Identity identification method, device and equipment and storage medium

Abstract

The present application provides an identity recognition method, device and equipment, and a storage medium, and relates to the technical field of biological recognition. The method comprises the following steps: obtaining a first face image and a first iris image of a first user; determining a plurality of first candidate identity identifiers of the first user and a face matching degree of each first candidate identity identifier with the first user according to the first face image; determining a plurality of second candidate identity identifiers of the first user and an iris matching degree of each second candidate identity identifier with the first user according to the first iris image; and determining a target identity identifier of the first user according to the plurality of first candidate identity identifiers, the plurality of second candidate identity identifiers, the face matching degree of each first candidate identity identifier with the first user, and the iris matching degree of each second candidate identity identifier with the first user. The method provided by the present application solves the problems of low recognition rate and low security when identity recognition is performed only by using face recognition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of biometric identification, and particularly relates to an identity recognition method and device, equipment and a storage medium. BACKGROUND

[0002] With the continuous development of biometric identification technology, more and more enterprises apply it in the production process.

[0003] In the financial system, when processing large amount of transfer business, for the sake of security and convenience, face recognition technology is widely used for identity recognition before transfer. The face recognition transfer method not only greatly improves the security compared with the traditional password transfer, but also brings great convenience to people's life.

[0004] The face recognition technology has the shortcomings of recognition rate decline caused by appearance change and security decline caused by face copy. SUMMARY

[0005] The present application provides an identity recognition method, device, equipment and storage medium, which solves the problem of low recognition rate and low security when only using face recognition for identity recognition, and provides convenient and easy-to-use services for some users with damaged biological characteristics, and optimizes the customer experience.

[0006] In one aspect, the present application provides an identity recognition method, comprising:

[0007] obtaining a first face image and a first iris image of a first user;

[0008] determining a plurality of first candidate identity identifiers of the first user and a face matching degree of each first candidate identity identifier according to the first face image;

[0009] determining a plurality of second candidate identity identifiers of the first user and an iris matching degree of each second candidate identity identifier according to the first iris image;

[0010] determining a target identity identifier of the first user from the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers according to the plurality of first candidate identity identifiers, the plurality of second candidate identity identifiers, the face matching degree of each first candidate identity identifier and the iris matching degree of each second candidate identity identifier.

[0011] Optionally, the target identity of the first user is determined from the plurality of first candidate identity and the plurality of second candidate identity according to the plurality of first candidate identity, the plurality of second candidate identity, the face matching degree of each first candidate identity with the first user, and the iris matching degree of each second candidate identity with the first user, comprising:

[0012] determining at least one coincident candidate identity existing in the plurality of first candidate identity and the plurality of second candidate identity;

[0013] determining the target identity from the at least one coincident candidate identity according to the face matching degree of each coincident candidate identity with the first user, and the iris matching degree of each coincident candidate identity with the first user.

[0014] Optionally, the target identity is determined from the at least one coincident candidate identity according to the face matching degree of each coincident candidate identity with the first user, and the iris matching degree of each coincident candidate identity with the first user, comprising:

[0015] for any one coincident candidate identity, determining a fusion matching degree of the coincident candidate identity with the first user according to the face matching degree of the coincident candidate identity with the first user, and the iris matching degree of the coincident candidate identity with the first user;

[0016] determining the target identity from the at least one coincident candidate identity according to the fusion matching degree of each coincident candidate identity with the first user.

[0017] Optionally, the target identity is determined from the at least one coincident candidate identity according to the fusion matching degree of each coincident candidate identity with the first user, comprising:

[0018] determining the target identity as the coincident candidate identity with the highest fusion matching degree with the first user from the at least one coincident candidate identity.

[0019] Optionally, the plurality of first candidate identity of the first user and the face matching degree of each first candidate identity with the first user are determined according to the first face image, comprising:

[0020] obtaining a first face feature vector of the first face image;

[0021] obtaining a face similarity of the first face feature vector with each face feature vector in a face database;

[0022] determine a plurality of candidate facial feature vectors in the facial database according to the first facial feature vector and a facial similarity between the first facial feature vector and each facial feature vector in the facial database;

[0023] determine the identity corresponding to the candidate facial feature vector as the first candidate identity;

[0024] determine a similarity between the candidate facial feature vector corresponding to the first candidate identity and the first facial feature vector as a facial matching degree between the first candidate identity and the first user.

[0025] Optionally, determining a plurality of second candidate identities of the first user and a degree of iris matching between each second candidate identity and the first user according to the first iris image comprises:

[0026] obtain a first iris feature vector of the first iris image;

[0027] obtain an iris similarity between the first iris feature vector and each iris feature vector in an iris database;

[0028] determine a plurality of candidate iris feature vectors in the iris database according to the first iris feature vector and the iris similarity between the first iris feature vector and each iris feature vector in the iris database;

[0029] determine the identity corresponding to the candidate iris feature vector as the second candidate identity;

[0030] determine a similarity between the candidate iris feature vector corresponding to the second candidate identity and the first iris feature vector as the degree of iris matching between the second candidate identity and the first user.

[0031] Optionally, obtaining the first facial image of the first user comprises

[0032] obtain an initial image of the first user captured by the camera device;

[0033] perform at least one image processing operation on the initial image to obtain the first facial image, the at least one image processing operation comprising cropping, rotating or angle adjusting.

