Personnel management system identity recognition method, device and terminal equipment

Abstract

The present application provides a kind of personnel management system identity recognition method, device and terminal equipment, the method comprises: obtaining the face data of target personnel, the face data is input into the preset face recognition model, obtain at least one candidate identity of target personnel;Each candidate identity corresponding local feature enhancement recognition model is obtained, the face data is input into each candidate identity corresponding local feature enhancement recognition model, and the matching probability corresponding to each candidate identity is obtained;Wherein, each candidate identity corresponding local feature enhancement recognition model is based on each candidate identity corresponding local fusion feature to the face recognition model and is obtained by retraining;The final identity of target personnel is determined based on the matching probability corresponding to each candidate identity.The personnel management system identity recognition method, device and terminal equipment provided by the present application can improve the identity recognition accuracy.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of identity recognition, and more particularly relates to a personnel management system identity recognition method, device and terminal equipment. BACKGROUND

[0002] In a personnel management system, personnel identity recognition is crucial, which directly affects the security of system data and the legitimate rights and interests of system personnel. With the development of information technology, facial recognition as an important means of identity recognition has been widely used in various scenarios, so an identity recognition scheme based on facial recognition has emerged in the personnel management system.

[0003] The existing facial recognition is realized based on the matching degree of facial feature points, that is, as long as the matching degree of facial features reaches a certain value, the person being measured is determined as a certain legal identity in the system. On this basis, for people with similar facial contours, the recognition accuracy of this method is low. SUMMARY

[0004] The present application aims to provide a personnel management system identity recognition method, device and terminal equipment to improve the accuracy of identity recognition.

[0005] The first aspect of the embodiment of the present application provides a personnel management system identity recognition method, comprising:

[0006] obtaining facial data of a target person, inputting the facial data into a preset facial recognition model, and obtaining at least one candidate identity of the target person;

[0007] obtaining a local feature enhancement recognition model corresponding to each candidate identity, inputting the facial data into the local feature enhancement recognition model corresponding to each candidate identity, and obtaining a matching probability corresponding to each candidate identity; wherein the local feature enhancement recognition model corresponding to each candidate identity is obtained by retraining the facial recognition model based on the local fusion features corresponding to each candidate identity;

[0008] determining the final identity of the target person based on the matching probability corresponding to each candidate identity.

[0009] In a possible implementation, the method for training a local feature enhancement recognition model corresponding to a certain identity comprises:

[0010] obtaining historical facial data of a target identity, extracting features of the historical facial data of the target identity, and obtaining target features; obtaining historical facial data of a target relative identity, extracting features of the historical facial data of the target relative identity, and obtaining target relative features; wherein the target relative identity is the identity of a person who has a relative relationship with the target identity;

[0011] obtaining a local fusion feature corresponding to the target identity based on the target feature and the target relative feature, and fusing the local fusion feature corresponding to the target identity and the target feature to obtain a fusion feature;

[0012] performing secondary training on the face recognition model based on the fusion feature to obtain a local feature enhanced recognition model corresponding to the target identity.

[0013] In a possible implementation, the obtaining of the local fusion feature corresponding to the target identity based on the target feature and the target relative feature comprises:

[0014] determining at least one local feature recognition part corresponding to the target identity based on a matching degree between the target feature and the target relative feature;

[0015] correspondingly extracting at least one set of local features from historical face data of the target identity based on the at least one local feature recognition part corresponding to the target identity, and correspondingly extracting at least one set of relative local features from historical face data of the target relative identity based on the at least one local feature recognition part corresponding to the target identity;

[0016] fusing the at least one set of local features and the at least one set of relative local features based on the at least one local feature recognition part corresponding to the target identity to obtain at least one set of local fusion features corresponding to the target identity.

[0017] In a possible implementation, the fusing of the local fusion feature corresponding to the target identity and the target feature to obtain a fusion feature comprises:

[0018] replacing features of corresponding parts in the target feature based on the local feature recognition parts corresponding to each local fusion feature in sequence, and taking each replaced target feature as a fusion feature.

