Multifactor-based biometric authentication device and method

Abstract

The present invention relates to a multifactor-based biometric authentication device. A multifactor-based biometric authentication device, according to one embodiment of the present invention, may comprise: a data acquisition unit that acquires fingerprint data and blood flow data from a subject; a data preprocessing unit that preprocesses the fingerprint data and the blood flow data; a feature extraction unit that extracts fingerprint features and blood flow features from the fingerprint data and the blood flow data, respectively; a representative feature value generation unit that generates a representative feature value for the subject by combining and averaging the fingerprint features and the blood flow features; and a user authentication unit that performs user authentication of the subject on the basis of the representative feature value.

Description

Multifactor-based biometric authentication device and method

[0001] The present invention relates to a multi-factor based biometric authentication device and method.

[0002] Biometric technologies identify individuals based on their physiological or behavioral characteristics. Physiological characteristics can include fingerprints, irises, retinas, faces, handwriting, and DNA, while behavioral characteristics include voice, gait, and handwriting. The most widely used biometric technology is fingerprint recognition. However, fingerprint recognition methods have drawbacks: they are susceptible to forgery and duplication, are inaccurate on contaminated surfaces, and fixed biometric information cannot be altered.

[0003] Meanwhile, PPG (Photoplethysmography) data measures blood flow under the skin by shining light on a subject's body part, potentially including unique heart rate characteristics. PPG data offers the advantages of easy acquisition, cost-effectiveness, and high measurement reliability. Because it includes an individual's average heart rate characteristics, it can be highly secure when applied to biometrics.

[0004] Accordingly, it is expected that the shortcomings of existing fingerprint-based biometric recognition technology can be complemented by implementing a multi-factor authentication technique that strengthens security by combining multiple authentication methods by applying PPG data to existing fingerprint-based biometric recognition technology.

[0005] The technology underlying this application is disclosed in Korean Patent Publication No. 10-1990285.

[0006] The present invention aims to solve the problems of the prior art described above, and to provide a multifactor-based biometric authentication device and method that can improve the accuracy of an individual's biometric characteristics included in fingerprint data and blood flow data by constructing and utilizing a deep learning-based feature extraction model capable of extracting each feature from fingerprint data and blood flow data.

[0007] The present invention is intended to solve the problems of the prior art described above, and to provide a multifactor-based biometric authentication device and method that can prevent abnormal access and improve authentication accuracy and security compared to the prior art by performing user authentication based on multifactor-based feature values ​​that combine features extracted from fingerprint data and blood flow data, respectively.

[0008] However, the technical tasks to be achieved by the embodiments of the present invention are not limited to the technical tasks described above, and other technical tasks may exist.

[0009] As a technical means for achieving the above-described technical task, a multifactor-based biometric authentication device according to one embodiment of the present invention may include a data acquisition unit that acquires fingerprint data and blood flow data from a subject, a data preprocessing unit that performs preprocessing on the fingerprint data and the blood flow data, a feature extraction unit that extracts fingerprint features and blood flow features from the fingerprint data and the blood flow data, respectively, a representative feature value generation unit that generates a representative feature value for the subject by combining and averaging the fingerprint features and the blood flow features, and a user authentication unit that performs user authentication of the subject based on the representative feature value.

[0010] According to one embodiment of the present invention, the feature extraction unit may extract the fingerprint feature in vector form from the fingerprint data and extract the blood flow feature in vector form from the blood flow data using a feature extraction model.

[0011] According to one embodiment of the present invention, the representative feature value generation unit may generate a feature vector set using a multi-factor feature generated by combining the fingerprint feature and the blood flow feature output through the feature extraction model, and perform an average on the feature vector set to generate the representative feature value in the form of a single vector.

[0012] According to one embodiment of the present invention, the feature extraction model may be learned to have a hierarchical structure including a plurality of convolution blocks.

[0013] According to one embodiment of the present invention, the user authentication unit may perform the user authentication by deriving a distance between the representative feature value and at least one identification feature value among a plurality of identification feature values ​​previously stored in a database, comparing the derived distance with a preset threshold value, and determining whether the subject is a previously stored user.

[0014] According to one embodiment of the present invention, the user authentication unit sequentially derives the distance for each of the representative feature value and the plurality of identification feature values ​​in a preset order, but if the distance between the representative feature value and any one of the plurality of identification feature values ​​is greater than or equal to the preset threshold value, the distance derivation for the identification feature value corresponding to the order after the representative feature value and any one of the identification feature values ​​may be omitted.

[0015] According to one embodiment of the present invention, the data preprocessing unit may include a first data preprocessing unit that performs image preprocessing on the fingerprint data and a second data preprocessing unit that performs preprocessing on the blood flow data.

[0016] According to one embodiment of the present invention, the first data preprocessing unit may perform at least one image preprocessing process among background removal, image scaling, image size adjustment, and data augmentation on the fingerprint data.

[0017] According to one embodiment of the present invention, the second data preprocessing unit may perform at least one signal preprocessing process among resampling, filtering, segmentation, and baseline correction on the blood flow data.

