WO2022174699A1

WO2022174699A1 - Image updating method and apparatus, and electronic device and computer-readable medium

Info

Publication number: WO2022174699A1
Application number: PCT/CN2022/071647
Authority: WO
Inventors: 陶训强; 郭彦东; 何苗
Original assignee: Oppo广东移动通信有限公司
Priority date: 2021-02-22
Filing date: 2022-01-12
Publication date: 2022-08-25
Also published as: CN112818909A

Abstract

The present application relates to the technical field of image processing. Disclosed are an image updating method and apparatus, and an electronic device and a computer-readable medium. The method comprises: acquiring a facial image to be subjected to recognition; matching said facial image with a facial image of a registered user in a database, so as to obtain a matching result; and if the matching result is less than a first threshold value and greater than a second threshold value, taking said facial image as a new facial image, and updating same to the database, wherein when the matching result is greater than the second threshold value, said facial image and the facial image of the registered user correspond to the same user, and when the matching result is less than the first threshold value, there is a difference between the two facial images. Therefore, when it is determined that a facial image to be subjected to recognition belongs to a registered user and has a relatively great difference over same, a facial image of the registered user in a database is updated, such that not only can the recognition performance be improved, but loss caused by frequent writing in the database can also be reduced.

Description

Image updating method, apparatus, electronic device, and computer-readable medium

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the priority of the Chinese Patent Application No. 202110199718.8 and entitled "Image Update Method, Apparatus, Electronic Device and Computer-Readable Medium" filed with the China Patent Office on February 22, 2021, the entire contents of which are by reference Incorporated in this application.

technical field

The present application relates to the technical field of image processing, and more particularly, to an image updating method, apparatus, electronic device, and computer-readable medium.

Background technique

The face recognition system in the prior art generally does not update the registered face image in the face registration database after the registration image is entered for the first time. In the process of using the face recognition system, factors such as lighting, whether to wear makeup, whether to wear glasses/sunglasses, whether to wear a mask, age growth, changes in hairstyles and other factors may cause changes in the user's facial features. It may lead to a decrease in the accuracy of face recognition.

SUMMARY OF THE INVENTION

The present application proposes an image updating method, apparatus, electronic device and computer-readable medium to improve the above-mentioned defects.

In a first aspect, an embodiment of the present application provides an image updating method, including: acquiring a face image to be identified; matching the face image to be identified with a face image of a registered user in a database to obtain a matching Result; if the matching result is less than the first threshold and greater than the second threshold, update the face image to be identified as a new face image to the database, wherein the matching result is greater than the second threshold In this case, the face image to be recognized and the face image of the registered user correspond to the same user, and when the matching result is less than the first threshold, the face image to be recognized corresponds to the registered user's face image. There are differences between face images.

In a second aspect, the embodiments of the present application further provide an image updating apparatus, including: an acquiring unit, a matching unit, and an updating unit. The acquiring unit is used to acquire the face image to be recognized. The matching unit is configured to match the face image to be recognized with the face image of the registered user in the database to obtain a matching result. An update unit, configured to update the face image to be recognized as a new face image to the database if the matching result is less than a first threshold and greater than a second threshold, wherein the matching result is greater than the first In the case of two thresholds, the face image to be recognized corresponds to the same user as the face image of the registered user, and when the matching result is less than the first threshold, the face image to be recognized corresponds to the registered user. There are differences between the user's face images.

In a third aspect, embodiments of the present application further provide an electronic device, one or more processors; a memory; one or more application programs, wherein the one or more application programs are stored in the memory and accessed by Configured to be executed by the one or more processors, the one or more programs are configured to perform the above-described method.

In a fourth aspect, an embodiment of the present application further provides a computer-readable medium, where the readable storage medium stores program code executable by a processor, and when the program code is executed by the processor, causes the processor to Perform the above method.

In a fifth aspect, an embodiment of the present application further provides a computer program product, including a computer program/instruction, wherein the computer program/instruction is executed by a processor to implement the above method.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the drawings that are used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those skilled in the art, other drawings can also be obtained from these drawings without creative effort.

1 shows a schematic diagram of a face recognition system provided by an embodiment of the present application;

FIG. 2 shows a method flowchart of an image updating method provided by an embodiment of the present application;

3 shows a schematic diagram of a face image storage table provided by an embodiment of the present application;

4 shows a schematic diagram of a face image storage table provided by another embodiment of the present application;

5 shows a schematic diagram of a face image storage table provided by another embodiment of the present application;

6 shows a schematic diagram of a face image storage table provided by still another embodiment of the present application;

FIG. 7 shows a method flowchart of an image updating method provided by another embodiment of the present application;

FIG. 8 shows a schematic diagram of prompt information provided by an embodiment of the present application;

Fig. 9 shows the flowchart of S740 in Fig. 7;

FIG. 10 shows a method flowchart of an image updating method provided by another embodiment of the present application;

11 shows a schematic diagram of a face image storage table provided by yet another embodiment of the present application;

12 shows a schematic diagram of a face image storage table provided by another embodiment of the present application;

FIG. 13 shows a method flowchart of an image updating method provided by still another embodiment of the present application;

Fig. 14 shows the flowchart of S1320 in Fig. 13;

15 shows a block diagram of a module of an image updating apparatus provided by an embodiment of the present application;

FIG. 16 shows a module block diagram of an electronic device provided by an embodiment of the present application;

FIG. 17 shows a storage unit for storing or carrying a program code for implementing the image updating method according to the embodiment of the present application according to the embodiment of the present application.

FIG. 18 shows a structural block diagram of a computer program product provided by an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application generally described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations. Thus, the following detailed description of the embodiments of the application provided in the accompanying drawings is not intended to limit the scope of the application as claimed, but is merely representative of selected embodiments of the application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

It should be noted that like numerals and letters refer to like items in the following figures, so once an item is defined in one figure, it does not require further definition and explanation in subsequent figures. Meanwhile, in the description of the present application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

The application of face recognition technology is gradually applied to people's work and life. For example, face images can be collected for payment authentication, unlock authentication, and facial images can be beautified. Through the face recognition technology, the face in the image can be detected, and the face in the image can also be identified to which person's face, so as to identify the user's identity.

As shown in FIG. 1 , FIG. 1 shows a schematic diagram of the face recognition system provided by the present application. The face recognition system includes a server 10 and a user terminal 20, the user terminal 20 and the server 10 are located in a wireless network or a wired network, and the server 10 and the user terminal 20 can perform data interaction, and the data interaction can include face recognition and face images. update etc.

In the embodiment of the present application, the user terminal 20 is provided with a biometric information collection device, which is used to collect the user's biometric information. Wherein, the biometric information collection device can be a fingerprint module, an iris collector and a face collection device, the fingerprint module is used to collect the user's fingerprint information, the iris collector is used to collect the user's iris information, and the face collection device is used for collecting the user's iris information. It is used to collect the user's face image information. As an embodiment, the user terminal 20 may be a mobile user terminal device, for example, may include a smart phone, a tablet computer, an e-book reader, a laptop computer, a vehicle-mounted computer, a wearable mobile user terminal, etc. Payment user terminal with biometric information collection device, etc.