[0019] In a possible implementation, the fusing of the at least one set of local features and the at least one set of relative local features based on the at least one local feature recognition part corresponding to the target identity comprises:

[0020] combining the local features and the relative local features corresponding to the same local feature recognition part into a complex vector to obtain at least one set of local fusion features corresponding to the target identity.

[0021] In a possible implementation, the determining of the final identity of the target person based on the matching probability corresponding to each candidate identity comprises:

[0022] If the matching probability corresponding to each candidate identity is greater than the preset probability value, it is determined that the identity of the target person is legal, and the candidate identity with the highest matching probability is taken as the final identity of the target person.

[0023] If the matching probability corresponding to each candidate identity is not greater than the preset probability value, it is determined that the identity of the target person is illegal.

[0024] In a second aspect, the embodiment of the present application provides a personnel management system identity recognition device, which comprises:

[0025] A first recognition module is configured to acquire face data of a target person, input the face data into a preset face recognition model, and obtain at least one candidate identity of the target person.

[0026] A second recognition module is configured to acquire a local feature enhancement recognition model corresponding to each candidate identity, input the face data into the local feature enhancement recognition model corresponding to each candidate identity, and obtain a matching probability corresponding to each candidate identity; wherein the local feature enhancement recognition model corresponding to each candidate identity is obtained by retraining the face recognition model based on the local fusion feature corresponding to each candidate identity.

[0027] An identity determination module is configured to determine a final identity of the target person based on the matching probability corresponding to each candidate identity.

[0028] In a possible implementation, the personnel management system identity recognition device further comprises:

[0029] A model training module is configured to train the local feature enhancement recognition model corresponding to each identity.

[0030] In a third aspect, the embodiment of the present application provides a terminal device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the personnel management system identity recognition method when executing the computer program.

[0031] In a fourth aspect, the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements the steps of the personnel management system identity recognition method when executed by a processor.

[0032] The personnel management system identity recognition method, device and terminal device provided by the embodiment of the present application have the following advantages:

[0033] Different from the scheme of directly matching the degree of facial feature points for identity recognition in the prior art, the application determines the candidate identity of the target personnel based on a preset facial recognition model, and further recognizes the face data according to a local feature enhancement recognition model corresponding to the candidate identity obtained through pre-training, and determines the identity recognition result based on the result of the further recognition. Compared with the prior art, the application performs enhancement processing on the local fusion features based on the facial recognition model, obtains the local feature enhancement recognition model, considers the matching degree of the facial features and the matching degree of the local features, and thus has higher identity recognition accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0035] Figure 1 The flowchart of the identity recognition method of the personnel management system provided by an embodiment of the application is shown.

[0036] Figure 2 The structural block diagram of the identity recognition device of the personnel management system provided by an embodiment of the application is shown.

[0037] Figure 3 The schematic block diagram of the terminal device provided by an embodiment of the application is shown. DETAILED DESCRIPTION

[0038] In the following description, specific details are set forth such as particular system configurations, techniques, etc., in order to provide a thorough understanding of the embodiments of the application. However, it should be apparent to those skilled in the art that the application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the application with unnecessary detail.

[0039] In order to make the objects, technical solutions and advantages of the application clearer, the following will be described with reference to the accompanying drawings through specific embodiments.

[0040] Reference Figure 1 , Figure 1 The flowchart of the identity recognition method of the personnel management system provided by an embodiment of the application is shown. The method comprises:

[0041] S101: Obtain face data of the target person, input the face data into a preset face recognition model, and obtain at least one candidate identity of the target person.

[0042] In the embodiment, the target person is a person to be identified. The embodiment can obtain face data of the target person through an image acquisition device, input the face data into a pre-trained face recognition model, and obtain possible identities of the target person, that is, at least one candidate identity.

[0043] Specifically, the face data can be matched with face data of each identity in a preset database, and identities with a face data matching degree greater than a preset threshold are all candidate identities of the target person.