[0018] As a technical means for achieving the above-described technical task, a multifactor-based biometric authentication method according to one embodiment of the present invention may include a step of obtaining fingerprint data and blood flow data from a subject, a step of performing preprocessing on the fingerprint data and the blood flow data, a step of extracting fingerprint features and blood flow features from the fingerprint data and the blood flow data, respectively, a step of generating a representative feature value for the subject by combining and averaging the fingerprint features and the blood flow features, and a step of performing user authentication of the subject based on the representative feature value.

[0019] According to one embodiment of the present invention, the step of extracting the fingerprint features and blood flow features may include extracting the fingerprint features in vector form from the fingerprint data and extracting the blood flow features in vector form from the blood flow data using a feature extraction model.

[0020] According to one embodiment of the present invention, the step of generating the representative feature value may include generating a feature vector set using a multifactor feature generated by combining the fingerprint feature and the blood flow feature output through the feature extraction model, and performing averaging on the feature vector set to generate the representative feature value in the form of a single vector.

[0021] According to one embodiment of the present invention, the feature extraction model may be learned to have a hierarchical structure including a plurality of convolution blocks.

[0022] According to one embodiment of the present invention, the step of performing the user authentication may be to derive a distance between the representative feature value and at least one identification feature value among a plurality of identification feature values ​​previously stored in a database, and compare the derived distance with a preset threshold value to determine whether the subject is a previously stored user, thereby performing the user authentication.

[0023] According to one embodiment of the present invention, the step of performing the user authentication may sequentially derive the distance for each of the representative feature value and the plurality of identification feature values ​​in a preset order, but if the distance between the representative feature value and any one of the plurality of identification feature values ​​is greater than or equal to the preset threshold value, the distance derivation for the identification feature value corresponding to the order after the representative feature value and any one of the identification feature values ​​may be omitted.

[0024] According to one embodiment of the present invention, the step of performing the preprocessing may include a step of performing preprocessing on an image of the fingerprint data and a step of performing preprocessing on the blood flow data.

[0025] According to one embodiment of the present invention, the step of performing image preprocessing on the fingerprint data may include performing at least one image preprocessing process among background removal, image scaling, image size adjustment, and data augmentation on the fingerprint data.

[0026] According to one embodiment of the present invention, the step of performing preprocessing on the blood flow data may include performing at least one signal preprocessing process among resampling, filtering, segmentation, and baseline correction on the blood flow data.

[0027] The above-described problem-solving methods are merely exemplary and should not be construed as limiting the present invention. In addition to the exemplary embodiments described above, additional embodiments may be included in the drawings and detailed description of the invention.

[0028] According to the aforementioned means for solving the problem of the present invention, by constructing and utilizing a deep learning-based feature extraction model capable of extracting each feature from fingerprint data and blood flow data, the accuracy of an individual's biological characteristics included in fingerprint data and blood flow data can be improved.

[0029] According to the aforementioned means for solving the problem of the present invention, user authentication is performed based on multifactor-based feature values ​​that combine features extracted from fingerprint data and blood flow data, thereby preventing abnormal access and improving authentication accuracy and security compared to conventional technologies.

[0030] However, the effects that can be obtained from this center are not limited to the effects described above, and other effects may exist.

[0031] Figure 1 is a schematic diagram of a multifactor-based biometric authentication system according to one embodiment of the present invention.

[0032] FIG. 2 is a block diagram schematically illustrating a user registration process and a user authentication process performed in a multifactor biometric authentication device according to one embodiment of the present invention.

[0033] FIG. 3 is a drawing showing in more detail a user registration process and a user authentication process performed by a multifactor biometric authentication device according to one embodiment of the present invention.

[0034] Figure 4 is a diagram exemplarily showing a preprocessing process of blood flow data according to one embodiment of the present invention.

[0035] FIG. 5 is a diagram exemplarily showing a preprocessing process of fingerprint data according to one embodiment of the present invention.

[0036] Fig. 6 is a diagram exemplifying an algorithm of a feature extraction model according to one embodiment of the present invention.

[0037] Figure 7 is a diagram showing a process for generating representative feature values ​​according to one embodiment of the present invention.

[0038] FIG. 8 is a diagram exemplifying a tree structure for a plurality of identification feature values ​​stored in a database according to one embodiment of the present invention.

[0039] FIG. 9 is a schematic block diagram of a multifactor-based biometric authentication device according to one embodiment of the present invention.

[0040] Figure 10 is a flowchart of a multifactor-based biometric authentication method according to one embodiment of the present invention.

[0041] Below, with reference to the attached drawings, embodiments of the present invention are described in detail to facilitate easy implementation by those skilled in the art. However, the present invention can be implemented in various different forms and is not limited to the embodiments described herein. In the drawings, irrelevant parts have been omitted for clarity, and similar reference numerals have been used throughout the specification to indicate similar elements.

[0042] Throughout this specification, when a part is said to be "connected" to another part, this includes not only the case where it is "directly connected," but also the case where it is "electrically connected" or "indirectly connected" with another element in between.

[0043] Throughout this specification, when it is said that a member is located “on,” “above,” “upper,” “lower,” “lower” or “lower” another member, this includes not only cases where the member is in contact with the other member, but also cases where another member exists between the two members.