A client terminal may be installed in the user terminal 20. Specifically, the client terminal may be an application program installed in the user terminal 20. The client terminal may acquire the face image collected by the user terminal 20 and send the face image to the server 10. , the server 10 recognizes the face image and feeds back the recognition result to the client. Of course, the client can also record the face image collected by the user terminal 20 into the server 10 to update the face image in the server 10 .

The face image of the registered user is stored in the server 10, and the face image can be used as the identity information of the registered user to verify the identity of the logged-in user in scenarios that require identity verification, such as payment and unlocking of user terminals. That is, the face recognition operation is performed.

Specifically, face recognition generally includes the following two processes: face registration and face verification and recognition. The user usually needs to perform face registration before the first use, that is, the user's face image is entered into the face registration database. After the entry, the corresponding face is recognized by matching the registered face image and the collected face image.

In the process of using the face recognition system, factors such as lighting, whether to wear makeup, whether to wear glasses/sunglasses, whether to wear a mask, age growth, changes in hairstyles and other factors may cause changes in the user's facial features. It may lead to a decrease in the accuracy of face recognition. So the registration image needs to be updated.

The inventor found in the research that the current face update often needs to match the face image to be recognized with the registered face image to obtain a matching score. When the obtained matching score is greater than the preset threshold, the registered image is updated according to different update strategies. Specifically, the current face update has the following two defects: if the preset threshold is set too low, it will lead to misidentification and frequent registration database updates; if the preset threshold is set too high, the input image and the registered image are similar , ignoring the changes, and cannot achieve the purpose of improving the recognition performance by updating the registration template; in addition, when updating the face image in the database, the new face image is directly updated into the database, and then the new face image is updated. The face image of the user may not be the face image of the registered user in the database, that is, there is a misrecognition. Furthermore, the image with the recent registration time should be similar to the real state of the user, and should be given when identifying. The more you think about it; in the end, you need to manually set the number of registered images. Too little may result in loss of performance, and too much will take up too much system resources.

Therefore, in order to improve the above-mentioned defects, the embodiments of the present application provide an image updating method, apparatus, electronic device, and computer-readable medium, which use double thresholds to judge the face image to be recognized, and filter out the matching scores that are too high and too low. The face image to be recognized is updated, and the face image of the registered user in the database is updated, which can not only improve the recognition performance, but also reduce the loss caused by frequent writing to the database.

Specifically, please refer to FIG. 2. FIG. 2 shows an image updating method provided by an embodiment of the present application. The execution body of the updating method may be the above-mentioned server or the above-mentioned user terminal, which is not limited here. . Specifically, the method includes: S201 to S203.

S201: Acquire a face image to be recognized.

In the embodiment of the present application, the scenario of executing the method for updating the face image may be, when the user terminal verifies the user's identity information through the user's face image, specifically, when the user logs in to the client of the user terminal. , input the user's face image through the camera of the user terminal, and the client of the user terminal obtains the user's face image, and performs identity verification on the face image. The update of the face image entered during registration, that is, the face image used for user identity recognition as the face image to be recognized this time.

As another implementation manner, the user can operate the client of the user terminal to make the client perform the face image update operation, that is, the face image update operation can be performed independently without performing the authentication operation. For example, the client provides a face image update function, and the administrator can obtain the face image input by the user, and perform the update operation of the face image based on the face image.

In some embodiments, the face image to be recognized may be a face image obtained by on-site shooting of the user's face area. For example, the user's face image is collected during the above-mentioned user authentication. In other embodiments, the face image to be recognized may also be a pre-stored face image including the user's face area. For example, the user collects the face image of the user through the user terminal, sends the face image to the server, the server stores the face images uploaded by multiple users, and obtains the face image of the user at the moment of updating the face image. The image update operation is performed as the face image to be identified this time, that is, the face image of the user pre-stored in the server is updated based on the face image to be identified.

S202: Match the face image to be recognized with the face image of the registered user in the database to obtain a matching result.

As an implementation manner, the database is arranged in the above-mentioned server, that is, the database is a storage space in the above-mentioned server, and when a user registers, at least one face image will be stored in the server. Specifically, when the user registers at the client of the user terminal, he will enter at least one face image and the user account, and then send the user account and the entered at least one image to the server, and the server is in the database The received at least one image sent by the client is stored corresponding to the user account.

As an implementation manner, the server stores the face images corresponding to different users in the database, such as the face image storage table shown in FIG. 3 , each user account corresponds to the face image corresponding to the user account Store, where "face 11" is used as the image identifier of the face image of user 1.

The format of the face image stored in the database may be a picture format, for example, bmp, jpg, png, tif, gif, pcx, tga, exif, fpx, svg, psd, cdr, pcd, dxf, ufo, eps, ai, raw, WMF, webp and other formats.

As another embodiment, the format of the face images stored in the database may be feature data, that is, the server extracts the feature values of the face images, and stores the feature values of each face image corresponding to the user account. As shown in Fig. 4, "feature 11" is used as the data identifier of the feature data of the face image of user 1, wherein the database stores the original image and feature data of the face image respectively, that is, the picture format and the data format are stored separately , wherein the image identifier corresponds to a face image in a picture format, and the data format corresponds to the feature data of the face image. As shown in FIG. 4 , the user account corresponding to the feature 11 is the user 1. Therefore, the feature data of the face image corresponding to the user account can be conveniently found through the user account and the feature identifier. In some embodiments, the user account can also be stored corresponding to the image identification and data identification of the face image at the same time, and the face image corresponding to the user account and the face can be conveniently found through the user account, image identification and feature identification. The feature data corresponding to the image. By directly storing the feature data, it is convenient to perform comparison, recognition, matching and image processing on face images.

After acquiring the face image to be recognized, the face image to be recognized is matched with the face image of the registered user in the database, wherein the registered user is pre-registered successfully and the face is entered in the database image user. As an implementation manner, the face image to be recognized is matched with the face image of the registered user in the database according to a preset matching algorithm. The preset matching algorithm may be to extract features from the face image to be recognized, and compare the features with the face features extracted from the face images of registered users in the database. Specifically, the Euclidean pattern between the features may be used. The distance, etc., is not limited here, so as to obtain the matching score between the image to be recognized and the face image of the registered user.

It should be noted that the higher the matching score between the face image to be identified and the face image of the registered user, the higher the probability that the user in the face image to be identified and the registered user are the same user. For example, the matching score between the face image to be recognized and the face image of the registered user is normalized to [0, 1], and the closer it is to 1, the more it indicates the same user. As an example, the matching score of the face image to be recognized and the registered user's face image may be expressed as the matching score of the face image to be recognized and the registered user's face image with the highest similarity.