[0034] In another aspect, the present application provides an identity recognition device, comprising:

[0035] an obtaining module configured to obtain a first facial image and a first iris image of a first user;

[0036] determine a plurality of first candidate identity identifiers of the first user according to the first face image, and a face matching degree of each first candidate identity identifier with the first user;

[0037] determine a plurality of second candidate identity identifiers of the first user according to the first iris image, and an iris matching degree of each second candidate identity identifier with the first user;

[0038] determine a target identity identifier of the first user in the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers according to the plurality of first candidate identity identifiers, the plurality of second candidate identity identifiers, the face matching degree of each first candidate identity identifier with the first user, and the iris matching degree of each second candidate identity identifier with the first user.

[0039] In a possible implementation, the determining module is specifically configured to:

[0040] determine at least one coincident candidate identity identifier existing in the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers;

[0041] determine the target identity identifier in the at least one coincident candidate identity identifier according to the face matching degree of each coincident candidate identity identifier with the first user, and the iris matching degree of each coincident candidate identity identifier with the first user.

[0042] In a possible implementation, the determining module is specifically configured to:

[0043] for any one coincident candidate identity identifier, determine a fusion matching degree of the coincident candidate identity identifier with the first user according to the face matching degree of the coincident candidate identity identifier with the first user, and the iris matching degree of the coincident candidate identity identifier with the first user;

[0044] determine the target identity identifier in the at least one coincident candidate identity identifier according to the fusion matching degree of each repeated candidate identity identifier with the first user.

[0045] In a possible implementation, the determining module is specifically configured to:

[0046] determine, as the target identity identifier, a coincident candidate identity identifier with the highest fusion matching degree of the first user in the at least one coincident candidate identity identifier.

[0047] In a possible implementation, the determining module is specifically configured to:

[0048] obtain a first face feature vector of the first face image;

[0049] obtain a face similarity between the first face feature vector and each face feature vector in a face database;

[0050] determine a plurality of candidate face feature vectors in the face database according to the face similarity between the first face feature vector and each face feature vector in the face database;

[0051] determine the plurality of second candidate identity identifiers as identity identifiers corresponding to the plurality of candidate face feature vectors;

[0052] determine a face matching degree between the second candidate identity identifier and the first user as a similarity between a candidate face feature vector corresponding to the second candidate identity identifier and the first face feature vector.

[0053] In a possible implementation, the determining module is specifically configured to:

[0054] obtain a first iris feature vector of the first iris image;

[0055] obtain an iris similarity between the first iris feature vector and each iris feature vector in an iris database;

[0056] determine a plurality of candidate iris feature vectors in the iris database according to the iris similarity between the first iris feature vector and each iris feature vector in the iris database;

[0057] determine the plurality of second candidate identity identifiers as identity identifiers corresponding to the plurality of candidate iris feature vectors;

[0058] determine an iris matching degree between the second candidate identity identifier and the first user as a similarity between a candidate iris feature vector corresponding to the second candidate identity identifier and the first iris feature vector.

[0059] In a possible implementation, the obtaining module is specifically configured to:

[0060] obtain an initial image of the first user collected by the camera device;

[0061] perform at least one image processing operation on the initial image to obtain the first face image, the at least one image processing operation including cropping processing, rotation processing, or angle adjustment processing.

[0062] a processor and a memory;

[0063] the memory stores computer execution instructions;

[0064] The processor executes computer-executed instructions stored in the memory, so that the electronic device executes the method of any one of the first aspect.

[0065] In a fourth aspect of the present application, a computer-readable storage medium is provided, and the computer-readable storage medium stores computer-executed instructions. When the computer-executed instructions are executed by a processor, the computer-executed instructions are used to implement the determination method of the driver of the hardware peripheral according to any one of the first aspect.

[0066] The embodiment provides an identity recognition method, device, equipment and storage medium. The method comprises the following steps: acquiring a first face image and a first iris image of a first user; determining a plurality of first candidate identity identifiers of the first user and a face matching degree of each first candidate identity identifier according to the first face image; determining a plurality of second candidate identity identifiers of the first user and an iris matching degree of each second candidate identity identifier according to the first iris image; and determining a target identity identifier of the first user in the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers according to the plurality of first candidate identity identifiers, the plurality of second candidate identity identifiers, the face matching degree of each first candidate identity identifier and the iris matching degree of each second candidate identity identifier. The method acquires the face image and the iris image of the first user, simultaneously determines the face matching degree and the iris matching degree, and then determines the target identity identifier, thereby solving the problems of low recognition rate and low security when identity recognition is performed only by using face recognition, providing convenient and easy-to-use services for some users with damaged biological characteristics, and optimizing the customer experience. BRIEF DESCRIPTION OF DRAWINGS

[0067] The accompanying drawings, which are incorporated herein and form part of the specification, illustrate embodiments consistent with the present application and, together with the description, further serve to explain the principles of the application.

[0068] Figure 1 is a structural diagram of an identity recognition method provided by the present application;

[0069] Figure 2 is a flowchart of an identity recognition method provided by an embodiment of the present application Figure 1 ;

[0070] Figure 3 is a flowchart of an identity recognition method provided by an embodiment of the present application Figure 2 ;

[0071] Figure 4 is a flowchart of an identity recognition method provided by an embodiment of the present application Figure 3 ;

[0072] Figure 4 is a flowchart of an identity recognition method provided by an embodiment of the present application Figure 5 ;

[0073] Figure 6 is a large transfer process chart of mobile phone bank provided by the present application;

[0074] Figure 7 is a structural schematic view of an identity recognition device provided by an embodiment of the present application;

[0075] Figure 8 is a structural schematic view of an electronic device provided by an embodiment of the present application.