[0044] Throughout this specification, whenever a part is said to "include" a component, this does not mean that it excludes other components, but rather that it may include other components, unless otherwise specifically stated.

[0045] The present invention relates to a multi-factor based biometric authentication device and method.

[0046] Figure 1 is a schematic diagram of a multifactor-based biometric authentication system according to one embodiment of the present invention.

[0047] Referring to FIG. 1, a multifactor-based biometric authentication system (10) (hereinafter referred to as “authentication system (10)”) may include a multifactor-based biometric authentication device (100) (hereinafter referred to as “authentication device (100)”), a biometric information measuring device (200), and a terminal (300).

[0048] According to one embodiment of the present invention, the authentication system (10) can determine whether the subject (1) is a user based on multifactor data including fingerprint data and blood flow data of the subject (1).

[0049] In this regard, the biometric information measuring device (200) may be for measuring fingerprint data and blood flow data of a subject (1). For example, the biometric information measuring device (200) may include a contact portion (not shown) with which a finger of the subject (1) comes into contact. The biometric information measuring device (200) may measure fingerprint data and blood flow data of the subject (1) through the finger of the subject (1) coming into contact with the contact portion (not shown) and transmit the measured data to the authentication device (100).

[0050] Additionally, the terminal (300) may display information about a subject who has attempted user authentication through the authentication device (100) and a registered user. For example, the terminal (300) may be possessed by the administrator of the authentication system (10) and may be used by the administrator to verify and manage user information stored in a database. In this case, the user information may include the user's name, age, registration date, etc., and may include other necessary information depending on the space and system in which the authentication system (10) is utilized.

[0051] The terminal (300) may be a device that is linked to the authentication device (100) and the biometric information measuring device (200) through a network, and may be, for example, a smartphone, a smart pad, a tablet PC, a wearable device, and all types of wireless communication devices such as a PCS (Personal Communication System), a GSM (Global System for Mobile communication), a PDC (Personal Digital Cellular), a PHS (Personal Handyphone System), a PDA (Personal Digital Assistant), an IMT (International Mobile Telecommunication)-2000, a CDMA (Code Division Multiple Access)-2000, a W-CDMA (W-Code Division Multiple Access), a Wibro (Wireless Broadband Internet) terminal, and a desktop computer, a smart TV, and a stationary terminal.

[0052] In addition, examples of networks for sharing information between the authentication device (100), the biometric information measuring device (200), and the terminal (300) may include, but are not limited to, a 3rd Generation Partnership Project (3GPP) network, a Long Term Evolution (LTE) network, a 5G network, a World Interoperability for Microwave Access (WIMAX) network, a wired / wireless Internet, a Local Area Network (LAN), a Wireless Local Area Network (Wireless LAN), a Wide Area Network (WAN), a Personal Area Network (PAN), a Bluetooth network, a Wifi network, an NFC (Near Field Communication) network, a satellite broadcasting network, an analog broadcasting network, a DMB (Digital Multimedia Broadcasting) network, etc.

[0053] FIG. 2 is a block diagram schematically illustrating a user registration process and a user authentication process performed in a multifactor biometric authentication device according to one embodiment of the present invention.

[0054] Referring to FIG. 2, the authentication device (100) can perform user registration and user authentication. The authentication device (100) can obtain fingerprint data and blood flow data from a subject (1), perform preprocessing on each of the obtained fingerprint data and blood flow data, and extract features for each of the fingerprint data and blood flow data.

[0055] At this time, referring to FIG. 2, the authentication device (100) may store the characteristics of each of the acquired fingerprint data and blood flow data in a database when performing user registration. Specifically, the authentication device (100) may receive user information through the biometric information measuring device (200) or a separate terminal provided with the biometric information measuring device (200), and store the user information in a database in connection with the characteristics of each of the acquired fingerprint data and blood flow data, thereby registering the subject (1) as a user.

[0056] In addition, referring to FIG. 2, when performing user authentication, the authentication device (100) can perform user authentication by comparing the multifactor data generated through data combination and averaging using the features of each of fingerprint data and blood flow data with biometric information (features of fingerprint data and features of blood flow data, etc.) about a registered user already stored in a database to determine whether the subject (1) is a user.

[0057] Below, the biometric authentication process performed in the authentication device (100) will be described in detail.

[0058] FIG. 3 is a drawing showing in more detail a user registration process and a user authentication process performed by a multifactor biometric authentication device according to one embodiment of the present invention.

[0059] Referring to FIG. 3, the authentication device (100) can obtain fingerprint data and blood flow data of a subject (1) through a biometric information measurement device (200). At this time, the fingerprint data may be image data obtained by photographing the fingerprint of the subject (1), and the blood flow data may be data obtained by measuring the optical blood flow flowing in the finger of the subject (1) by irradiating the finger with light, and may be a PPG (Photoplethysmography) signal.

