pixel located at a preset distance from the reference pixel.

Meanwhile, the comparison target pixel may be a pixel to be set as a reference pixel next to the current reference pixel or a pixel set as a reference pixel before the current reference pixel. In this specification, a case in which the comparison target pixel is a pixel set as a reference pixel prior to the current reference pixel is described, but is not limited thereto.

Meanwhile, a plurality of preset directions may be set. The controller 160 may set a pixel located at one point of the face image as a reference pixel, and compare temperatures between the reference pixel and pixels adjacent to the reference pixel while changing the reference pixel along one of a plurality of directions. . After the temperature comparison between pixels in one direction is completed, the control unit 160 resets the reference pixel to the starting pixel, and then performs temperature comparison while changing the reference pixel along the one direction and the other direction. can do. That is, the present invention compares the temperature values of pixels disposed along one direction centered on a point of the face image, and the temperature comparison operation may be performed in a plurality of directions centered on a point of the face image. there is.

The controller 160 may perform a temperature comparison operation for each of the plurality of preset directions. As described above, the controller 160 may perform a temperature comparison operation in another one of the plurality of preset directions after completing the temperature comparison operation in any one of the plurality of preset directions. Unlike this, the controller 160 may perform a temperature comparison operation in another one of the plurality of preset directions while performing the temperature comparison operation in any one of the plurality of preset directions. . That is, the temperature comparison operation for each of a plurality of predetermined directions may be performed sequentially or simultaneously.

For example, the controller 160 sets a unit increment amount in the x direction (hereinafter referred to as x-axis change amount) and a unit increment amount in the y-direction (hereinafter referred to as y-axis change amount), and sets the x included in the coordinate information of the reference pixel. Set a pixel whose x-coordinate is the value obtained by adding the x-axis change amount to the coordinates and whose y-coordinate is the value obtained by adding the y-axis change amount to the y-coordinate included in the coordinate information of the reference pixel as a new reference pixel, Temperature comparison between a pixel that was a reference pixel before setting the reference pixel (hereinafter referred to as a previous pixel) and the new reference pixel may be performed. That is, the controller 160 may reset a reference pixel to a pixel located in one direction centering on one point of the face image, and then perform temperature comparison between the reset reference pixel and the previous pixel.

Furthermore, the reference pixel may move in each of a plurality of directions based on a point of the face image. Specifically, the reference pixel is (x standard + x change, y standard + y change), (x standard - x change, y standard + y change), (x standard + x change, y standard - y change), ( It may be reset according to at least one criterion of x criterion-x change amount, y criterion-y change amount). Here, the x reference is the x coordinate of the current reference pixel, and the y reference is the y coordinate of the current reference pixel.

For example, the controller 160 may reset the reference pixel according to a criterion (x criterion + x change amount, y criterion + y change amount). In this case, the temperature comparison operation may be performed in a diagonal direction in the upper right corner based on a point of the face image.

In one embodiment, the control unit 160 resets the reference pixel according to a criterion (x criterion + x change amount, y criterion + y change amount), and after completing the temperature comparison operation, resets the reference pixel to the starting pixel, (x standard - x change amount, y standard + y change amount) Depending on the standard, the reference pixel can be reset and temperature comparison can be performed. In this case, the temperature comparison operation is performed in a diagonal direction in the upper right corner based on a point of the face image, and then in a diagonal direction in the upper left corner based on the point in the face image.

For example, referring to FIG. 5, while moving reference pixels in four diagonal directions (directions where pixels 1, 3, 7, and 9 are located) based on the starting pixel (pixel 5) set in area A, The temperature value of the reference pixel and the previous pixel can be compared. On the other hand, when a mask is recognized in the face image, the reference pixel is moved in two diagonal directions (the direction where pixels 1 and 3 are located) based on the starting pixel (pixel 5) set in area A, and the reference pixel and The temperature value of the previous pixel can be compared.