As an implementation manner, if there are multiple face images corresponding to the registered user, the face image to be recognized is matched with each face image of the registered user, and the similarity corresponding to each face image is obtained. The similarity of the face images determines the matching result between the face image to be recognized and the face image of the registered user. Specifically, the similarity between the face image to be recognized and each face image is obtained through the above-mentioned preset algorithm, and the average value of all the similarities may be used as the matching result, or the lowest among all the similarities may be used. or the highest similarity as the matching result.

S203: If the matching result is less than the first threshold and greater than the second threshold, update the face image to be recognized to the database as a new face image.

As an embodiment, when the database includes a plurality of registered users, obtain a matching result corresponding to the face image to be recognized and the face image of each registered user, and search for the matching results that are smaller than the first The matching result whose threshold is greater than the second threshold is regarded as the target matching result, and the registered user corresponding to the target matching result is regarded as the target user.

As an embodiment, when the matching result is greater than the second threshold, and the face image to be identified corresponds to the same user as the face image of the registered user, it can be considered that the user corresponding to the face image to be identified corresponds to the user It is the same as the registered user in the database, that is, the user corresponding to the face image to be recognized is the registered user in the database. Specifically, for the implementation manner of obtaining the matching result, reference may be made to the foregoing steps, which will not be repeated here. In the case where the matching result is less than or equal to the second threshold, the face image to be identified corresponds to a different user from the face image of the registered user, then it can be considered that the user corresponding to the face image to be identified is the same as the user already in the database. The registered users are not the same, that is, the user corresponding to the face image to be recognized is not a registered user in the database.

As another implementation, when the matching result is greater than or equal to the first threshold, it can be considered that the user corresponding to the face image to be recognized is the same as the registered user in the database, and the face image to be recognized is the same as the registered user in the database. The difference between the registered user's face images is very small, that is, it can be considered that there is no difference between the two, that is, the difference can be ignored, or the difference does not meet the update condition, that is, the matching When the result is greater than or equal to the first threshold, it means that the probability that the face image to be recognized and the registered user in the database are the same person is very high, the face image is very similar, and cannot reflect the change of the face image, and the current pending recognition is discarded. The face image of the registered user in the database is not updated based on the face image to be recognized.

If the matching result is less than the first threshold, there is a difference between the face image to be identified and the face image of the registered user, that is, it can be determined that the face image to be identified and the face of the registered user are different The images are not the same image, or the facial features of the two are not the same.

Therefore, when the matching result is less than the first threshold and greater than the second threshold, it can be determined that the face image to be identified corresponds to the same user as the face image of the registered user, and the face image to be identified is the same as the face image of the registered user. There are differences between the face images of the registered users, so the face image to be recognized can be used as a supplementary image of the face image of the registered user in the database, that is, a new face image, and then the The operation of updating the recognized face image to the database as a new face image.

As an embodiment, the embodiment of updating the face image to be recognized to the database as a new face image may be that the face image to be recognized is written into the database, and the database uses the face image to be recognized. The recognized face image is stored corresponding to the target user. Specifically, on the basis of the face image that has been stored correspondingly by the target user, the face image to be recognized can be additionally stored. As shown in FIG. 5 , feature 14 is the data of the feature data of the face image to be recognized. It can be seen that, on the basis of the features 11, 12 and 13 corresponding to the user 1 in the database, the data identification of the feature data of the face image to be recognized, that is, the feature 14 is newly added.

As another implementation, at least one face image in all face images corresponding to the target user in the database may be replaced with the face image to be recognized. Specifically, as shown in FIG. 6 , feature 13 ′ is the feature data after the feature 13 is replaced, and the feature 13' is the feature data of the face image to be recognized. As an embodiment, when the target user corresponds to multiple face images, the target face image can be determined from the multiple face images, and the target face image is updated to the face image to be recognized. As an embodiment, the face image with the lowest matching result among the multiple face images may be used as the target face image, or the face image with the highest matching result among the multiple face images may be used as the target face image .

Therefore, in the above method, the face image to be identified is obtained; the face image to be identified is matched with the face image of the registered user in the database, and a matching result is obtained. The user's face image is updated, for example, adding a new image or modifying an existing image. Specifically, the method of determining whether to update is as follows: if the matching result is less than a first threshold and greater than a second threshold, update the face image to be recognized to the database as a new face image, wherein if If the matching result is greater than the second threshold, it indicates that the face image to be recognized corresponds to the same user as the registered user’s face image, and if the matching result is less than the first threshold, it indicates that the face image to be recognized is the same as the registered user’s face image. There are differences between face images. Therefore, when it is determined according to the matching result that the face image to be recognized belongs to a registered user and is quite different from the face image in the database and meets the requirements for updating, the existing face image in the database is determined. The face image update of registered users can not only improve the recognition performance, but also reduce the loss caused by frequent writing to the database.

In addition, the image updating method provided by the embodiments of the present application can reduce misrecognition and improve recognition stability and performance, in addition to using double thresholds to improve the recognition performance and reduce the loss caused by frequent writing to the database. 7 , which shows an image updating method provided by an embodiment of the present application. The execution body of the updating method may be the above-mentioned server or the above-mentioned user terminal, which is not limited here. Specifically, the method includes: S710 to S740.

S710: Acquire a face image to be recognized.

S720: Match the face image to be recognized with the face image of the registered user in the database to obtain a matching result.

For the implementation of S710 and S720, reference may be made to the foregoing embodiments, and details are not described herein again.

S730: If the matching result is less than the first threshold and greater than the second threshold, obtain the first feature data of the face image to be recognized and the second feature data of the face image of the registered user.

As an implementation manner, the execution body of this embodiment of the present application may be a server, and when the server determines that the matching result is less than the first threshold and greater than the second threshold, the server may directly execute the face image to be recognized as a new face image The operation of updating to the database, for example, directly executes S730. As another implementation, the server feeds back prompt information to the user terminal after determining that the matching result is less than the first threshold and greater than the second threshold, and the user terminal displays the prompt information in the designated interface. When the confirmation instruction is performed, the operation of updating the face image to be recognized to the database as a new face image or the operation of S730 is performed based on the confirmation instruction.

In some embodiments, the designated interface may be a face recognition interface, and specifically, the face recognition interface may be an interface corresponding to a payment operation or an unlock operation, which is not limited herein. The payment operation may be an operation performed by the user on the payment module of the application, and the unlocking operation may be an operation performed when the user unlocks the user terminal. Therefore, the designated interface may be a payment interface or an unlocking interface displayed by the user terminal.

As shown in FIG. 8 , a prompt box 810 is displayed on the designation interface 80, and “whether to allow updating the face image database” is displayed in the prompt box 810, that is, as shown in FIG. 8, the prompt information can be displayed in the figure. "Whether to allow updating face image library". In scenarios such as payment or unlocking of the terminal, the user performs face recognition in the designated interface, and the designated interface 80 displays prompt information when the face recognition passes and the matching result satisfies less than the first threshold and greater than the second threshold. The user inputs a confirmation instruction in the prompt message to instruct the server to perform the operation of updating the face image to be recognized to the database as a new face image or the operation of S730. As an implementation manner, the input method of the confirmation instruction may be operating a confirmation button on the designated interface, or may be input by voice, or may be input by facial expression.