[0060] In addition, the authentication device (100) may acquire a plurality of fingerprint data and a plurality of blood flow data. Specifically, the authentication device (100) may receive fingerprint data and blood flow data that are measured continuously multiple times from the biometric information measuring device (200). At this time, the plurality of fingerprint data and the plurality of blood flow data may be measured for a preset time by the biometric information measuring device (200). Hereinafter, for the convenience of explanation, the word "plural" is omitted, but it should be interpreted that the meaning is included even if the fingerprint data and blood flow data are not modified with the word "plural."

[0061] Referring to FIG. 3, the authentication device (100) may obtain fingerprint data and blood flow data as either registration data for registration of the subject (1) or authentication data for user authentication of the subject (1).

[0062] However, since the registration process and authentication process performed in the authentication device (100) are performed in the same process up to the process of generating a representative feature value for the subject (1), the description up to the process of generating a representative feature value will be described below without distinguishing between the registration process and the authentication process.

[0063] Referring to FIG. 3, the authentication device (100) can perform preprocessing on fingerprint data and blood flow data. Specifically, since the data formats of the fingerprint data and blood flow data are different, the authentication device (100) can perform image preprocessing on the fingerprint data and preprocessing on the blood flow data.

[0064] Figure 4 is a diagram exemplarily showing a preprocessing process of blood flow data according to one embodiment of the present invention.

[0065] Referring to FIG. 4, the authentication device (100) can perform at least one image preprocessing process among normalization, background removal, image scaling, image resizing, sharpening, data augmentation, binarization, and skeletonization (thinning) on ​​fingerprint data.

[0066] FIG. 5 is a diagram exemplarily showing a preprocessing process of fingerprint data according to one embodiment of the present invention.

[0067] Referring to FIG. 5, the authentication device (100) can perform at least one signal preprocessing process among resampling, differentiation, filtering, segmentation, baseline correction, and normalization on blood flow data.

[0068] Referring again to FIG. 3, the authentication device (100) can extract fingerprint features and blood flow features from fingerprint data and blood flow data, respectively.

[0069] The authentication device (100) can extract fingerprint features in vector form from fingerprint data and blood flow features in vector form from blood flow data using a feature extraction model. In this regard, the authentication device (100) can utilize two separate feature extraction models to extract features for fingerprint data and blood flow data, which are different types of data, respectively. However, the present invention is not limited thereto, and a single model capable of extracting features from two types of data may be utilized.

[0070] Fig. 6 is a diagram exemplifying an algorithm of a feature extraction model according to one embodiment of the present invention.

[0071] At this time, the feature extraction model may be a multimodal synthetic deep learning model combining a CNN neural network (CNN) and a Long Short Term Memory (LSTM), but is not limited thereto, and for example, various neural network models for machine learning that are already known in the past or will be developed in the future, such as a convolutional neural network (CNN), a recurrent neural network (RNN), a deep neural network, a Residual Network (ResNet), a transformer model, and an attention machine learning (Attention mechanism) algorithm, may be applied, and may be replaced with a machine learning (machine learning) model, a neural network model (artificial neural network model), a neuro-fuzzy model, etc.

[0072] Referring to FIG. 6, the feature extraction model may be trained to have a hierarchical structure including a plurality of convolution blocks. The plurality of convolution blocks may be for extracting features in the form of a single vector including a plurality of elements. In other words, the feature extraction model may extract a feature vector of the corresponding data by inputting at least one of fingerprint data and blood flow data and outputting at least one single vector comprehensively including a plurality of elements corresponding to the input data. In other words, the fingerprint features and blood flow features may be in the form of vectors.

[0073] Additionally, each of the plurality of convolution blocks may include, but is not limited to, a convolutional layer, a leaky ReLU activation function, an L1L2 regularization layer, an average pooling layer, and a layer normalization layer.

[0074] In other words, the authentication device (100) may acquire a plurality of fingerprint features and a plurality of blood flow features for each of the acquired plurality of fingerprint data and the plurality of blood flow data. Accordingly, the authentication device (100) may improve the accuracy of the individual's biometric characteristics included in the fingerprint data and blood flow data by constructing and utilizing a deep learning-based feature extraction model capable of extracting each feature from the fingerprint data and the blood flow data.

[0075] According to one embodiment of the present invention, the authentication device (100) can generate a representative feature value for a subject by combining and averaging fingerprint features and blood flow features.

[0076] Figure 7 is a diagram showing a process for generating representative feature values ​​according to one embodiment of the present invention.

[0077] Referring to FIG. 7, as described above, the authentication device (100) extracts a plurality of fingerprint features and a plurality of blood flow features through a feature extraction model, and generates a multi-factor feature by combining the fingerprint features and blood flow features for each of the plurality of fingerprint data and the plurality of blood flow data that correspond to each other (data measured simultaneously or within the same time interval), thereby generating a plurality of multi-factor features for the plurality of fingerprint data and the plurality of blood flow data.

[0078] At this time, the authentication device (100) can group multiple multi-factor features for the subject (1) and generate a feature vector set that comprehensively includes features of biometric information for the subject (1). In other words, the authentication device (100) can generate a feature vector set using multi-factor features generated by combining fingerprint features and blood flow features output through a feature extraction model.