[0076] The specific embodiments of the present application have been shown in the above-described drawings, and will be described in more detail hereinafter. These drawings and the following description are not intended to limit the scope of the present application in any way, but to explain the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0077] The exemplary embodiments will be described in detail herein with reference to the attached drawings. The same numbers are used in different drawings to represent the same or similar elements. The following detailed description is not intended to limit the scope of the present application, but to explain the principles of the present application as detailed in the appended claims.

[0078] Figure 1 is a structural view of an identity recognition method provided by the present application. As shown in Figure 1 the method of the present application first acquires a face image and an iris image of a user through an image acquisition device, respectively, and then processes the two images, respectively. On one hand, the face image is pre-processed to obtain a normalized image which can be used for feature extraction. On this basis, the normalized face image is extracted to obtain a face feature. Thereafter, face recognition is performed according to the face feature to obtain a face similarity. The face similarity here refers to the similarity between the face feature of the user and the face feature stored in the database in advance. The process of obtaining the iris similarity by processing the user's iris image is the same as the above process, which will not be described herein. Finally, the face similarity and the iris similarity are fused and matched to obtain a fused similarity, and then the identity of the user is determined.

[0079] When identity recognition is performed based on face recognition technology and then a transfer is performed, the recognition rate is prone to decrease or the security is not high. With the growth of age, the appearance also changes to a certain extent, and the recognition rate of the face recognition algorithm is different for different age groups. If the face is injured, made up or plastic surgery, etc., the face features will change greatly, thereby affecting the accuracy of face recognition, and even leading to failure of recognition. With the wide application of cameras in life, the security of personal face features has brought new challenges, and there is a great security problem in large amount transfer by face recognition.

[0080] The present application provides an identity recognition method, which collects a first face image and a first iris image of a first user, simultaneously determines a face matching degree and an iris matching degree, and then determines a target identity, thereby solving the problem of low recognition rate and low security when only face recognition is used for identity recognition, and providing convenient and easy-to-use services for some users with damaged biological features, and optimizing the customer experience.

[0081] The identity recognition method provided by the present application aims to solve the above technical problems of the prior art.

[0082] The technical solutions of the present application and how the technical solutions of the present application solve the above technical problems will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described again in some embodiments. The embodiments of the present application will be described below with reference to the drawings.

[0083] Figure 2 Identity recognition method provided by the present application Figure 1 . As shown in Figure 2 , the method of the present embodiment comprises:

[0084] S201, acquiring a first face image and a first iris image of a first user;

[0085] The execution subject of the present embodiment can be an electronic device, or an identity recognition device provided in the electronic device. Optionally, the identity recognition device can be realized by software, or realized by the combination of software and hardware.

[0086] In this embodiment, the first face image and the first iris image of the first user are generally obtained by different devices. The face image can be obtained by a general camera, while the iris texture is generally small and dark, and the texture is not very clear. Using a general color camera and visible light will cause the light intensity to be too strong due to the sensitivity of the human eye to visible light, and it is difficult to obtain a clear and contrast iris image. Therefore, a special iris image acquisition device must be used, including an infrared optical imaging system, an electronic control unit and appropriate software algorithms. At the same time, the iris image acquisition device is generally a "feedback" device, which requires the user to adjust his position and angle according to the image feedback from the device to the user to adapt to the device to shoot the iris image.

[0087] Optionally, obtaining the first face image of the first user comprises

[0088] Obtaining an initial image of the first user collected by the camera device;

[0089] Performing at least one image processing operation on the initial image to obtain the first face image, the at least one image processing operation comprising: cropping processing, rotation processing or angle adjustment processing.

[0090] In this embodiment, the position of the camera device for obtaining the initial image relative to the first user is random, for example, most of the monitoring camera is above the user, and the user uses the mobile phone to collect the face image, and different users collect different image postures and angles. This leads to that the initial image cannot be well extracted.

[0091] Therefore, after obtaining the initial image of the first user collected by the camera device, at least one image processing operation needs to be performed on the initial image, such as cropping processing, rotation processing or angle adjustment processing. For example, if the initial image includes the face image of the first user and the face image of another user at the same time, cropping processing is needed at this time. When the face image of the first user in the initial image is not at a suitable angle, rotation processing or angle adjustment processing is needed.

[0092] Those skilled in the art can understand that the image processing operation of the initial image of the first user includes but is not limited to cropping processing, rotation processing or angle adjustment processing, and also includes operations such as adjusting image brightness or denoising processing.

[0093] Those skilled in the art can understand that after obtaining the initial image of the iris of the first user, the initial image of the iris also needs to be processed to obtain an image that can obtain the iris feature.

[0094] In this embodiment, the pre-processing of the first face image needs to locate the key points such as eyes, nose, and mouth, and the image normalization needs to consider compensating for the changes in illumination and posture. At the same time, the first iris image normalization needs to consider compensating for the changes in illumination and posture. The iris image pre-processing generally includes live detection, quality evaluation to remove images with poor quality that cannot be identified, iris inner and outer circle positioning and normalization. Among them, the iris positioning refers to determining the positions of the inner circle, the outer circle and the quadratic curve in the image. Among them, the inner circle is the boundary between the iris and the pupil, the outer circle is the boundary between the iris and the sclera, and the quadratic curve is the boundary between the iris and the upper and lower eyelids. The iris image normalization refers to adjusting the size of the iris in the image to a fixed size set by the identification system.