As an embodiment, a confirmation button 811 and a reject button 812 are displayed in the prompt box 810. The user can input a confirmation instruction through the confirmation button 811, or can input a rejection instruction through the reject button 812. When the server obtains the rejection instruction , do not perform the operation of updating the face image to be identified as a new face image to the database or the operation of S730, but perform the operation of the application function corresponding to the designated interface after the face recognition is successful, such as , if the specified interface is a payment interface, the subsequent operation is to display the payment success interface or display the payment order, return to the interface before payment, etc. If the specified interface is an unlock interface, then the subsequent operation is to display the system desktop or The interface displayed on the screen before the user terminal locks the screen this time.

As another implementation manner, after the prompt information is displayed in the designated interface, it is not necessary to display the confirmation button 811 and the reject button 812 in the interface, and then acquire a plurality of people collected by the user terminal within the first preset time period Face image, based on the collected multiple face images to determine the user's facial motion information, the facial motion information may include dynamic changes of eyes, dynamic changes of mouth corners, and dynamic changes of ears, etc. It is determined whether the facial motion information is the first facial motion, and if so, it is determined that a confirmation instruction is received, and if the first facial motion is not received, it is determined that a rejection instruction is received. As an embodiment, a face image may be collected in real time within a first preset time period, and facial motion information may be determined in real time according to the collected multiple face images, if the facial motion information is determined to be the first facial motion, then It is determined that a confirmation instruction has been received, and if the facial motion information is the second facial motion, it is determined that a rejection instruction has been received. If the first facial motion has not been detected at the end of the first preset time period, it can be determined that Deny the order. The first facial action and the second facial action may be two different expressions determined by blinking, moving the pupil from the first position to the second position, raising the corner of the mouth, opening and closing the mouth, etc. The position is the leftmost position of the eye, and the second position is the rightmost position of the eye, which is not limited here. Illustratively, the first facial motion and the second facial motion may be opposite motions, eg, the first facial motion is mouth opening and the second facial motion is mouth closing. By setting the first facial action and the second facial action to be opposite actions, it is easier for the user to remember the first facial action and the second facial action, and the operation complexity is reduced.

As another embodiment, after the prompt information is displayed in the designated interface, it is not necessary to display the confirmation button 811 and the reject button 812 in the interface, and then the user terminal collected by the user terminal in the second preset time period is acquired. Limb images, determine the user's body motion information based on the collected multiple user body images, where the body motion information may include the actions of the user's hands, arms, etc. angle range, etc. It is determined whether the body motion information is the first body motion, if so, it is determined that a confirmation instruction is received, and if the first body motion is not received, it is determined that a rejection instruction is received. As an embodiment, a face image may be collected in real time within a second preset time period, and facial motion information may be determined in real time according to the collected multiple face images, and if it is determined that the body motion information is the first body motion If the confirmation instruction is received, if the body motion information is the second body motion, it is determined that a rejection instruction is received. If the first body motion is still not detected at the end of the second preset time period, it can be determined that the rejection instruction is received. . Wherein, the first body movement and the second body movement may be two different movements determined from actions such as clenching a fist, clenching a fist and extending it, turning a hand, or other gestures made by the hand, which are not limited herein.

As a further embodiment, the above-mentioned facial motion information and body motion information may also be used in combination. Specifically, when it is determined that the facial motion information is the first facial motion and the body motion information is the first body motion, It is determined that a confirmation instruction is received, and a rejection instruction is determined to be received when it is determined that at least one of the situations in which the facial motion information is not the first facial motion and the body motion information is not the first body motion occurs. Specifically, for the determination of the facial motion information and the body motion information, reference may be made to the foregoing embodiments, which will not be repeated here.

As another embodiment, when the confirmation button 811 and the reject button 812 are displayed in the designated interface, the designated action information input by the user can be detected at the same time as the user's response to the confirmation button 811 and the reject button 812, and the designated action can be detected. The information can be at least one of the above-mentioned facial motion information and body motion information. Specifically, when it is detected that the user triggers the confirmation button 811 and the specified motion information satisfies the specified conditions, it is determined that a confirmation instruction is received, and when the user is detected When the reject button 812 is triggered and the designated action information does not satisfy the designated condition, it is determined that a reject instruction has been received. Wherein, the specified action information satisfies the specified condition may be that the body action information is a first body action or if the facial action information is at least one of the first facial actions, specifically, please refer to the foregoing embodiment, which is not repeated here. Repeat.

As an implementation manner, the implementation of acquiring the first feature data of the face image to be recognized and the second feature data of the registered user's face image may be: processing the person to be recognized by a preset feature extraction algorithm The face image obtains the first feature data. If the feature data of the face image of the registered user is pre-stored in the database, the second feature data of the face image of the registered user can be searched in advance from the database. If there is feature data of the face image of the registered user, the second feature data is obtained by processing the face image of the registered user through a preset feature extraction algorithm. The preset algorithm may be a traditional machine learning method, or a deep learning method based on a CNN (Convolutional Neural Network, convolutional neural network) model, etc., which is not limited here.

S740: Acquire feature data of a new face image of a registered user according to the first feature data and the second feature data, and update the new feature data to the database.

It should be noted that the first feature data can be directly used as the feature data of the new face image, and the new feature data can be updated to the database, and the step of S740 can also be performed, wherein the features of the new face image The data can be named as third feature data, then the third feature data is related to the first feature data and the second feature data, so that the third feature data can retain the first feature data and the second feature data at the same time. Therefore, after the new feature data is updated to the database, the face image corresponding to the registered user not only has the feature of the face image to be recognized, but also has the feature of the face image that has been stored in advance, which can To avoid the problem of misrecognition, the face image to be recognized is directly used as the new face image, resulting in an error in the new face image corresponding to the registered user.

Wherein, the first feature data and the second feature data may be fused to obtain the third feature by weighted summation, multiplication, subtraction or convolution of the first feature data and the second feature data by a preset method. data.

As an implementation manner, the third feature data may be obtained by means of weighted fusion. Specifically, as shown in FIG. 9 , the implementation of S740 may include S741 to S743.

S741: Acquire a first weight value of the first feature data and a second weight value of the second feature data.

As an implementation manner, the first weight value and the second weight value can be set according to experience and user requirements. The sum of the first weight value and the second weight value is 1. For example, the first weight is α, and the second weight value is 1-α.

S742: Obtain third characteristic data according to the first weight value, the first characteristic data, the second weight value and the second characteristic data.