[0079] In addition, referring to FIG. 7, the authentication device (100) can obtain a single feature vector that comprehensively includes multiple feature information about the subject (1) by averaging multiple multi-factor features. That is, the authentication device (100) can perform averaging on a set of feature vectors to generate a representative feature value in the form of a single vector.

[0080] The authentication device (100) has the advantage of ensuring storage efficiency and computational economy when performing authentication by generating only one feature vector that comprehensively includes biometric information of one subject (1) using multiple features and utilizing this for user registration and user authentication.

[0081] Referring again to FIG. 3, the authentication device (100) can perform user authentication of the subject based on the representative feature value.

[0082] According to one embodiment of the present invention, the authentication device (100) can perform user authentication by deriving a distance between a representative feature value and at least one identification feature value among a plurality of identification feature values ​​previously stored in a database, comparing the derived distance with a preset threshold value, and determining whether the subject is a previously stored user.

[0083] At this time, multiple identification feature values, which are representative feature values ​​for each registered user already stored in the database, are c s , and the representative feature value for the subject (1) attempting user authentication is c j, and multiple identification feature values ​​c s and representative feature value c j The minimum distance among the distances between d j When expressed as , d is the distance between the representative feature value of the subject (1) and the most similar identification feature value, that is, the distance between the most similar identification feature value and the representative feature value of the subject (1). j can be expressed as in [Formula 1] below.

[0084] [Formula 1]

[0085]

[0086] Here, f may represent a distance function. The distance function may include various distance functions that are already known in the past or will be developed in the future, such as cosine similarity and Euclidean.

[0087] According to one embodiment of the present invention, the authentication device (100) is configured to determine a distance d between a representative feature value and one of a plurality of identification feature values ​​having the closest distance. j If the value is equal to or less than a preset threshold value, the subject (1) corresponding to the representative feature value can be authenticated as a user corresponding to any one of the identification feature values.

[0088] In addition, the authentication device (100) is configured to determine the distance d between the representative feature value and one of the plurality of identification feature values ​​having the closest distance. j If it is greater than a preset threshold value, it can be determined that the subject (1) corresponding to the representative feature value is not a pre-registered user.

[0089] At this time, when the preset threshold is expressed as t, the authentication variable R for the subject (1) j can be expressed as in [Formula 2] below.

[0090] [Formula 2]

[0091]

[0092] In other words, the authentication device (100) provides an authentication variable R for the subject (1). j If it is 1, it can be determined that the target (1) is a pre-registered user, and if it is 0, it can be determined that the target (1) is not a pre-registered user.

[0093] In this regard, the authentication device (100) may have an optimal threshold value preset through a linear regression model. The preset optimal threshold value may be set as shown in [Equation 3] below.

[0094] [Formula 3]

[0095]

[0096] Here, t is the optimal threshold, β0, β1, and β2 are linear regression coefficients, 1 may mean the amount of data used for learning, and 2 may mean the amount of data used for authentication.

[0097] As described above, the authentication device (100) can perform user authentication by deriving all distances between the representative feature value and all identification feature values ​​stored in the database and comparing the identification feature value with the closest distance with a preset threshold value, but is not limited thereto. If one identification feature value is derived that has a distance close to the preset threshold value according to user settings, the authentication device (100) may authenticate the subject (1) as a user corresponding to the identification feature value. However, the threshold value at this time may be a smaller value than the threshold value used when deriving the identification feature value with the minimum distance.

[0098] In this regard, the authentication device (100) sequentially derives the distances for the representative feature value and each of the plurality of identification feature values ​​in a preset order, but if the distance between the representative feature value and any one of the plurality of identification feature values ​​is greater than or equal to a preset threshold value, the distance derivation for the identification feature value corresponding to the order after the representative feature value and any one of the identification feature values ​​may be omitted.

[0099] Specifically, the authentication device (100) may sequentially derive the distance between the representative feature value and each of the plurality of identification feature values ​​previously stored in the database according to a predetermined distance derivation order, and when an identification feature value having a smaller distance than the representative feature value and a preset threshold value is found, the authentication device (100) may determine that the user corresponding to the identification feature value is the same person as the subject (1), and may complete user authentication of the subject. Accordingly, the authentication device (100) may determine that deriving distances with other identification feature values ​​corresponding to the distance derivation order after the identification feature value is unnecessary and may omit it.

[0100] Accordingly, since the authentication device (100) does not always have to perform distance derivation with all identification feature values ​​previously stored in the database, the time required for user authentication can be reduced. However, in this case, the later the predetermined distance derivation order of the identification feature value is, the more distances need to be derived with respect to the identification feature value when authenticating the user corresponding to the identification feature value, and the time required for user authentication for each user varies depending on the distance derivation order.

[0101] To solve this problem, the authentication device (100) can reduce the difference in time required for user authentication for each user by performing distance derivation according to a random comparison order for each of the plurality of identification features. In other words, the authentication device (100) sequentially performs distance derivation between the representative feature value and the identification feature value previously stored in the database according to a predetermined distance derivation order, but the predetermined distance derivation order may be randomly changed each time user authentication is performed.