[0044] Optionally, if there is no identity with a face data matching degree greater than the preset threshold, the target person can be directly determined as an illegal identity.

[0045] S102: Obtain a local feature enhancement recognition model corresponding to each candidate identity, input the face data into the local feature enhancement recognition model corresponding to each candidate identity, and obtain a matching probability corresponding to each candidate identity.

[0046] In the embodiment, the local feature enhancement recognition model corresponding to each candidate identity is obtained by retraining a face recognition model based on a local fusion feature corresponding to each candidate identity.

[0047] In the embodiment, the local feature enhancement recognition model corresponding to each candidate identity is different, and the local feature enhancement recognition model of each identity can be pre-trained and directly obtained when needed.

[0048] S103: Determine a final identity of the target person based on the matching probability corresponding to each candidate identity.

[0049] In a possible implementation manner of the embodiment, determining the final identity of the target person based on the matching probability corresponding to each candidate identity includes:

[0050] If the matching probability corresponding to each candidate identity is greater than a preset probability value, it is determined that the identity of the target person is legal, and the candidate identity with the highest matching probability is taken as the final identity of the target person.

[0051] If the matching probability corresponding to each candidate identity is not greater than the preset probability value, it is determined that the identity of the target person is illegal.

[0052] In the embodiment, the matching probability represents a matching degree of a face data feature and corresponding face data in a preset database.

[0053] From the above, different from the prior art scheme of directly matching the degree of facial feature points for identity recognition, the embodiment of the present application will further recognize the face data according to the pre-trained local feature enhancement recognition model corresponding to the candidate identity after determining the candidate identity of the target personnel based on the preset face recognition model, and determine the identity recognition result based on the re-identified result. Compared with the prior art, the embodiment of the present application performs enhancement processing on the local fusion feature based on the face recognition model, obtains the local feature enhancement recognition model, considers both the matching degree of the facial features and the matching degree of the local features, and thus has higher identity recognition accuracy.

[0054] In a possible implementation, the method for training the local feature enhancement recognition model corresponding to a certain identity includes:

[0055] The historical face data of the target identity is obtained, and the features of the historical face data of the target identity are extracted to obtain target features. The historical face data of the target relative identity is obtained, and the features of the historical face data of the target relative identity are extracted to obtain target relative features. The target relative identity is the identity of a person who has a relative relationship with the target identity.

[0056] The local fusion features corresponding to the target identity are obtained based on the target features and the target relative features, and the local fusion features corresponding to the target identity and the target features are fused to obtain fusion features.

[0057] The face recognition model is retrained based on the fusion features to obtain the local feature enhancement recognition model corresponding to the target identity.

[0058] In the embodiment, the face recognition model is retrained based on the fusion features to obtain the local feature enhancement recognition model corresponding to the target identity, including:

[0059] The weight coefficients of the face recognition model are used as the initial weight coefficients of this training, the fusion features corresponding to the target identity are input into the face recognition model, the weight coefficients of the face recognition model are updated, and the local feature enhancement recognition model corresponding to the target identity is obtained.

[0060] In a possible implementation, the local fusion features corresponding to the target identity are obtained based on the target features and the target relative features, including:

[0061] At least one local feature recognition part corresponding to the target identity is determined based on the matching degree of the target features and the target relative features.

[0062] corresponding to the target identity are extracted from historical face data of the target identity, and at least one set of relative local features corresponding to the at least one local feature recognition part corresponding to the target identity are extracted from historical face data of the target relative identity.

[0063] The at least one set of local features and the at least one set of relative local features are fused based on the at least one local feature recognition part corresponding to the target identity to obtain at least one set of local fusion features corresponding to the target identity.

[0064] In the embodiment, the at least one local feature recognition part corresponding to the target identity is determined based on matching degrees of the target feature and the target relative feature, and includes:

[0065] If the feature matching degree of the target feature and the target relative feature at a certain face part is greater than a preset matching degree, the face part is taken as the local feature recognition part corresponding to the target identity.