As an embodiment, the first reference data is determined according to the first weight value and the first feature data, the second reference data is obtained according to the second weight value and the second feature data, and the first reference data is obtained according to the first reference data and the second reference data. Three characteristic data. Specifically, an implementation manner of determining the first reference data according to the first weight value and the first feature data may be to multiply the first weight value and the first feature data to obtain the first reference data. Similarly, the second reference data is obtained by multiplying the second weight value and the second characteristic data. Then, the first reference data and the second reference data are added to obtain third characteristic data.

S743: Use the third feature data as feature data of a new face image of the registered user.

As an implementation manner, if there are multiple face images corresponding to the registered users in the database, all face images corresponding to the registered users can be obtained; Obtaining the second feature data means obtaining the second feature data according to all face images of the registered user.

As an embodiment, an embodiment of processing the feature data of all the face images may be, when obtaining the matching result between the face image to be identified and each face image of the registered user in the database, that is, the matching score When , find the feature data of the face image with the highest matching score, and use the feature data as the second feature data, so that the face image corresponding to the second feature data can maintain the most similar characteristics to the face image to be recognized, It is avoided that the setting of the second threshold is unreasonable, such as being too low, so that the feature data of the face images of different users is used as the face images of the registered user.

As another implementation manner, the implementation manner of processing the characteristic data of all the face images may be, when obtaining the matching result between the face image to be recognized and each face image of the registered user in the database, that is, matching When the score is obtained, the feature data of the face image with the lowest matching score is searched, and the feature data is used as the second feature data, so that the face image corresponding to the second feature data can be kept the most different from the face image to be recognized. feature, so that the face image of the registered user in the updated database can better reflect the real face characteristics of the user, and it can also avoid the unreasonable setting of the first threshold, for example, if it is too large, which leads to an excessively high matching result. The face image has little difference with the face image in the database in essence, and using the face image with the lowest matching score can improve the difference, so that the updated face image can better complement the face image in the database It improves the comprehensiveness of face images in the data.

As yet another embodiment, it is also possible to perform an averaging process on the feature data of all the face images to obtain the second feature data, that is, to average the feature data of all the face images to obtain the second feature data, so as to obtain the second feature data. The second characteristic data can retain the characteristics of each face image of the registered user, so that the face image corresponding to the second characteristic data can maintain the average characteristics of the original registered pictures, which is equivalent to the average face of the user.

It should be noted that, for the content not described in detail in the foregoing steps, reference may be made to the foregoing embodiments, and details are not described herein again.

Therefore, the embodiment of the present application improves the update strategy, and always retains the image that is determined to be the first registration of the registrant. Then features are extracted from the image to be recognized and the registered face image set of the target registered object, and feature fusion is performed. In addition, all feature data are stored and updated, which can reduce the amount of calculation.

In addition, the image update method provided by the embodiment of the present application, in addition to using double thresholds to improve the recognition performance and reduce the loss caused by frequent writing to the database, it can further reduce the time of collecting the face image to be recognized. Frequent writes to the database. Specifically, please refer to FIG. 10. FIG. 10 shows an image updating method provided by an embodiment of the present application. The execution body of the updating method may be the above-mentioned server or the above-mentioned user terminal, which is not limited here. . Specifically, the method includes: S1001 to S1005.

S1001: Acquire a face image to be recognized.

S1002: Match the face image to be recognized with the face image of the registered user in the database to obtain a matching result.

For the implementation of S1001 and S1002, reference may be made to the foregoing embodiments, and details are not described herein again.

S1003: If the matching result is less than the first threshold and greater than the second threshold, obtain the first time of the last updated face image of the registered user in the database.

As an embodiment, the database not only stores the face images corresponding to the registered users, but also stores the time points of each entry or update of the face images. As shown in FIG. 11 , as an embodiment, the database The data identifiers of the feature data of the face image can also be numbered in the order of update time. For example, the time sequence of the update operations corresponding to feature 11, feature 12, and feature 13 is later in order, that is, the larger the number of the data identifier, the corresponding The later the update time of , so that the latest operation of updating the face image can be quickly found according to the data identification, and then the update time corresponding to the operation can be obtained. For example, the time of the latest update of the face image corresponding to user 1 is 2020.03.25, which is March 25, 2020. To find the update time of the latest and sequentially updated faces by means of data identification, compared with directly determining the sequence of each update time, the search rate is faster and the effect is higher. Of course, it is also possible to directly determine the sequence of each update time, so as to find the time of the latest update of the face image.

S1004: Determine the acquisition time of the face image to be recognized as the second time.

As an implementation manner, the system time when the face image to be recognized is acquired may be used as the second time.

S1005: If the time interval between the first time and the second time is greater than a specified threshold, update the face image to be recognized to the database as a new face image.

Wherein, the specified threshold can be set according to experience, for example, it can be 3-60 days, for example, it can be 15 days. If the time interval between the first time and the second time is greater than the specified threshold, it indicates that the length of time between the face image to be recognized and the face image updated in the latest update of the registered user in the database is relatively large. In this case, the difference between the face image to be recognized and the face image of the registered user in the database will be relatively large, for example, the user does not cut his hair or shave his beard for a long time.

As an embodiment, as shown in FIG. 12 , the face image corresponding to the feature 22 is a new face image, that is, the face image to be recognized, and the corresponding update time is 2021.01.28, as an embodiment , the update time may be the above-mentioned second time. As another implementation manner, the server may need to determine whether the data volume of the face image to be written is greater than the specified data volume, and if so, write the face image to be recognized into the database. For the number of writes, you can wait until the total amount of face images to be written by multiple users is greater than the specified amount of data, and then perform the write operation. Then the update time may be the time when the face image to be recognized is written into the database.

Therefore, in the embodiment of the present application, when the time interval between the face image to be recognized and the face image of the registered user in the database is relatively large, the registered user's person in the database can be compared. Face image update avoids frequent update of face images in the database, which causes high hardware loss.

In addition, in the image updating method provided by the embodiment of the present application, the determination of the first threshold and the second threshold may be set by the user according to experience, or may be a fixed value preset by the server, or may be determined by using sample data , In addition, it is also possible to judge the quality of the face image to be recognized before performing the update, and only perform the update when the quality is relatively high. Specifically, please refer to FIG. 13. FIG. 13 shows an image updating method provided by an embodiment of the present application. The execution body of the updating method may be the above-mentioned server or the above-mentioned user terminal, which is not limited here. . Specifically, the method includes: S1310 to S1350.

S1310: Acquire a face image to be recognized.

S1320: Determine a first threshold and a second threshold based on the plurality of sample images.

As an embodiment, the sample image may be a pre-collected face image, and by matching the face image with the face image of the registered user in the database, it can be verified whether the settings of the first threshold and the second threshold are reasonable. 14, the S1320 includes S1321 to S1324.

S1321: Acquire multiple sample images.