[0102] Meanwhile, when the authentication device (100) is used for user authentication for a subject (1) to enter a specific institution or access a specific platform, the frequency of user authentication may be higher for users who have recently entered or accessed the institution or system, depending on the characteristics of the institution or system. To prepare for such cases, the authentication device (100) may prioritize the comparison order of identification characteristic values ​​corresponding to a user who has been recently authenticated, thereby enabling user authentication to be completed in a shorter time for a subject who has been recently authenticated.

[0103] In other words, if the number of identification feature values ​​stored in the database exceeds a preset threshold number, the authentication device (100) may perform distance derivation according to a preset distance derivation order based on recent authentication records for each of the plurality of identification feature values.

[0104] That is, if the number of identification feature values ​​already stored in the database is less than or equal to a preset threshold number, the authentication device (100) sequentially performs distance derivation with respect to the representative feature value for all identification feature values, and if the number of identification feature values ​​already stored in the database exceeds the preset threshold number, distance derivation is sequentially performed according to the distance derivation order only until an identification feature value having a distance less than the preset threshold value from the representative feature value is found, and similarity comparison for the identification feature values ​​for the subsequent comparison order can be omitted.

[0105] Additionally, the distance derivation order may be random and may be set based on recent authentication records. However, this is not limited to this, and the distance derivation order may be set in various ways other than the above examples, and may be set flexibly depending on the characteristics of the space and system to which the authentication device (100) is applied.

[0106] According to the above, the authentication device (100) derives a distance between a representative feature value generated using the subject's fingerprint data and blood flow data and a plurality of previously stored identification feature values, and performs user authentication based on the derived distance. However, by sequentially performing distance derivation according to a preset distance derivation order, the distance derivation and comparison process for user authentication can be simplified and the time required for user authentication can be reduced.

[0107] Meanwhile, if it is determined that the subject (1) is not a registered user (unregistered user) (the identification characteristic value having a close distance from the representative characteristic value of the subject (1) to a preset threshold value is not stored in the database), the authentication device (100) may, through the biometric information measuring device (200), notify that the subject (1) is not a user (unregistered), receive a selection input from the subject (1) as to whether to register as a user, and store the representative characteristic value for the subject in the database based on the received selection input, thereby newly registering the subject (1) as a new user.

[0108] In this regard, the database may be sequentially stored in the order in which multiple identification characteristic values ​​are registered (the order in which new users are registered), but is not limited thereto, and the authentication device (100) may store the values ​​in such a way that the comparison order of users with a high frequency of user authentication is located at the top according to the number of times user authentication is performed based on a tree structure.

[0109] FIG. 8 is a diagram exemplifying a tree structure for a plurality of identification feature values ​​stored in a database according to one embodiment of the present invention.

[0110] Referring to Figure 8, it can be confirmed that the higher the node is located in the tree structure, the higher the comparison order is. In this regard, the number of each node in the tree structure shown in Figure 8 may indicate the order in which identification features are registered (the order in which new users are registered).

[0111] Referring to Fig. 8 as an example, the user corresponding to the identification feature value 2 may be the user with the highest frequency of user authentication among the plurality of users for the plurality of identification feature values ​​previously stored in the database. In this way, the authentication device (100) utilizes an algorithm that sorts and stores users with a high frequency of user authentication in the upper layer of the database based on a tree structure, so that even if the comparison order is the same as the order stored in the database, the more frequently the identification feature value of the user with a high frequency of user authentication is used, the more preferentially the distance from the representative feature value of the target (1) is derived, and through this, the user authentication can be processed more quickly for the user with a high frequency of user authentication.

[0112] However, it is not limited thereto, and the authentication device (100) according to one embodiment of the present invention may store the final characteristic value for the subject in a database (register the subject as a new user) by applying various scheduling algorithms such as priority queue and round robin in the field of computer science as well as the tree structure exemplified above.

[0113] According to the above, the authentication device (100) can perform user authentication based on a multifactor-based feature value that combines features extracted from fingerprint data and blood flow data, thereby preventing abnormal access and improving authentication accuracy and security compared to the prior art.

[0114] FIG. 9 is a schematic block diagram of a multifactor-based biometric authentication device according to one embodiment of the present invention.

[0115] Referring to FIG. 9, the authentication device (100) may include a data acquisition unit (110), a data preprocessing unit (120), a feature extraction unit (130), a representative feature value generation unit (140), and a user authentication unit (150).

[0116] According to one embodiment of the present invention, the data acquisition unit (110) can acquire fingerprint data and blood flow data from the subject (1).

[0117] According to one embodiment of the present invention, the data preprocessing unit (120) may perform preprocessing on fingerprint data and blood flow data. Specifically, the data preprocessing unit (120) may include a first data preprocessing unit (121) that performs image preprocessing on fingerprint data and a second data preprocessing unit (122) that performs preprocessing on blood flow data.

[0118] At this time, the first data preprocessing unit (121) may perform at least one image preprocessing process among background removal, image scaling, image size adjustment, and data augmentation on the fingerprint data. In addition, the second data preprocessing unit (122) may perform at least one signal preprocessing process among resampling, filtering, segmentation, and baseline correction on the blood flow data.