[0066] That is, all face parts with a feature matching degree greater than a preset matching degree are taken as the local feature recognition part corresponding to the target identity.

[0067] In a possible implementation, after the local fusion features corresponding to the target identity and the target feature are fused, a fusion feature is obtained, including:

[0068] Features of corresponding parts in the target feature are replaced in sequence based on the local feature recognition part corresponding to each local fusion feature, and each replaced target feature is taken as a fusion feature.

[0069] In a possible implementation, the at least one set of local features and the at least one set of relative local features are fused based on the at least one local feature recognition part corresponding to the target identity, including:

[0070] The local feature and the relative local feature corresponding to the same local feature recognition part are combined into a complex vector to obtain the at least one set of local fusion features corresponding to the target identity.

[0071] In the embodiment, the local feature recognition model is trained based on the features of the target identity and the target relative identity, which can improve the diversity of the features and further improve the recognition accuracy of the local feature recognition model.

[0072] The personnel management system identity recognition method according to the above embodiment, Figure 2 A structural block diagram of a personnel management system identity recognition device according to an embodiment of the present application is provided. For ease of illustration, only parts related to the embodiment of the present application are shown. For reference Figure 2The personnel management system identity recognition device 20 comprises a first recognition module 21, a second recognition module 22, and an identity determination module 23.

[0073] The first recognition module 21 is configured to obtain facial data of a target person, input the facial data into a preset face recognition model, and obtain at least one candidate identity of the target person.

[0074] The second recognition module 22 is configured to obtain a local feature enhanced recognition model corresponding to each candidate identity, input the facial data into the local feature enhanced recognition model corresponding to each candidate identity, and obtain a matching probability corresponding to each candidate identity. The local feature enhanced recognition model corresponding to each candidate identity is obtained by retraining the face recognition model based on a local fusion feature corresponding to each candidate identity.

[0075] The identity determination module 23 is configured to determine a final identity of the target person based on the matching probability corresponding to each candidate identity.

[0076] In a possible implementation, the personnel management system identity recognition device 20 further comprises:

[0077] The model training module 24 is configured to train the local feature enhanced recognition model corresponding to each identity.

[0078] In a possible implementation, the model training module 24 is specifically configured to:

[0079] Obtain historical facial data of a target identity, extract features of the historical facial data of the target identity, and obtain target features. Obtain historical facial data of a target relative identity, extract features of the historical facial data of the target relative identity, and obtain target relative features. The target relative identity is an identity of a person who has a relative relationship with the target identity.

[0080] Fuse the local fusion features corresponding to the target identity and the target features to obtain fusion features.

[0081] Perform secondary training on the face recognition model based on the fusion features to obtain the local feature enhanced recognition model corresponding to the target identity.

[0082] In a possible implementation, the model training module 24 is specifically configured to:

[0083] Determine at least one local feature recognition part corresponding to the target identity based on a matching degree between the target features and the target relative features.

[0084] Based on at least one local feature recognition part corresponding to the target identity, at least one set of local features are extracted from the historical facial data of the target identity. Based on at least one local feature recognition part corresponding to the target identity, at least one set of relative local features are extracted from the historical facial data of the target's relative identity.

[0085] Based on at least one local feature recognition part corresponding to the target identity, at least one set of local features are fused with at least one set of kinship local features to obtain at least one set of local fused features corresponding to the target identity.

[0086] In one possible implementation, the model training module 24 is specifically used for:

[0087] Based on the local feature recognition parts corresponding to each local fusion feature, the features of the corresponding parts in the target feature are replaced in turn, and the replaced target features are used as fusion features.

[0088] In one possible implementation, the model training module 24 is specifically used for:

[0089] By combining local features corresponding to the same local feature recognition location and kinship local features into a complex vector, at least one set of local fusion features corresponding to the target identity can be obtained.

[0090] In one possible implementation, the identity determination module 23 is specifically used for:

[0091] If the matching probability of each candidate identity is greater than the preset probability value, then the identity of the target person is determined to be legitimate, and the candidate identity with the highest matching probability is taken as the final identity of the target person.