The plurality of sample images include a first image marked as a positive example and a second image marked as a negative example, wherein the positive example indicates that the sample image matches the face image in the database, and the negative example indicates that the sample image matches the person in the database. Face images do not match. As an embodiment, the collection scene of the sample image is similar to the registration scene when the user registers the face image in the database, for example, it may be of the same type or under the same environmental parameters, and the environmental parameters may include lighting, weather, time period etc., the type may include the type of environment, eg, the type of scene, eg, indoor, or mall, office, bedroom, street, etc.

S1322: Obtain a matching result between each of the sample images and the face images of the registered users in the database.

Match the first image marked as a positive example with the face image of the registered user in the database to obtain multiple matching results, and match the second image marked as a negative example with the face of the registered user in the database Image matching to get multiple matching results.

As an implementation manner, four matching results are defined in all the matching results, namely the first result, the second result, the third result and the fourth result. Wherein, the first result indicates that the matching result in the first image is a positive example, that is, after obtaining multiple matches obtained by matching the first image marked as a positive example with the face image of the registered user in the database Among the results, find the matching result whose matching result is greater than the specified threshold, and record it as the first result, and record the matching result less than or equal to the specified threshold as the second result, and the second result indicates that the matching result in the first image is a negative example. The matching result greater than the specified threshold indicates that the image and the face image of the registered user correspond to the same user, that is, it indicates that the user corresponding to the first image exists in the database.

Assuming that the first result is TP (True positive), it is used to indicate that the user corresponding to the image is predicted to exist in the database, and is marked as a positive example in the sample image, that is, in a real situation, the user corresponding to the image actually exists in the In-database, that is, the predicted results are the same as the real results, that is, the real positives are predicted as positives. Assuming that the second result is FP (False positive), it is used to represent that the user corresponding to the image is predicted not to exist in the database, and is marked as a positive example in the sample image, that is, in a real situation, the user corresponding to the image really exists In the database, the predicted results are not the same as the real results, that is, the real positive examples are predicted as negative examples.

In the obtained multiple matching results obtained by matching the second image marked as a negative example with the face image of the registered user in the database, find the matching result whose matching result is greater than the specified threshold, and record it as the third result, which is less than The matching result that is equal to or equal to the specified threshold is denoted as the fourth result, and the fourth result indicates that the matching result in the second image is a negative example.

Assuming that the third result is FN (False negative), it is used to represent that the user corresponding to the image is predicted to exist in the database, and is marked as a negative example in the sample image, that is, in the real situation, the user corresponding to the image does not actually exist in the database. In-database, that is, the predicted results are not the same as the real results, that is, the real negative examples are predicted as positive examples. Assuming that the fourth result is TN (True negative), it is used to indicate that the user corresponding to the image is predicted not to exist in the database, and is marked as a negative example in the sample image, that is, in the real situation, the user corresponding to the image does not really exist In the database, that is, the predicted results are the same as the real results, that is, the real counterexamples are predicted as counterexamples.

S1323: Obtain a first ratio and a second ratio based on the matching result of each of the sample images.

The first ratio is used to represent the ratio of the number of positive matching results in the first image to the number of positive matching results in all sample images, and the second ratio is used to represent the matching results in the second image: The ratio of the number of positive examples to the number of negative examples in all sample images.

Specifically, the results in which the matching results in all sample images are positive examples may include the first result and the third result, and the number of matching results in all sample images that are positive examples is the sum of the numbers of the first and third results , then the first ratio may be, TPR=TP/(TP+FN), where TPR (True Positive Rate) represents the first ratio. The results in which the matching results in all sample images are negative examples can include the second result and the fourth result, then the number of matching results in all sample images that are negative examples is the sum of the numbers of the second and fourth results, then the second ratio It can be, FPR=FP/(TN+FP), wherein, FPR (False Positive Rate) represents the second ratio.

S1324: Set the first threshold and the second threshold based on the first ratio and the second ratio.

As an embodiment, the first ratio can reflect the ratio of the predicted positive examples and the real situation to be positive examples, accounting for the ratio of all the positive examples in the real situation, and the second ratio can reflect the predicted positive examples but the real situation is a negative example, accounting for all the positive examples. The ratio of counterexamples in the real case. Both are affected by the specified threshold, and the first threshold and the second threshold can be flexibly selected according to the actual task requirements for FPR and TPR. For example, the specified thresholds obtained under different task requirements are used as the first threshold and the second threshold, respectively.

Specifically, acquiring the first index and the second index of the first ratio and the third index of the second ratio, and determining the first threshold based on the first index of the first ratio and the third index of the second ratio; A second threshold is determined based on the second index of the first ratio and the third index of the second ratio.

As an embodiment, the second ratio can reflect the error rate caused by the setting of the threshold. Therefore, under the condition that a certain error rate requirement is met, the first index and the second index of the first ratio are determined, and then according to the first index and the second index of the first ratio are determined. Different metrics of a ratio adjust the first threshold and the second threshold.

Specifically, the first threshold is determined based on the first index of the first ratio and the third index of the second ratio. As an embodiment, the above specified threshold is used as the first threshold, and when the second ratio satisfies the third index, the first threshold is set so that the first ratio satisfies the first index. For example, the third instruction is that the second ratio is not greater than 10 ^-4 , specifically, the third indicator is that the second ratio is 10 ^-4 , and the first indicator of the first ratio is TPR>99%, that is, at FPR In the case of 10 ⁻⁴ , the first threshold is set so that TPR>99%.

Similarly, the second threshold may be determined based on the second index of the first ratio and the third index of the second ratio. As an embodiment, the above specified threshold is used as the second threshold, and when the second ratio satisfies the third index, the second threshold is set so that the first ratio satisfies the second index. For example, the third instruction is that the second ratio is not greater than 10 ^-4 , specifically, the third indicator is that the second ratio is 10 ^-4 , and the second indicator of the first ratio is TPR>95%, that is, at FPR In the case of 10 ⁻⁴ , the second threshold is set so that TPR>95%.

S1330: Perform image quality evaluation on the face image to be recognized to obtain an image quality evaluation result.

S1340: If the image quality evaluation result satisfies the preset condition, match the face image to be recognized with the face image of the registered user in the database to obtain a matching result.

As an embodiment, comprehensive grade classification is performed according to one or more factors such as occlusion, blur, illumination, expression, Euler angle, etc. of the face image to be recognized, such as low quality, average quality, and high quality. For example, it can be determined that A rating value, and the rating value is used as the above evaluation result. The rating value can be a continuous numerical distribution, for example, 1-100. Then, multiple intervals are set, and each interval corresponds to a quality level, so that the identification to be identified can be determined. The quality level of the face image. Then, an implementation manner in which the image quality evaluation result satisfies the preset condition may be that the quality level of the face image to be recognized is higher than a specified level.

As an implementation manner, the execution data of S1320 to S1340 may not be limited to the sequence shown in FIG. 13 , and may also be after S1310 , performing image quality evaluation on the to-be-recognized face image to obtain an image quality evaluation result , if the image quality evaluation result satisfies the preset conditions, determine the first threshold and the second threshold based on a plurality of sample images, and then match the face image to be recognized with the face image of the registered user in the database , to obtain a matching result, as long as the first threshold and the second threshold are determined based on a plurality of sample images before performing the matching between the face image to be recognized and the registered user's face image in the database.