[0119] According to one embodiment of the present invention, the feature extraction unit (130) can extract fingerprint features and blood flow features from fingerprint data and blood flow data, respectively. Specifically, the feature extraction unit (130) can extract fingerprint features in vector form from fingerprint data and blood flow features in vector form from blood flow data using a feature extraction model. At this time, the feature extraction model may be trained to have a hierarchical structure including a plurality of convolution blocks.

[0120] According to one embodiment of the present invention, the representative feature value generation unit (140) can generate a representative feature value for a subject by combining and averaging fingerprint features and blood flow features. Specifically, the representative feature value generation unit (140) can generate a feature vector set using multi-factor features generated by combining fingerprint features and blood flow features output through a feature extraction model, and perform an average on the feature vector set to generate a representative feature value in the form of a single vector.

[0121] According to one embodiment of the present invention, the user authentication unit (150) can perform user authentication of a subject based on a representative feature value. Specifically, the user authentication unit (150) can perform user authentication by deriving a distance between the representative feature value and at least one identification feature value among a plurality of identification feature values ​​previously stored in a database, comparing the derived distance with a preset threshold value, and determining whether the subject (1) is a previously stored user.

[0122] In addition, the user authentication unit (150) sequentially derives the distances for the representative feature value and each of the plurality of identification feature values ​​in a preset order, but if the distance between the representative feature value and any one of the plurality of identification feature values ​​is greater than or equal to a preset threshold value, the distance derivation for the identification feature value corresponding to the order after the representative feature value and any one of the identification feature values ​​may be omitted.

[0123] Below, we will briefly review the operating flow of the present invention based on the detailed description above.

[0124] Figure 10 is a flowchart of a multifactor-based biometric authentication method according to one embodiment of the present invention.

[0125] The multifactor-based biometric authentication method illustrated in FIG. 10 can be performed by the multifactor-based biometric authentication device (100) described above. Therefore, even if omitted below, the description of the multifactor-based biometric authentication device (100) can be equally applied to the description of the method.

[0126] Referring to FIG. 10, in step S11, the data acquisition unit (110) can acquire fingerprint data and blood flow data from the subject (1).

[0127] Next, in step S12, the data preprocessing unit (120) may perform preprocessing on fingerprint data and blood flow data. Specifically, the data preprocessing unit (120) may include a first data preprocessing unit (121) that performs image preprocessing on fingerprint data and a second data preprocessing unit (122) that performs preprocessing on blood flow data.

[0128] At this time, the first data preprocessing unit (121) may perform at least one image preprocessing process among background removal, image scaling, image size adjustment, and data augmentation on the fingerprint data. In addition, the second data preprocessing unit (122) may perform at least one signal preprocessing process among resampling, filtering, segmentation, and baseline correction on the blood flow data.

[0129] Next, in step S13, the feature extraction unit (130) can extract fingerprint features and blood flow features from the fingerprint data and blood flow data, respectively. Specifically, in step S13, the feature extraction unit (130) can extract fingerprint features in vector form from the fingerprint data and blood flow features in vector form from the blood flow data using a feature extraction model. At this time, the feature extraction model may be trained to have a hierarchical structure including a plurality of convolution blocks.

[0130] Next, in step S14, the representative feature value generation unit (140) can generate a representative feature value for the subject by combining and averaging the fingerprint features and blood flow features. Specifically, in step S14, the representative feature value generation unit (140) can generate a feature vector set using multi-factor features generated by combining the fingerprint features and blood flow features output through the feature extraction model, and perform averaging on the feature vector set to generate a representative feature value in the form of a single vector.

[0131] Next, in step S15, the user authentication unit (150) can perform user authentication of the subject based on the representative feature value. Specifically, in step S15, the user authentication unit (150) can perform user authentication by deriving a distance between the representative feature value and at least one identification feature value among a plurality of identification feature values ​​previously stored in the database, comparing the derived distance with a preset threshold value, and determining whether the subject (1) is a previously stored user.

[0132] In addition, in step S15, the user authentication unit (150) sequentially derives the distance for each of the representative feature value and the plurality of identification feature values ​​in a preset order, but if the distance between the representative feature value and any one of the plurality of identification feature values ​​is greater than or equal to a preset threshold value, the distance derivation for the identification feature value corresponding to the order after the representative feature value and any one of the identification feature values ​​may be omitted.

[0133] In the above description, steps S11 to S15 may be further divided into additional steps or combined into fewer steps, depending on the implementation example of the present invention. Furthermore, some steps may be omitted as needed, and the order of the steps may be changed.

[0134] A multifactor-based biometric authentication method according to an embodiment of the present invention may be implemented in the form of program commands that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program commands, data files, data structures, etc., either singly or in combination. The program commands recorded on the medium may be those specially designed and configured for the present invention or may be those known and available 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 media such as CD-ROMs and DVDs, magneto-optical media such as floptical disks, and hardware devices specially configured to store and execute program commands, such as ROMs, RAMs, and flash memories. Examples of program commands include not only machine language codes generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc. The above hardware devices may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

[0135] Additionally, the aforementioned multifactor-based biometric authentication method can also be implemented in the form of a computer program or application executed by a computer and stored in a recording medium.