[0095] The iris positioning algorithm based on Hough transform can be used for normalization and positioning of the iris and other pre-processing operations.

[0096] S202, determining a plurality of first candidate identity identifiers of the first user and a face matching degree of each first candidate identity identifier according to the first face image;

[0097] In this embodiment, after the first face image is pre-processed and normalized, face feature extraction is needed. Face feature extraction, also known as face representation, is a process of modeling the features of a face. The face feature extraction methods that can be taken include two categories: one is a knowledge-based representation method; the other is an algebraic feature or statistical learning representation method. Among them, the knowledge-based representation method mainly obtains feature data that is helpful for face classification according to the shape description of face organs and their distance characteristics, and the feature components usually include Euclidean distance, curvature and angle between feature points. The representation method based on algebraic features or statistical learning mainly uses algebraic features to extract face identity information. It can be roughly divided into two categories of linear and nonlinear feature extraction methods.

[0098] After obtaining the face features of the first face image, the face features of the first user need to be matched with the face features in the database. The database pre-acquires the face images of all users, then extracts the face features from the face images and stores them, thereby constructing a feature library. After the face features of the first user are matched with the feature library in the database, a plurality of face features with high matching degree with the face features of the first user and corresponding face matching degrees are obtained, and then the user identifiers corresponding to these face features, i.e., the first candidate identity identifiers, are determined.

[0099] S203, determining a plurality of second candidate identity identifiers of the first user and an iris matching degree of each second candidate identity identifier according to the first iris image;

[0100] In this embodiment, after the first iris image is preprocessed and normalized, iris feature extraction is needed. The iris is mainly a texture feature, and a texture analysis method is usually used to extract the feature. The texture analysis methods that can be used include statistical method, structural method, and spectral method. The statistical method is described by the moments of the gray level histogram or the gray level co-occurrence matrix. The basic idea of the structural method is that a complex texture can be composed of simple textures in a certain rule, that is, it is assumed that the texture pattern is composed of the spatial arrangement of texture primitives, and then a syntax analysis method is used for recognition. The spectral method describes the texture characteristics of a periodic or nearly periodic 2D image by means of the frequency characteristics of Fourier transform.

[0101] After the iris feature of the first iris image is obtained, the iris feature of the first user needs to be matched with the iris features in the database. The database pre-acquires the iris images of all users, then extracts the iris features from the iris images and stores them, and then constructs a feature library. After the iris feature of the first user is matched with the feature library in the database, a plurality of iris features with high matching degrees with the iris feature of the first user and corresponding iris matching degrees are obtained, and then the user identifiers corresponding to these face features, i.e., the second candidate identity identifiers, are determined.

[0102] S204, according to the plurality of first candidate identity identifiers, the plurality of second candidate identity identifiers, the face matching degree of each first candidate identity identifier with the first user, and the iris matching degree of each second candidate identity identifier with the first user, determining the target identity identifier of the first user from the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers.

[0103] In this embodiment, after the plurality of first candidate identity identifiers and the plurality of candidate identity identifiers are determined, when the target identity identifier of the first user is further determined, the face matching degree of each first candidate identity identifier with the first user and the iris matching degree of each second candidate identity identifier with the first user need to be considered. The features of the face and the iris are used to determine the target identity identifier of the first user, which improves the accuracy of the identity recognition result and can make up for the defect that some users with damaged biological characteristics cannot conveniently and quickly perform identity recognition when only the face is used for biological recognition technology.

[0104] The range of the first candidate identity identifier and the second candidate identity identifier can be completely coincident or not completely coincident. When the range of the first candidate identity identifier and the second candidate identity identifier is coincident, for example, the three first candidate identity identifiers are user A, user B, and user C, and the three second candidate identity identifiers are also user A, user B, and user C. At this time, the target identity identifier can be determined according to the face matching degree and the iris matching degree value of the candidate identity identifier.

[0105] The target identity can be determined by using the matching degree fusion method. The fusion at this time belongs to the score layer fusion, and the process includes the following steps: firstly, the matching degrees of various modes are obtained, then they are normalized to the same scale, and finally the matching degrees after fusion are obtained by using the minimum, maximum, mean value and other operations.

[0106] The obtained matching degrees can be regarded as feature vectors, and a classifier is trained based on this to determine the target identity. Commonly used are support vector machines and likelihood ratios. Based on the obtained matching degrees after fusion, the selected identity corresponding to the face matching degree or the iris matching degree with better evaluation is determined as the target identity. In the financial system, the identity of the user after the above operation is recognized, and then more operations similar to large amount transfer can be performed.

[0107] The embodiment provides an identity recognition method. The method comprises the following steps: obtaining a first face image and a first iris image of a first user; determining a plurality of first selected identities of the first user and a face matching degree of each first selected identity with the first user according to the first face image; determining a plurality of second selected identities of the first user and an iris matching degree of each second selected identity with the first user according to the first iris image; and determining a target identity of the first user in the plurality of first selected identities and the plurality of second selected identities according to the plurality of first selected identities, the plurality of second selected identities, the face matching degree of each first selected identity with the first user and the iris matching degree of each second selected identity with the first user. The method simultaneously determines the face matching degree and the iris matching degree by collecting the first face image and the first iris image of the first user, and then determines the target identity, thereby solving the problems of low recognition rate and low security when only using face recognition to recognize the identity, simultaneously providing convenient and easy-to-use services for some users with damaged biological characteristics, and optimizing the customer experience.