S1350: If the matching result is less than the first threshold and greater than the second threshold, update the face image to be recognized to the database as a new face image.

Please refer to FIG. 15 , which shows a structural block diagram of an image updating apparatus 1500 provided by an embodiment of the present application. The apparatus may include: an acquiring unit 1501 , a matching unit 1502 , and an updating unit 1503 .

The obtaining unit 1501 is configured to obtain a face image to be recognized.

The matching unit 1502 is configured to match the face image to be recognized with the face image of the registered user in the database to obtain a matching result.

Further, the matching unit 1502 is further configured to evaluate the image quality of the face image to be identified, so as to obtain an image quality evaluation result; if the image quality evaluation result satisfies a preset condition, the face image to be identified The face image is matched with the face image of the registered user in the database, and the matching result is obtained.

The updating unit 1503 is configured to update the face image to be recognized as a new face image to the database if the matching result is less than the first threshold and greater than the second threshold, wherein the matching result is greater than In the case of the second threshold, the face image to be recognized corresponds to the same user as the face image of the registered user, and in the case that the matching result is less than the first threshold, the face image to be recognized is the same as that of the registered user. There are differences between the face images of registered users.

Further, the updating unit 1503 is also used to obtain the first feature data of the face image to be recognized and the second feature data of the registered user's face image; according to the first feature data and the second feature The data acquires the feature data of the new face image of the registered user, and updates the new feature data to the database.

Further, the updating unit 1503 is further configured to obtain the first weight value of the first characteristic data and the second weight value of the second characteristic data; according to the first weight value, the first characteristic data, the second weight value and the second feature data to obtain the third feature data; the third feature data is used as the feature data of the new face image of the registered user.

Further, the updating unit 1503 is further configured to acquire all face images corresponding to the registered user; and process the feature data of all the face images to obtain second feature data.

Further, the updating unit 1503 is further configured to perform averaging processing on the feature data of all the face images to obtain second feature data.

Further, the updating unit 1503 is also used to obtain the first time of the last updated face image of the registered user in the database; determine the obtaining time of the face image to be recognized as the second time; The time interval between a first time and the second time is greater than a specified threshold, and the face image to be recognized is updated to the database as a new face image.

Further, the image updating apparatus 1500 further includes a setting unit for acquiring a plurality of sample images, wherein the plurality of sample images include a first image marked as a positive example and a second image marked as a negative example, wherein the positive The example indicates that the sample image matches the face image of the database, and the negative example indicates that the sample image does not match the face image of the database; obtain the matching result between each of the sample images and the face image of the registered user in the database Obtain the first ratio and the second ratio based on the matching result of each described sample image, wherein, the first ratio is used to characterize that the matching result in the first image is the number of positive examples and the matching result in all sample images is positive The ratio of the number of examples, the second ratio is used to characterize the ratio of the number of positive examples in the second image to the number of negative examples in all sample images; based on the first ratio and the second A ratio sets the first threshold and the second threshold.

Further, the setting unit is further configured to obtain the first index and the second index of the first ratio and the third index of the second ratio; based on the first index of the first ratio and the third index of the second ratio, determining a first threshold; determining a second threshold based on the second index of the first ratio and the third index of the second ratio.

Those skilled in the art can clearly understand that, for the convenience and brevity of description, for the specific working process of the above-described devices and modules, reference may be made to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

In several embodiments provided in this application, the coupling between the modules may be electrical, mechanical or other forms of coupling.

In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically alone, or two or more modules may be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware, and can also be implemented in the form of software function modules.

Please refer to FIG. 16 , which shows a structural block diagram of an electronic device provided by an embodiment of the present application. The electronic device 100 may be the above-mentioned user terminal or server. The electronic device 100 in the present application may include one or more of the following components: a processor 110, a memory 120, and one or more application programs, wherein the one or more application programs may be stored in the memory 120 and configured to be executed by One or more processors 110 execute, and one or more programs are configured to execute the methods described in the foregoing method embodiments.

The processor 110 may include one or more processing cores. The processor 110 uses various interfaces and lines to connect various parts of the entire electronic device 100, and executes by running or executing the instructions, programs, code sets or instruction sets stored in the memory 120, and calling the data stored in the memory 120. Various functions of the electronic device 100 and processing data. Optionally, the processor 110 may adopt at least one of a digital signal processing (Digital Signal Processing, DSP), a Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA), and a Programmable Logic Array (Programmable Logic Array, PLA). A hardware form is implemented. The processor 110 may integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU), a modem, and the like. Among them, the CPU mainly handles the operating system, user interface and application programs, etc.; the GPU is used for rendering and drawing of the display content; the modem is used to handle wireless communication. It can be understood that, the above-mentioned modem may also not be integrated into the processor 110, and is implemented by a communication chip alone.

The memory 120 may include random access memory (Random Access Memory, RAM), or may include read-only memory (Read-Only Memory). Memory 120 may be used to store instructions, programs, codes, sets of codes, or sets of instructions. The memory 120 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playback function, an image playback function, etc.) , instructions for implementing the following method embodiments, and the like. The storage data area may also store data created by the terminal 100 during use (such as phone book, audio and video data, chat record data) and the like.

Please refer to FIG. 17 , which shows a structural block diagram of a computer-readable storage medium provided by an embodiment of the present application. The computer-readable medium 1700 stores program codes, and the program codes can be invoked by the processor to execute the methods described in the above method embodiments.

The computer-readable storage medium 1700 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), EPROM, hard disk, or ROM. Optionally, the computer-readable storage medium 1700 includes a non-transitory computer-readable storage medium. Computer readable storage medium 1700 has storage space for program code 1710 to perform any of the method steps in the above-described methods. These program codes can be read from or written to one or more computer program products. Program code 1710 may be compressed, for example, in a suitable form.