[0092] If the matching probability of each candidate identity is not greater than the preset probability value, then the target person's identity is determined to be an illegitimate identity.

[0093] See Figure 3 , Figure 3 This is a schematic block diagram of a terminal device provided according to an embodiment of the present invention. Figure 3 The terminal 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules / units in the above-described device embodiments, such as... Figure 2 The functions of modules 21 to 24 are shown.

[0094] It should be understood that the processor 301 in the embodiments of the present application can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or can also be any conventional processor.

[0095] The input device 302 can include a touchpad, a fingerprint acquisition sensor (for acquiring fingerprint information and direction information of a fingerprint of a user), a microphone, etc., and the output device 303 can include a display (LCD, etc.), a speaker, etc.

[0096] The memory 304 can include a read-only memory and a random access memory, and provide instructions and data for the processor 301. A portion of the memory 304 can also include a non-volatile random access memory. For example, the memory 304 can also store device type information.

[0097] In specific implementations, the processor 301, the input device 302 and the output device 303 described in the embodiments of the present application can execute the implementation manners described in the first embodiment and the second embodiment of the personnel management system identity recognition method provided by the embodiments of the present application, and can also execute the implementation manners of the terminal described in the embodiments of the present application, which will not be described here.

[0098] In another embodiment of the present application, a computer readable storage medium is provided, which stores a computer program. The computer program includes program instructions, which, when executed by a processor, implement all or part of the processes of the above-mentioned embodiments. The computer program can also instruct related hardware to complete the implementation. The computer program can be stored in a computer readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable code, or some intermediate form. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the computer readable medium can include or exclude some contents according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0099] The computer readable storage medium can be an internal storage unit of the terminal, such as a hard disk or a memory of the terminal. The computer readable storage medium can also be an external storage device of the terminal, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the computer readable storage medium can include both the internal storage unit and the external storage device of the terminal. The computer readable storage medium is used to store the computer program and other programs and data required by the terminal. The computer readable storage medium can also be used to temporarily store data that has been output or will be output.

[0100] Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been described in general terms in the above description. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0101] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the terminal and the unit described above can refer to the corresponding process in the foregoing method embodiment, and will not be repeated here.

[0102] In several embodiments provided in the present application, it should be understood that the disclosed terminal and method can be implemented in other ways. For example, the apparatus embodiments described above are merely schematic, for example, the division of units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces or units, and can also be electrical, mechanical or other form of connection.

[0103] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment of the present application.

[0104] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software functional unit.

[0105] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any skilled person in the art can easily think of various equivalent modifications or replacements within the technical range disclosed by the present application, and these modifications or replacements should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method of identity recognition for a people management system, the method comprising: The method comprises the following steps: obtaining facial data of a target person, inputting the facial data into a preset face recognition model, and obtaining at least one candidate identity of the target person; obtaining a local feature enhancement recognition model corresponding to each candidate identity, inputting the facial data into the local feature enhancement recognition model corresponding to each candidate identity, and obtaining a matching probability corresponding to each candidate identity; wherein the local feature enhancement recognition model corresponding to each candidate identity is obtained by retraining the face recognition model based on a local fusion feature corresponding to each candidate identity; determining a final identity of the target person based on the matching probability corresponding to each candidate identity; The method for training the local feature enhancement recognition model corresponding to a certain identity comprises the following steps: obtaining historical facial data of a target identity, extracting features of the historical facial data of the target identity, and obtaining target features; obtaining historical facial data of a target relative identity, extracting features of the historical facial data of the target relative identity, and obtaining target relative features; wherein the target relative identity is an identity of a person having a relative relationship with the target identity; obtaining a local fusion feature corresponding to the target identity based on the target features and the target relative features, and obtaining a fusion feature after fusing the local fusion feature corresponding to the target identity and the target features; performing secondary training on the face recognition model based on the fusion feature, and obtaining the local feature enhancement recognition model corresponding to the target identity; The method for obtaining the local fusion feature corresponding to the target identity based on the target features and the target relative features comprises the following steps: determining at least one local feature recognition part corresponding to the target identity based on a matching degree between the target features and the target relative features; correspondingly extracting at least one set of local features from the historical facial data of the target identity based on the at least one local feature recognition part corresponding to the target identity, and correspondingly extracting at least one set of relative local features from the historical facial data of the target relative identity based on the at least one local feature recognition part corresponding to the target identity; fusing the at least one set of local features and the at least one set of relative local features based on the at least one local feature recognition part corresponding to the target identity, and obtaining at least one set of local fusion features corresponding to the target identity.