Here, the x-axis change amount and the y-axis change amount may be the same value or different values. In addition, the amount of change in the x-axis and the amount of change in the y-axis may be set based on at least one of the size of the face image, the aspect ratio of the face image, and the resolution of the face image. Accordingly, the amount of change in the x-axis and the amount of change in the y-axis may vary according to the size, ratio, and resolution of the face image.

The movement direction of the reference pixel may vary according to the ratio of the x-axis change amount and the y-axis change amount.

In one embodiment, each of the x-axis change amount and the y-axis change amount may not be 0. Accordingly, the direction in which the reference pixel moves may cross the longitudinal direction of the face image or a direction perpendicular to the longitudinal direction. More specifically, a virtual line connecting the reference pixel and the comparison target pixel is not parallel to each of the virtual x-axis and y-axis corresponding to the image information. In other words, the reference pixel may perform temperature comparison while proceeding in a diagonal direction centered on a point of the face image.

In an embodiment, the controller 160 may no longer change the temperature when a temperature comparison operation in a preset number of directions centered on a point of the face image is completed or when a reference pixel is changed according to a preset reference pixel change rule. The comparison can be performed until there are no pixels left. Accordingly, a difference value between the temperature value of the reference pixel and the previous pixel is calculated, and the difference value may be plural.

The controller 160 may determine that the face image satisfies the first condition when a maximum value among the plurality of difference values is greater than a preset first reference value.

For example, after setting a point of the face image as a reference pixel (start pixel), the controller 160 compares the temperature of the reference pixel and the previous pixel while changing the reference pixel along one direction centering on the start pixel. After that, if one or more reference pixels having a temperature difference value exceeding the first reference value exist, it may be determined that the face image satisfies the first condition.

Furthermore, after setting a point of the face image as a reference pixel (start pixel), the controller 160 compares the temperature of the reference pixel and the previous pixel while changing the reference pixel along one direction centering on the start pixel. When the temperature comparison operation in one direction is completed, the control unit 160 resets the reference pixel to the starting pixel and compares the temperature of the reference pixel and the previous pixel while changing the reference pixel along the one direction and the other direction. do. The controller 160 determines that the face image satisfies the first condition when there is one or more reference pixels having temperature difference values greater than the first reference value after temperature comparison in a plurality of preset directions is completed. can do.

Meanwhile, the number of preset directions may vary depending on whether a preset object exists in the face image. Specifically, the controller 160 may determine whether a preset object exists in the face image.

For example, the controller 160 may determine whether a preset object exists based on a temperature value matched to each of pixels included in a specific region of the face image. Specifically, the controller 160 may determine whether a preset object exists based on an average temperature and variance values calculated based on temperature values matched to each of the pixels included in the specific area.

Here, the preset object may be an object corresponding to an accessory worn on the face. For example, the preset object may be an object corresponding to a mask.

When a preset object is recognized in the face image, the controller 160 sets the number of directions in which temperature comparison is performed based on a point of the face image to be relatively small, and the preset object is not recognized in the face image. If not, the number of directions in which temperature comparison is performed based on one point of the face image may be set relatively large.

For example, when a predetermined object is recognized in the face image, the control unit 160 controls (x standard + x change amount, y standard + y change amount), (x standard - x change amount, y standard + y change amount) The temperature comparison can be performed by moving the reference pixel for each criterion. That is, when a preset object is recognized in the face image, the control unit 160 moves a reference pixel in a diagonal direction in the upper left corner based on a point of the face image, performs temperature comparison, and then uses the point as a reference point. By moving the reference pixel in the upper right diagonal direction, the temperature comparison can be performed.

In contrast, when a preset object is recognized in the face image, centered on the start pixel (x standard + x change amount, y standard + y change amount), (x standard - x change amount, y standard + y change amount), (x Temperature comparison may be performed while moving a reference pixel for each criterion of reference + x change, y reference - y change, (x reference - x change, y reference - y change). That is, the controller 160 may perform temperature comparison in each of the upper left diagonal, upper right diagonal, lower left diagonal, and lower right diagonal directions based on a point of the face image.