Please refer to FIG. 18, which shows a computer program product 1800 provided by an embodiment of the present application, including a computer program/instruction 1810, which implements the above method when the computer program/instruction is executed by a processor.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or some technical features thereof are equivalently replaced; and these modifications or replacements do not drive the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

An image updating method, comprising:

Obtain the face image to be recognized;

Matching the face image to be identified with the face image of the registered user in the database to obtain a matching result;

If the matching result is less than the first threshold and greater than the second threshold, the face image to be recognized is updated to the database as a new face image, where the matching result is greater than the second threshold , the to-be-recognized face image and the registered user's face image correspond to the same user, and when the matching result is less than the first threshold, the to-be-recognized face image and the registered user's face image There are differences.
The method according to claim 1, wherein the updating the face image to be recognized to the database as a new face image comprises:

Obtain the first feature data of the face image to be recognized and the second feature data of the registered user's face image;

The feature data of the new face image of the registered user is acquired according to the first feature data and the second feature data, and the new feature data is updated to the database.
The method according to claim 2, wherein the acquiring the feature data of the new face image of the registered user according to the first feature data and the second feature data comprises:

obtaining the first weight value of the first characteristic data and the second weight value of the second characteristic data;

Obtain third characteristic data according to the first weight value, the first characteristic data, the second weight value and the second characteristic data;

The third feature data is used as feature data of a new face image of the registered user.
The method according to claim 3, wherein obtaining the third characteristic data according to the first weight value, the first characteristic data, the second weight value and the second characteristic data comprises:

Determine first reference data according to the first weight value and the first feature data;

Obtain second reference data according to the second weight value and the second characteristic data;

The third characteristic data is obtained according to the first reference data and the second reference data.
The method according to claim 2, wherein the acquiring the second feature data of the face image of the registered user comprises:

Obtain all face images corresponding to the registered user;

The feature data of all the face images are processed to obtain second feature data.
The method according to claim 5, wherein the processing of the feature data of the all face images to obtain the second feature data comprises:

Averaging processing is performed on the feature data of all the face images to obtain second feature data.
The method according to any one of claims 1-6, wherein the updating the face image to be recognized to the database as a new face image comprises:

Obtain the first time of the last updated face image of the registered user in the database;

Determine the acquisition time of the face image to be recognized as the second time;

If the time interval between the first time and the second time is greater than a specified threshold, the face image to be recognized is updated to the database as a new face image.
The method according to any one of claims 1-7, wherein, if the matching result is less than a first threshold and greater than a second threshold, the to-be-recognized face image is used as a new face Before the image is updated to the database, it also includes:

Acquiring a plurality of sample images, the plurality of sample images include a first image marked as a positive example and a second image marked as a negative example, the positive example represents that the sample image matches the face image in the database, and the negative example represents the sample image Does not match the face image in the database;

Obtain a matching result between each of the sample images and the face images of the registered users in the database;

A first ratio and a second ratio are obtained based on the matching results of each of the sample images, wherein the first ratio is used to characterize the number of positive matching results in the first image and the positive matching results in all sample images The second ratio is used to represent the ratio of the number of positive examples in the second image and the number of negative examples in all sample images;

The first threshold and the second threshold are set based on the first ratio and the second ratio.
The method of claim 8, wherein the setting the first threshold and the second threshold based on the first ratio and the second ratio comprises:

obtaining the first index and the second index of the first ratio and the third index of the second ratio;

determining a first threshold based on the first index of the first ratio and the third index of the second ratio;

A second threshold is determined based on the second index of the first ratio and the third index of the second ratio.
The method according to claim 8, wherein the obtaining the first ratio and the second ratio based on the matching result of each of the sample images comprises:

Among all the matching results of the first image, the matching results greater than the specified threshold are recorded as the first result TP, and the matching results less than or equal to the specified threshold are recorded as the second result FP, and the second results represent the The matching result is a negative example;

Among all the matching results of the second image, the matching results greater than the specified threshold are recorded as the third result FN, and the matching results less than or equal to the specified threshold are recorded as the fourth result TN, and the fourth result represents the The matching result is a negative example;

The first ratio is TPR=TP/(TP+FN), and the second ratio is FPR=FP/(TN+FP).
The method according to any one of claims 1-10, wherein the matching of the face image to be identified with the face image of the registered user in the database to obtain a matching result, comprising:

Perform image quality evaluation on the face image to be recognized to obtain an image quality evaluation result;

If the image quality evaluation result satisfies the preset condition, the face image to be recognized is matched with the face image of the registered user in the database to obtain a matching result.
The method according to any one of claims 1-11, wherein the matching of the face image to be identified with the face image of a registered user in a database to obtain a matching result, comprising:

If there are multiple face images corresponding to the registered user, match the face image to be recognized with each face image of the registered user to obtain the similarity corresponding to each face image;

The matching result between the face image to be recognized and the face image of the registered user is determined according to the similarity of each face image.
The method according to any one of claims 1-12, wherein the database includes a plurality of registered users, and the face image to be recognized is compared with the face images of the registered users in the database. Matching to obtain a matching result, if the matching result is less than the first threshold and greater than the second threshold, then the face image to be recognized is updated to the database as a new face image, including:

Obtain the matching result corresponding to the face image to be recognized and the face image of each registered user, and find the matching result that is smaller than the first threshold and greater than the second threshold from all the matching results. The registered user corresponding to the target matching result is regarded as the target user;

The face image to be recognized is written into the database, and the face image to be recognized is stored in the database corresponding to the target user.
The method according to claim 13, characterized in that, writing the face image to be recognized into the database, wherein the face image to be recognized corresponds to the target user in the database storage, including:

The historical face image stored correspondingly to the target user in the database is retained, and the face image to be recognized is stored as a new face image of the target user.
The method according to claim 13, wherein the face image to be recognized is written into the database, and the face image to be recognized corresponds to the target user in the database storage, including:

At least one face image in all face images corresponding to the target user in the database is replaced with the face image to be recognized.
The method according to claim 15, wherein the replacing at least one face image in all face images corresponding to the target user in the database with the face image to be recognized comprises:

When the target user corresponds to multiple face images, the target face image is determined from the multiple face images, and the target face image is updated to the face image to be recognized.
The method according to claim 16, wherein the determining the target face image from the plurality of face images comprises:

The face image with the lowest matching result among the plurality of face images is used as the target face image.
The method according to claim 16, wherein the determining the target face image from the plurality of face images comprises:

The face image with the highest matching result among the plurality of face images is used as the target face image.
The method according to any one of claims 1-18, wherein, if the matching result is less than a first threshold and greater than a second threshold, the to-be-recognized face image is used as a new face Images are updated to the database, including:

If the matching result is less than the first threshold and greater than the second threshold, sending prompt information to the user terminal, so that the user terminal displays the prompt information in the designated interface;

When a confirmation instruction input by the user based on the prompt information is detected, the face image to be recognized is updated to the database as a new face image based on the confirmation instruction.
An image updating device, comprising:

an acquisition unit for acquiring a face image to be recognized;

a matching unit for matching the face image to be identified with the face image of the registered user in the database to obtain a matching result;

An update unit, configured to update the face image to be recognized as a new face image to the database if the matching result is less than a first threshold and greater than a second threshold, wherein the matching result is greater than the first In the case of two thresholds, the face image to be recognized corresponds to the same user as the face image of the registered user, and when the matching result is less than the first threshold, the face image to be recognized corresponds to the registered user. There are differences between the user's face images.
An electronic device, comprising:

one or more processors;

memory;

One or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs are configured to perform such as The method of any one of claims 1-19.
A computer-readable medium, characterized in that the computer-readable storage medium stores program codes, and the program codes can be invoked by a processor to execute the method of any one of claims 1-19.
A computer program product, characterized in that it includes a computer program/instruction, characterized in that, when the computer program/instruction is executed by a processor, the method according to any one of claims 1-19 is implemented.