2. The personnel management system identification method of claim 1, wherein The method for fusing the local fusion feature corresponding to the target identity and the target features to obtain a fusion feature comprises the following steps: sequentially replacing features of corresponding parts in the target features based on local feature recognition parts corresponding to each local fusion feature, and taking each replaced target feature as a fusion feature.

3. The personnel management system identification method of claim 1, wherein The method for fusing the at least one set of local features and the at least one set of relative local features based on the at least one local feature recognition part corresponding to the target identity comprises the following steps: combining local features and relative local features corresponding to the same local feature recognition part into a complex vector to obtain at least one set of local fusion features corresponding to the target identity.

4. The personnel management system identification method of claim 1, wherein The method for determining the final identity of the target person based on the matching probability corresponding to each candidate identity comprises the following steps: If the matching probability corresponding to each candidate identity is greater than the preset probability value, it is determined that the identity of the target person is legal, and the candidate identity with the highest matching probability is taken as the final identity of the target person. If the matching probability corresponding to each candidate identity is not greater than the preset probability value, it is determined that the identity of the target person is illegal.

5. A personnel management system identification device, characterized by, Comprise: The first identification module is used for obtaining face data of the target person, inputting the face data into a preset face recognition model, and obtaining at least one candidate identity of the target person; The second identification module is used for obtaining a local feature enhanced recognition model corresponding to each candidate identity, inputting the face data into the local feature enhanced recognition model corresponding to each candidate identity, and obtaining a matching probability corresponding to each candidate identity; wherein the local feature enhanced recognition model corresponding to each candidate identity is obtained by retraining the face recognition model based on a local fusion feature corresponding to each candidate identity; the method for training the local feature enhanced recognition model corresponding to a certain identity comprises: obtaining historical face data of a target identity, and extracting features of the historical face data of the target identity to obtain target features; obtaining historical face data of a target relative identity, and extracting features of the historical face data of the target relative identity to obtain target relative features; wherein the target relative identity is an identity of a person who has a relative relationship with the target identity; obtaining a local fusion feature corresponding to the target identity based on the target features and the target relative features, and obtaining a fusion feature after fusing the local fusion feature corresponding to the target identity and the target features; performing secondary training on the face recognition model based on the fusion feature to obtain the local feature enhanced recognition model corresponding to the target identity; the method for obtaining the local fusion feature corresponding to the target identity based on the target features and the target relative features comprises: determining at least one local feature recognition part corresponding to the target identity based on a matching degree between the target features and the target relative features; correspondingly extracting at least one set of local features from the historical face data of the target identity based on the at least one local feature recognition part corresponding to the target identity, and correspondingly extracting at least one set of relative local features from the historical face data of the target relative identity based on the at least one local feature recognition part corresponding to the target identity; fusing the at least one set of local features and the at least one set of relative local features based on the at least one local feature recognition part corresponding to the target identity to obtain at least one set of local fusion features corresponding to the target identity; The identity determination module is used for determining the final identity of the target person based on the matching probability corresponding to each candidate identity.

6. The personnel management system identification apparatus of claim 5, wherein, Further comprise: The model training module is used for training the local feature enhanced recognition model corresponding to each identity.

7. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the steps of the method of any one of claims 1 to 4.

8. A computer-readable storage medium storing a computer program, the computer-readable storage medium comprising: The computer program is executed by the processor to realize the steps of the method of any one of claims 1 to 4.

Images