Referring to FIG. 6 , after receiving a face image (S210), the controller 160 sets a reference pixel (start pixel) and a reference direction (eg, diagonal direction in the upper left corner) (S220). Meanwhile, the controller 160 It is determined whether a mask is detected in the face image (S220).

When a mask is detected in the face image, the controller 160 performs temperature comparison (S242) while moving a reference pixel in a left diagonal direction with respect to the start pixel (S241). If there is no pixel that can be changed in the left diagonal direction (S243), the controller 160 determines whether the temperature comparison operation has been performed for two preset directions (eg, upper left diagonal direction and upper right diagonal direction). . The controller 160 ends the operation when the temperature comparison in the upper left diagonal direction and the upper right diagonal direction are respectively completed. That is, when the preset object is recognized in the face image, a temperature comparison operation may not be performed with respect to a lower direction of a point of the face image.

In contrast, when the mask is not detected in the face image, the controller 160 performs temperature comparison (S252) while moving a reference pixel in a left diagonal direction with respect to the start pixel (S251). If there is no pixel changeable in the left diagonal direction (S253), the controller 160 selects four preset directions (eg, upper left diagonal direction, upper right diagonal direction, lower left diagonal direction, lower right diagonal direction) It is determined whether the temperature comparison operation has been performed for . The controller 160 ends the operation when the temperature comparison in the upper left diagonal direction, the upper right diagonal direction, the lower left diagonal direction, and the lower right diagonal direction are respectively completed.

As described above, in FIG. 6, N may be an integer smaller than M. That is, when a preset object is not recognized from the face image, the controller 160 may perform a temperature comparison operation in more directions than when a preset object is recognized from the face image.

As described above, the present invention can detect spoofing simply, quickly, and with high accuracy by performing a temperature comparison operation in a specific direction based on a point of a face image.

A point of the face image and a direction in which comparison is performed may be set differently according to circumstances. In particular, spoofing detection errors may occur due to body parts (eg, hair) having different temperatures from the user's body temperature and accessories worn by the user (eg, hats, earrings, masks, etc.). For example, if there are accessories such as hats, earrings, and masks in the up, down, left, and right directions based on a point (for example, the center of the face) of a face image, it is effective to select a diagonal direction based on the center of the face. can be As described above, the present invention can detect spoofing by utilizing the first condition.

On the other hand, the spoofing detection method using the above-described first condition analyzes the inter-pixel pattern, so the accuracy is very high for the frontal face image, but there may be a weakness in the case where the face rotates. The present invention can provide a spoofing detection method using a second condition to complement the spoofing detection method using the first condition.

Next, a step of determining whether the face image satisfies the second condition may be performed (S130).

Here, the second condition may be a condition related to pixels included in the face image. Specifically, the second condition may be related to a comparison result between a temperature value of a pixel constituting one region of the face image and a temperature value of a pixel constituting another region of the face image.

The controller 160 may determine whether the face image satisfies the second condition based on a preset algorithm.

The controller 160 calculates a temperature value matched to each of a plurality of pixels constituting the face image, compares a specific temperature value among the plurality of temperature values, and determines whether the face image satisfies the second condition according to the comparison result. can judge

The controller 160 may divide the face image into a plurality of regions and calculate a representative temperature value representing each of the plurality of regions.

The representative temperature value may be a maximum value or a minimum value among temperature values matched to each of a plurality of pixels included in a specific area.

Meanwhile, the controller 160 may compare representative temperature values of each of a plurality of regions and determine whether the face image satisfies the second condition based on the comparison result.

In an embodiment, the controller 160 may determine that the face image satisfies the second condition when a difference between a maximum value and a minimum value among the representative temperature values of each of the plurality of regions exceeds a second reference value. .

Meanwhile, in setting the plurality of regions, the controller 160 includes one of the plurality of regions including the left side of the face image, another region including the right side of the face image, and another region including the face image. It can be set to include the center of the image. Accordingly, the number of the plurality of areas may be at least three.