[0108] Figure 3 The identity recognition method process provided by the embodiment Figure 2 . As Figure 3 indicated, the method of the embodiment is based on the embodiment as Figure 2 indicated, and a detailed description is provided for the process of determining the target identity of the first user in the plurality of first selected identities and the pluralityity of second selected identities according to the plurality of first selected identities, the plurality of second selected identities, the face matching degree of each first selected identity with the first user and the iris matching degree of each second selected identity with the first user.

[0109] S301, determining at least one coincident selected identity existing in the plurality of first selected identities and the plurality of second selected identities;

[0110] When the ranges of the first candidate identity and the second candidate identity are not completely coincident, at least one coincident candidate identity needs to be determined. For example, three first candidate identities are user A, user B and user C, and three second candidate identities are user A, user D and user F, and the coincident candidate identity is user A.

[0111] On the basis of determining the coincident candidate identity, a target identity is determined from the at least one coincident candidate identity according to a face matching degree and an iris matching degree of each coincident candidate identity and the first user.

[0112] S302, for any one coincident candidate identity, a fusion matching degree of the coincident candidate identity and the first user is determined according to a face matching degree of the coincident candidate identity and the first user and an iris matching degree of the coincident candidate identity and the first user.

[0113] In this embodiment, the fusion matching degree of the first user is determined by using a score layer fusion method.

[0114] The fusion recognition algorithm based on the C-SVC linear kernel function can be used to determine the fusion matching degree of the coincident candidate identity and the first user. The kernel function is used to implicitly map from a low-dimensional space to a high-dimensional space, and the mapping can make two classes of points in the low-dimensional space linearly inseparable.

[0115] The fusion recognition algorithm of kernel canonical correlation analysis (KCCA) can be used to determine the fusion matching degree of the coincident candidate identity and the first user. The KCAA algorithm is a new way to provide a learning method of nonlinearity by mapping the sample set into a high-dimensional feature space and replacing the inner product operation in the feature space with a regenerative kernel defined in advance without increasing the calculation amount, and can accurately and efficiently express the nonlinear relationship contained in the sample.

[0116] It is worth noting that the face matching degree and the iris matching degree of the coincident candidate identity and the first user are obtained respectively, and therefore the matching degrees obtained are similar in distribution type, but not in the same range in value. Therefore, they need to be distributed in the same numerical range before feature fusion. The minimum-maximum normalization processing method is used for linear change, and the formula is as follows:

[0117]

[0118] wherein, S N is the normalized matching score value; S is the score value before processing; S minis the minimum value of the same kind of biological feature score value; S max is the maximum value of the same kind of biological feature score; x represents the lower limit of the transformed interval, and is set to 0 when applied; y represents the upper limit of the transformed interval, and is set to 1 when applied.

[0119] After determining the fusion matching degrees of the coincident candidate identity and the first user, the target identity is determined from the at least one coincident candidate identity according to the fusion matching degrees of each coincident candidate identity and the first user.

[0120] S303, the coincident candidate identity with the highest fusion matching degree of the first user is determined as the target identity from the at least one coincident candidate identity.

[0121] The fusion matching degree of the coincident candidate identity and the first user is generally a vector, including the face matching degree and the iris matching degree after fusion. The coincident candidate identity with the highest fusion matching degree of the first user can be determined in multiple ways.

[0122] For example, there are three coincident candidate identities, user A, user B and user C, and their corresponding fusion matching degrees are (a1, b1), (a2, b2) and (a3, b3), where ai represents the face matching degree after fusion, and bi represents the iris matching degree after fusion. The weight of the face matching degree can be set as w1, and the weight of the iris matching degree can be set as w2. After calculating the weighted sum ai*w1+bi*w2, the coincident candidate identity corresponding to the maximum weighted sum is selected as the target identity.

[0123] The embodiment provides an identity recognition method. The method comprises the following steps: determining at least one coincident candidate identity from a plurality of first candidate identities and a plurality of second candidate identities; determining a fusion matching degree of the coincident candidate identity and a first user according to a face matching degree of the coincident candidate identity and the first user and an iris matching degree of the coincident candidate identity and the first user for any one of the coincident candidate identities; and determining a target identity as the coincident candidate identity with the highest fusion matching degree of the first user from the at least one coincident candidate identity. The method considers that the ranges of the first candidate identities and the second candidate identities are not completely coincident, determines the coincident candidate identity first, and then determines the target identity through fusion, thereby greatly improving the practicability of the method.

[0124] Figure 4 An identity recognition method provided by the embodiment Figure 3 . As Figure 4 shown, the method of the embodiment comprises the following steps: Figure 2Based on the embodiment shown, the process of determining a plurality of first candidate identity identifiers of the first user according to the first face image and the face matching degree of each first candidate identity identifier to the face of the first user is described in detail.

[0125] S401, obtaining a first face feature vector of the first face image;

[0126] In this embodiment, the first face feature vector of the first face image is obtained by using a face feature extraction method. A statistical learning-based representation method can be used. The statistical learning-based representation method is roughly divided into linear and nonlinear two categories, wherein the linear feature extraction method mainly includes principal component analysis (PCA), linear discriminant analysis (LDA) and independent component analysis (ICA) and the like, and the nonlinear feature extraction method includes kernel principal component analysis (KPCA) and locally linear embedding (LLE) and the like.