[0136] The above description of the present invention is for illustrative purposes only, and those skilled in the art will readily appreciate that the present invention can be readily modified into other specific forms without altering the technical spirit or essential characteristics of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single entity may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined manner.

[0137] The scope of the present invention is indicated by the claims described below rather than the detailed description above, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention.

Claims

1. In a multifactor-based biometric authentication device, A data acquisition unit that acquires fingerprint data and blood flow data from a subject; A data preprocessing unit that performs preprocessing on the fingerprint data and the blood flow data; A feature extraction unit that extracts fingerprint features and blood flow features from the fingerprint data and blood flow data, respectively; A representative feature value generation unit that generates a representative feature value for the subject by combining and averaging the fingerprint features and the blood flow features; and A user authentication unit that performs user authentication of the subject based on the representative feature value; An authentication device including:

2. In paragraph 1, The above feature extraction unit, An authentication device that extracts fingerprint features in vector form from the fingerprint data and blood flow features in vector form from the blood flow data using a feature extraction model.

3. In paragraph 2, The representative feature value generation unit above is, An authentication device that generates a feature vector set using multifactor features generated by combining the fingerprint features and blood flow features output through the feature extraction model, and performs averaging on the feature vector set to generate the representative feature value in the form of a single vector.

4. In paragraph 2, The above feature extraction model is, An authentication device characterized in that it is trained to have a hierarchical structure including a plurality of convolution blocks.

5. In paragraph 1, The above user authentication unit is, An authentication device that performs user authentication by deriving a distance between the representative feature value and at least one identification feature value among a plurality of identification feature values ​​previously stored in a database, comparing the derived distance with a preset threshold value, and determining whether the subject is a previously stored user.

6. In paragraph 5, The above user authentication unit is, The distances for the representative feature values ​​and the plurality of identification feature values ​​are sequentially derived in a preset order. An authentication device, wherein if the distance between the representative feature value and any one of the plurality of identification feature values ​​is greater than or equal to the preset threshold value, distance derivation for the identification feature value corresponding to the order after the representative feature value and any one of the identification feature values ​​is omitted.

7. In paragraph 1, The above data preprocessing unit, A first data preprocessing unit that performs image preprocessing on the above fingerprint data; and A second data preprocessing unit that performs preprocessing on the above blood flow data, An authentication device comprising:

8. In paragraph 7, The above first data preprocessing unit, An authentication device that performs at least one image preprocessing process among background removal, image scaling, image size adjustment, and data augmentation on the above fingerprint data.

9. In paragraph 7, The above second data preprocessing unit, An authentication device that performs at least one signal preprocessing process among resampling, filtering, segmentation, and baseline correction on the above blood flow data.

10. In a multifactor-based biometric authentication method, A step of obtaining fingerprint data and blood flow data from a subject; A step of performing preprocessing on the fingerprint data and the blood flow data; A step of extracting fingerprint features and blood flow features from the fingerprint data and the blood flow data, respectively; A step of combining and averaging the fingerprint features and the blood flow features to generate a representative feature value for the subject; and A step of performing user authentication of the subject based on the representative feature value; Authentication method including.

11. In paragraph 10, The steps of extracting the above fingerprint features and blood flow features are: An authentication method comprising: extracting fingerprint features in vector form from the fingerprint data and extracting blood flow features in vector form from the blood flow data using a feature extraction model.

12. In paragraph 11, The step of generating the above representative feature value is: An authentication method comprising: generating a feature vector set using multifactor features generated by combining the fingerprint features and blood flow features output through the feature extraction model; and performing averaging on the feature vector set to generate the representative feature value in the form of a single vector.

13. In paragraph 11, The above feature extraction model is, An authentication method characterized in that it is learned to have a hierarchical structure including a plurality of convolution blocks.

14. In paragraph 10, The steps for performing the above user authentication are: An authentication method comprising: deriving a distance between the representative feature value and at least one identification feature value among a plurality of identification feature values ​​already stored in a database; comparing the derived distance with a preset threshold value to determine whether the subject is a previously stored user; and performing the user authentication.

15. In paragraph 14, The steps for performing the above user authentication are: The distances for the representative feature values ​​and the plurality of identification feature values ​​are sequentially derived in a preset order. An authentication method, wherein if the distance between the representative feature value and any one of the plurality of identification feature values ​​is greater than or equal to the preset threshold value, distance derivation for the identification feature value corresponding to the order after the representative feature value and any one of the identification feature values ​​is omitted.

16. In paragraph 10, The step of performing the above preprocessing is: A step of performing image preprocessing on the above fingerprint data; and A step of performing preprocessing on the above blood flow data, An authentication method including:

17. In paragraph 16, The step of performing image preprocessing on the above fingerprint data is as follows: An authentication method comprising performing at least one image preprocessing process among background removal, image scaling, image size adjustment, and data augmentation on the above fingerprint data.

18. In paragraph 16, The step of performing preprocessing on the above blood flow data is: An authentication method comprising performing at least one signal preprocessing process among resampling, filtering, segmentation, and baseline correction on the above blood flow data.

19. A computer-readable recording medium recording a program for executing the method of any one of clauses 10 to 18 on a computer.

Images