For example, when there are three of the plurality of regions, the controller 160 controls the interval between the representative temperature value of the central region and the smaller of the representative temperature values of the left region and the right region and the representative temperature value of the central region. When the difference value exceeds the second reference value, it may be determined that the face image satisfies the second condition.

For another example, the controller 160 may determine that the representative temperature value of the left region is the largest, the representative temperature value of the right region is the smallest, and the difference between the representative temperature value of the left region and the representative temperature value of the right region is the second. When the reference value is exceeded, it may be determined that the face image satisfies the second condition.

For another example, the controller 160 may determine that the representative temperature value of the right region is the largest, the representative temperature value of the left region is the smallest, and the difference between the representative temperature value of the right region and the representative temperature value of the left region is the second. When the reference value is exceeded, it may be determined that the face image satisfies the second condition.

Meanwhile, each of the plurality of regions is set to extend from the upper side to the lower side in the face direction, and the length of the plurality of regions extends depending on whether the preset object is recognized in the face image and whether the preset object is It may be set based on at least one of the positions of the recognized area.

In an embodiment, the length of the region may vary depending on whether a predetermined object is recognized in the face image. Specifically, when a preset object is recognized in the face image, the controller 160 may set the plurality of regions so that the region where the recognized object is located is excluded from the plurality of regions.

For example, referring to FIG. 7 , when a predetermined object is recognized on the lower side of the face image 710, the plurality of regions 721, 722, and 723 may be set only on the upper side of the face image.

For another example, the controller 160 may set an area only on the upper side of the face image regardless of whether a preset object is recognized in the face image. Lengths of the plurality of regions may be set to a predetermined ratio (eg, half) of the total length of the face image.

Referring to FIG. 8 , after receiving a face image (S310), the controller 160 divides the face image into a plurality of regions (S320). Thereafter, the controller 160 calculates the representative temperature values for each of the plurality of regions (S330), and determines whether a difference between the maximum value and the minimum value among the plurality of representative temperature values for each region exceeds a second reference value (S340). . As a result of the determination, when the difference between the maximum value and the minimum value among the representative temperature values for each region exceeds the second reference value, it may be determined that the face image satisfies the second condition.

Finally, based on the determination result corresponding to each of the first and second conditions, a step of determining authenticity of the face image is performed (S140).

The spoofing detection method using the second condition described above may be used independently for spoofing detection according to the present invention or may be used together with the spoofing detection method using the first condition. That is, when the received face image satisfies any one of the first and second conditions, the controller 160 determines that the face image is normal, or when the received face image satisfies both the first and second conditions. It can be determined that the face image is normal only for .

As described above, since the face recognition system and its control method according to the present invention determine authenticity of a face image only with pixel values included in the face image, fast authenticity determination can be made without separate computational resources and artificial intelligence models. it could be possible

In addition, the face recognition system and method for controlling the same according to the present invention refer to both the comparison result between adjacent pixels included in the face image and the comparison result between the area where the face is located and the area where the face is not located to determine the authenticity of the face image. Therefore, it is possible to accurately distinguish between a spoofing image and a normal image obtained by photographing a face wearing an accessory that partially covers the face.

Since the spoofing detection method using the second condition utilizes the temperature difference for each region of the face image, it can have higher average accuracy than the spoofing detection method using the first condition. However, in the case of an image in which the front of the face is accurately photographed, the spoofing detection method using the first condition has higher accuracy than the spoofing detection method using the second condition. When the first and second conditions are used together, it is possible to quickly and accurately determine the authenticity of a face image without separate computational resources and an artificial intelligence model.

On the other hand, the computer-readable medium includes all types of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable media include Hard Disk Drive (HDD), Solid State Disk (SSD), Silicon Disk Drive (SDD), ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc. there is

Furthermore, the computer-readable medium may be a server or cloud storage that includes a storage and can be accessed by a terminal through communication. In this case, the computer may download the program according to the present invention from a server or cloud storage through wired or wireless communication.

Furthermore, in the present invention, the above-described computer is a processor, that is, a terminal equipped with a CPU (Central Processing Unit), and there is no particular limitation on its type.