[0127] The LLE and LDA combined nonlinear dimension reduction algorithm can be used to extract the LDA-LLE feature of the face image.

[0128] S402, obtaining a face similarity between the first face feature vector and each face feature vector in the face database;

[0129] The face similarity between the first face feature vector and each face feature vector in the face database is obtained by using a feature matching recognition method. Feature matching recognition is to input the extracted features into a classification decision maker for final training decision, so as to complete face recognition. The principle is to compare and classify the extracted feature information, if the similarity of two samples is high, it is determined as the same class. The feature matching method that can be used can be roughly divided into distance-based classification method and sample distribution statistical characteristic-based classification method.

[0130] The distance-based classification method can be used to calculate the face similarity between the first face feature vector and each face feature vector in the face database. First, a feature point in the first image is selected, then the two feature points with the closest Euclidean distance to the feature point in all database images are found, and the feature matching rate is used to reflect the similarity calculation formula of the image as follows:

[0131]

[0132] dE represents the Euclidean distance; x, y represent the feature vectors of the face; k represents the dimension of the feature vector.

[0133] S403, determining a plurality of candidate face feature vectors in the face database according to the face similarity between the first face feature vector and each face feature vector in the face database.

[0134] After determining the face similarity between the first face feature vector and each face feature vector in the face database, the face feature vector corresponding to the good face similarity can be selected as the candidate face feature vector.

[0135] The method of setting a threshold value can be adopted to compare the size of the face similarity and the threshold value. When the face similarity is greater than the threshold value, the face feature vector corresponding thereto is the candidate face feature vector.

[0136] Those skilled in the art can understand that the method of determining a plurality of candidate face feature vectors in the face database according to the face similarity between the first face feature vector and each face feature vector in the face database includes but is not limited to the above method.

[0137] S404, determining a plurality of first candidate identity identifiers corresponding to the plurality of candidate face feature vectors.

[0138] In the database, all face feature vectors are stored, and the mapping relationship between the face feature vectors and the corresponding identity identifiers is also stored. Therefore, after determining the plurality of candidate face feature vectors, the identity identifiers corresponding to the plurality of candidate face feature vectors can be queried, and then the plurality of first candidate identity identifiers are determined.

[0139] S405, determining the face matching degree between the first candidate identity identifier and the first user according to the similarity between the candidate face feature vector corresponding to the first candidate identity identifier and the first face feature vector.

[0140] The embodiment provides an identity recognition method. The method comprises the following steps: obtaining a first face feature vector of a first face image; obtaining a face similarity between the first face feature vector and each face feature vector in a face database; determining a plurality of candidate face feature vectors in the face database according to the face similarity between the first face feature vector and each face feature vector in the face database; determining a plurality of first candidate identity identifiers corresponding to the plurality of candidate face feature vectors as the plurality of first candidate identity identifiers; and determining a face matching degree between the first candidate identity identifiers and a face of a first user as a similarity between a candidate face feature vector corresponding to the first candidate identity identifier and the first face feature vector. The method can determine the face matching degree between the first candidate identity identifiers and the face of the first user through feature vector extraction and feature matching recognition operations on the first face image, so that the method has high accuracy and practicability.

[0141] Figure 5 Identity recognition method process provided by the embodiment Figure 4 As shown in Figure 5 The method of the embodiment is based on the embodiment shown in Figure 2 The embodiment provides a detailed description of the process of determining a plurality of second candidate identity identifiers of a first user according to a first iris image and an iris matching degree between each second candidate identity identifier and an iris of the first user.

[0142] S501, obtaining a first iris feature vector of a first iris image;

[0143] In the embodiment, the first iris feature vector of the first face image is obtained by using an iris feature extraction method. The phase two-dimensional Gabor filter method, the wavelet zero-crossing point detection method based on zero-crossing point description, the Gaussian Laplacian pyramid method based on texture analysis and the shape analysis method can be used.

[0144] The Gabor and PCA combined algorithm can be used to extract the feature vector of the iris image.

[0145] S502, obtaining an iris similarity between the first iris feature vector and each iris feature vector in an iris database;

[0146] In the embodiment, the iris similarity between the first iris feature vector and each iris feature vector in the iris database is obtained by using a feature matching recognition method. The Euclidean distance classification method can also be used to calculate the iris similarity between the first iris feature vector and each iris feature vector in the iris database. The process has been described above, and will not be described here.

[0147] S503, determining a plurality of candidate iris feature vectors in the iris database according to the iris similarity between the first iris feature vector and each iris feature vector in the iris database;

[0148] After the iris similarity of the first iris feature vector and each iris feature vector in the iris database is determined, the iris feature vector corresponding to the good iris similarity can be selected as the candidate iris feature vector.

[0149] The threshold setting method can be adopted to compare the size of the iris similarity and the threshold, and the iris feature vector corresponding to the iris similarity greater than the threshold is the candidate iris feature vector.

[0150] S504, the identity identifiers corresponding to the plurality of candidate iris feature vectors are determined as a plurality of second candidate identity identifiers;

[0151] S505, the similarity of the candidate iris feature vector corresponding to the second candidate identity identifier and the first iris feature vector is determined as the iris matching degree of the second candidate identity identifier and the first user.

[0152] The embodiment provides an identity recognition method, which comprises the following steps: acquiring a first iris feature vector of a first iris image; acquiring an iris similarity of the first iris feature vector and each iris feature vector in an iris database; determining a plurality of candidate iris feature vectors in the iris database according to the iris similarity of the first iris feature vector and each iris feature vector in the iris database; determining a plurality of second candidate identity identifiers corresponding to the plurality of candidate iris feature vectors; and determining an iris matching degree of the second candidate identity identifier and the first user according to the similarity of the candidate iris feature vector corresponding to the second candidate identity identifier and the first iris feature vector. The method can extract the feature vector and perform the feature matching recognition operation through the first iris image, and determine the iris matching degree of the second candidate identity identifier and the first user, so that the method has high accuracy and practicability.

[0153] The technical scheme of the present application will be described in detail below with reference to a specific embodiment.

[0154] Figure 6 is a large amount of transfer process chart of the mobile phone bank provided by the present application. As shown in FIG. 1, the process of the mobile phone bank large amount of transfer comprises the following steps: Figure 6The method is illustrated by taking the scenario of identity recognition of a bank using the multi-modal biometric recognition technology to realize large-amount transfer as an example. Before using the method, the following conditions need to be met: firstly, the face information and iris information of the account opening person need to be collected and stored in the face and iris feature database of the bank when the bank card is opened. Secondly, the threshold information of judging the large-amount transfer amount and the threshold information of the face and iris fusion matching degree are set in the bank system. When the user uses the mobile banking APP to perform large-amount transfer, the mobile phone main interface is entered, and the transfer function is clicked. After the user fills in the account number, name and transfer amount of the user to be transferred, the mobile banking APP judges whether the input amount belongs to large-amount transfer. If it is judged that the transfer amount belongs to large-amount transfer and the user agrees, the face image and iris image of the user are collected by the mobile phone camera when the user transfers, and identity verification is performed according to the face image and iris image.

[0155] Figure 7 A structural schematic diagram of an identity recognition device provided by an embodiment of the application is shown. The device of the embodiment can be in the form of software and / or hardware. As shown in the figure, Figure 7 An identity recognition device 700 provided by an embodiment of the application includes an acquisition module 701 and a determination module 702,

[0156] The acquisition module 701 is configured to acquire a first face image and a first iris image of a first user.

[0157] The determination module 702 is configured to determine, according to the first face image, a plurality of first candidate identity identifiers of the first user and a face matching degree of each first candidate identity identifier with the first user.

[0158] The determination module 702 is further configured to determine, according to the first iris image, a plurality of second candidate identity identifiers of the first user and an iris matching degree of each second candidate identity identifier with the first user.

[0159] The determination module 702 is further configured to determine, according to the plurality of first candidate identity identifiers, the plurality of second candidate identity identifiers, the face matching degree of each first candidate identity identifier with the first user and the iris matching degree of each second candidate identity identifier with the first user, a target identity identifier of the first user in the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers.

[0160] In a possible implementation manner, the determination module is specifically configured to:

[0161] determine at least one coincident candidate identity identifier existing in the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers.

[0162] According to the face matching degree of each coincident candidate identity with the first user and the iris matching degree of each coincident candidate identity with the first user, a target identity is determined from the at least one coincident candidate identity.

[0163] In a possible implementation, the determining module is specifically configured to:

[0164] For any one of the coincident candidate identities, according to the face matching degree of the coincident candidate identity with the first user and the iris matching degree of the coincident candidate identity with the first user, a fusion matching degree of the coincident candidate identity with the first user is determined.

[0165] According to the fusion matching degree of each repeated candidate identity with the first user, a target identity is determined from the at least one coincident candidate identity.

[0166] In a possible implementation, the determining module is specifically configured to:

[0167] The coincident candidate identity with the highest fusion matching degree with the first user from the at least one coincident candidate identity is determined as the target identity.

[0168] In a possible implementation, the determining module is specifically configured to:

[0169] A first face feature vector of the first face image is obtained.

[0170] A face similarity between the first face feature vector and each face feature vector in a face database is obtained.

[0171] According to the face similarity between the first face feature vector and each face feature vector in the face database, a plurality of candidate face feature vectors are determined from the face database.

[0172] Identity identifiers corresponding to the plurality of candidate face feature vectors are determined as a plurality of first candidate identities.

[0173] A similarity between the candidate face feature vector corresponding to the first candidate identity and the first face feature vector is determined as a face matching degree of the first candidate identity with the first user.

[0174] In a possible implementation, the determining module is specifically configured to:

[0175] A first iris feature vector of the first iris image is obtained.

[0176] An iris similarity between the first iris feature vector and each iris feature vector in an iris database is obtained.

[0177] According to the iris similarity between the first iris feature vector and each iris feature vector in the iris database, a plurality of candidate iris feature vectors are determined in the iris database;

[0178] The identity identifiers corresponding to the plurality of candidate iris feature vectors are determined as a plurality of second candidate identity identifiers;

[0179] The similarity between the candidate iris feature vector corresponding to the second candidate identity identifier and the first iris feature vector is determined as the iris matching degree between the second candidate identity identifier and the first user.

[0180] In a possible implementation manner, the obtaining module is specifically configured to:

[0181] Obtain an initial image of the first user collected by the camera device;

[0182] At least one image processing operation is performed on the initial image to obtain the first face image, and the at least one image processing operation includes cropping processing, rotation processing or angle adjustment processing.

[0183] The device for identity recognition provided in the embodiment can be used to execute the method embodiments, and has similar implementation principles and technical effects, which will not be described here again.