On the other hand, the above detailed description should not be construed as limiting in all respects and should be considered as illustrative. The scope of the present invention should be determined by reasonable interpretation of the appended claims, and all changes within the equivalent scope of the present invention are included in the scope of the present invention.

Claims (12)

Receiving a face image; and
Determining whether the face image is authentic based on a temperature value corresponding to each of a plurality of pixels included in the face image;
The step of determining the authenticity of the video,
Comparing the temperature values of each of the pixels arranged along a predetermined direction with respect to one point of the face image and the temperature values of pixels constituting one area of the face image and another area of the face image A face recognition method characterized in that it is performed based on at least one of the comparison results between the temperature values of the pixels. According to claim 1,
The step of determining the authenticity of the video,
Comparing temperature values of each of a plurality of pixels arranged along the predetermined direction based on the one point,
Comparing the temperature values,
setting one of a plurality of pixels disposed along the predetermined direction based on the one point as a reference pixel; and
Calculating a difference between the temperature value of the reference pixel and the temperature value of a pixel located at a predetermined distance from the reference pixel among a plurality of pixels disposed along the predetermined direction based on the one point Characterized face recognition method. According to claim 2,
Comparing the temperature values is repeatedly performed by resetting the reference pixel,
The reference pixel is reset to be away from the one point along the predetermined direction,
The face recognition method of claim 1 , wherein the predetermined direction is set to intersect with a longitudinal direction of the face image or a direction perpendicular to the longitudinal direction. According to claim 3,
The preset direction is plural,
Comparing the temperature values may be performed for each of the plurality of preset directions. According to claim 4,
Further comprising recognizing a preset object in the face image,
When the predetermined object is recognized in the face image, the comparing of the temperature values is not performed in a downward direction of a point of the face image. According to claim 1,
The step of determining the authenticity of the video,
dividing the face image into a plurality of regions;
setting a representative temperature value for each of the divided images; and
and determining whether the face image is authentic based on a comparison result between representative temperature values corresponding to each of the plurality of regions. According to claim 6,
A representative temperature value for any one of the divided regions is,
The face recognition method, characterized in that the maximum value or minimum value among the temperature values matched to each of the plurality of pixels constituting the one area. According to claim 7,
and determining that the face image is a normal image when a difference between a maximum value and a minimum value among the representative temperature values of each of the plurality of regions exceeds a second reference value. According to claim 8,
Each of the plurality of areas,
It is set to extend from the upper side to the lower side in the face direction, and the length of the plurality of regions depends on at least one of whether the preset object is recognized in the face image and the location of the region where the preset object is recognized. Face recognition method, characterized in that set based on. a camera unit that acquires a face image; and
A control unit for determining authenticity of the face image based on temperature values corresponding to each of a plurality of pixels included in the face image;
The control unit,
Comparing the temperature values of each of the pixels arranged along a predetermined direction with respect to one point of the face image and the temperature values of pixels constituting one area of the face image and another area of the face image A face recognition system, characterized in that for determining the authenticity of the image based on at least one of the comparison results between the temperature values of the pixels. a communication unit receiving a face image; and
A server including a controller that determines authenticity of the image based on a temperature value corresponding to each of a plurality of pixels included in the face image;
The control unit,
The comparison result of the temperature values of each of the pixels arranged along a predetermined direction based on one point of the face image and the temperature values of pixels constituting one region of the face image and another region of the face image A face recognition system, characterized in that for determining the authenticity of the image based on at least one of the comparison results between the temperature values of the pixels. A program that is executed by one or more processes in an electronic device and stored in a computer-readable recording medium,
Receiving a face image; and
Including instructions for performing a step of determining authenticity of the image based on a temperature value corresponding to each of a plurality of pixels included in the face image,
The step of determining the authenticity of the video,
The comparison result of the temperature values of each of the pixels arranged along a predetermined direction based on one point of the face image and the temperature values of pixels constituting one region of the face image and another region of the face image A program stored in a computer-readable recording medium, characterized in that it is performed based on at least one of the comparison results between the temperature values of the pixels.

Images