[0184] An example of the structure of an electronic device is provided in the embodiment, which is shown in FIG. 2. Figure 8 The electronic device 20 can include a processor 21 and a memory 22. Exemplarily, the processor 21, the memory 22, and parts thereof are connected to each other through a bus 23.

[0185] The memory 22 stores computer execution instructions;

[0186] The processor 21 executes the computer execution instructions stored in the memory 22, so that the electronic device executes the identity recognition method as described above.

[0187] It should be appreciated that the processor 21 described above can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the application can be directly embodied as hardware processor execution, or executed by a combination of hardware and software modules in the processor. The memory 22 can include a high-speed random access memory (RAM), and can also include a non-volatile memory (NVM), such as at least one disk memory, and can also be a U disk, a mobile hard disk, a read-only memory, a magnetic disk or an optical disk, etc.

[0188] The embodiments of the present application also provide a computer readable storage medium, which stores computer execution instructions. When the computer execution instructions are executed by a processor, the identity recognition method is implemented.

[0189] It should be noted that the identity recognition method and device of the present disclosure can be used in the financial field. It can also be used in any field other than the financial field. The application field of the identity recognition method and device of the present disclosure is not limited.

[0190] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation portal for user to choose authorization or refusal.

[0191] Other embodiments of the present application will be readily apparent to those skilled in the art upon considering the description of the application and practicing the application as disclosed herein. The present application is intended to cover any variations, uses or adaptive changes of the present application following the general principles of the present application and including common knowledge or conventional technical means in the art which are not disclosed in the present application. The specification and examples are only regarded as exemplary, and the true scope and spirit of the present application are indicated by the following claims.

[0192] It is to be understood that the application is not limited to the precise construction already described above and shown in the drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the application. The scope of the application should only be limited by the claims appended hereto.

Claims

1. An identity recognition method, characterized in that, include: Obtain the first face image and first iris image of the first user; Based on the first facial image, determine multiple first candidate identity identifiers of the first user, and the facial matching degree between each first candidate identity identifier and the first user; Based on the first iris image, a plurality of second candidate identity identifiers for the first user are determined, as well as the matching degree between each second candidate identity identifier and the first user's iris. Determine at least one overlapping candidate identity among the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers; For any overlapping candidate identity identifier, the fusion matching degree between the overlapping candidate identity identifier and the first user is determined based on the face matching degree between the overlapping candidate identity identifier and the first user, and the iris matching degree between the overlapping candidate identity identifier and the first user. Based on the fusion matching degree between each duplicate candidate identity identifier and the first user, the candidate identity identifier with the highest fusion matching degree among the at least one duplicate candidate identity identifiers is determined as the target identity identifier of the first user.

2. The method according to claim 1, characterized in that, Determining multiple first candidate identity identifiers of the first user based on the first facial image, and the facial matching degree between each first candidate identity identifier and the first user, including: Obtain the first facial feature vector of the first face image; Obtain the face similarity between the first face feature vector and each face feature vector in the face database; Based on the face similarity between the first face feature vector and each face feature vector in the face database, multiple candidate face feature vectors are determined in the face database; The identity identifiers corresponding to the plurality of candidate facial feature vectors are determined as the plurality of first candidate identity identifiers; The similarity between the candidate face feature vector corresponding to the first candidate identity and the first face feature vector is determined as the face matching degree between the first candidate identity and the first user.

3. The method according to any one of claims 1-2, characterized in that, Determining multiple second candidate identity identifiers of the first user based on the first iris image, and the matching degree between each second candidate identity identifier and the first user's iris, including: Obtain the first iris feature vector of the first iris image; Obtain the iris similarity between the first iris feature vector and each iris feature vector in the iris database; Based on the iris similarity between the first iris feature vector and each iris feature vector in the iris database, multiple candidate iris feature vectors are determined in the iris database. The identity identifiers corresponding to the plurality of candidate iris feature vectors are determined as the plurality of second candidate identity identifiers; The similarity between the candidate iris feature vector corresponding to the second candidate identity and the first iris feature vector is determined as the iris matching degree between the second candidate identity and the first user.

4. The method according to any one of claims 1-2, characterized in that, Obtain the first user's first facial image, including Acquire the initial image of the first user captured by the camera device; The initial image is subjected to at least one image processing operation to obtain the first face image, wherein the at least one image processing operation includes: cropping, rotation or angle adjustment.

5. An identity recognition device, characterized in that, include: The acquisition module is used to acquire the first face image and the first iris image of the first user; The determination module is used to determine multiple first candidate identity identifiers of the first user and the face matching degree of each first candidate identity identifier with the first user based on the first face image; The determining module is further configured to determine, based on the first iris image, a plurality of second candidate identity identifiers of the first user, and the matching degree between each second candidate identity identifier and the iris of the first user; The determining module is further configured to determine at least one overlapping candidate identity among the plurality of first candidate identity identifiers and the plurality of second candidate identity identifiers; For any overlapping candidate identity identifier, the fusion matching degree between the overlapping candidate identity identifier and the first user is determined based on the face matching degree between the overlapping candidate identity identifier and the first user, and the iris matching degree between the overlapping candidate identity identifier and the first user. Based on the fusion matching degree between each duplicate candidate identity identifier and the first user, the candidate identity identifier with the highest fusion matching degree among the at least one duplicate candidate identity identifiers is determined as the target identity identifier of the first user.

6. An electronic device, comprising: Processor and memory; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the electronic device to perform the method of any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement an identity recognition method as described in any one of claims 1 to 4.

Images