WO2023279783A1

WO2023279783A1 - Facial recognition method, device, and storage medium

Info

Publication number: WO2023279783A1
Application number: PCT/CN2022/083552
Authority: WO
Inventors: 徐天宇
Original assignee: 中兴通讯股份有限公司
Priority date: 2021-07-06
Filing date: 2022-03-29
Publication date: 2023-01-12
Also published as: CN115588216A

Abstract

The present disclosure belongs to the technical field of facial recognition, and provided in embodiments thereof are a facial recognition method, a device, and a storage medium. The method comprises: obtaining a plurality of similarity measure values in instances of historical facial recognition, wherein each similarity measure value is a similarity measure of a registration facial image and a facial image in an instance of historical facial recognition; determining an adjustment policy for a preset threshold according to a comparison value of the plurality of similarity measure values and said preset threshold; adjusting the preset threshold according to the determined adjustment policy and using the adjusted preset threshold for facial recognition comparison.

Description

Face recognition method, device and storage medium

Cross References to Related Applications

This disclosure claims the priority of the Chinese patent application CN202110765422.8 filed on July 06, 2021, entitled "Face Recognition Method, Device and Storage Medium", the entire content of which is incorporated by reference into this disclosure.

technical field

The present disclosure relates to the technical field of face recognition, and in particular to a face recognition method, a face recognition device and a storage medium.

Background technique

With the continuous development of 5G technology, benefiting from its own characteristics of low latency, high bandwidth, and wide connection, a large number of industrial applications relying on 5G technology have emerged in recent years, such as the popularization of 5G smart mobile terminals, relying on smart cameras Smart classrooms and the autonomous driving industry relying on vehicle-road collaboration technology, etc., and face recognition technology is an important part of it. Face recognition is a biometric technology for identity recognition through facial feature information. It mainly extracts facial features from collected facial images, and calculates the similarity value between the extracted facial features and registered facial features. Compare the similarity value with the preset threshold, if it is greater than the preset threshold, the recognition is passed, otherwise it is not passed. However, in practical applications, the environment of face recognition is complex and changeable, such as changes in the user's face posture, makeup, or changes in light intensity, so the fixed preset threshold cannot adapt to the changeable environment .

Contents of the invention

The main purpose of the embodiments of the present disclosure is to provide a face recognition method, a face recognition device, and a storage medium, aiming to dynamically adjust the preset threshold according to the similarity data of the user in the historical face recognition process. In the case of changing the face recognition model, the face recognition method can be adapted to different scenarios and improve user experience.

In the first aspect, an embodiment of the present disclosure provides a face recognition method, including: obtaining a plurality of similarity values in historical face recognition, wherein the similarity value is the difference between the face image in the historical face recognition and the registrant The similarity of the face image; the adjustment strategy of the preset threshold is determined according to the comparison result of the multiple similarity values and the preset threshold; the preset threshold is adjusted according to the determined adjustment strategy, and the adjusted Preset thresholds are used for face recognition comparison.

In the second aspect, the embodiment of the present disclosure also provides a face recognition device, the face recognition device includes a processor, a memory, a computer program stored on the memory and executable by the processor, and used to implement A data bus connecting and communicating between the processor and the memory, wherein when the computer program is executed by the processor, the steps of any one face recognition method provided in this disclosure specification are realized.

In a third aspect, an embodiment of the present disclosure further provides a storage medium for computer-readable storage, the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors , so as to realize the steps of any face recognition method provided in this disclosure specification.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present application more clearly, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present application. Ordinary technicians can also obtain other drawings based on these drawings on the premise of not paying creative work.

FIG. 1 is a schematic flow diagram of a face recognition method provided by an embodiment of the present disclosure;

FIG. 2 is a schematic structural block diagram of a face recognition device provided by an embodiment of the present disclosure.

detailed description

The following will clearly and completely describe the technical solutions in the embodiments of the present disclosure with reference to the accompanying drawings in the embodiments of the present disclosure. Apparently, the described embodiments are part of the embodiments of the present disclosure, not all of them. Based on the embodiments in the present disclosure, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present disclosure.

The flow charts shown in the drawings are just illustrations, and do not necessarily include all contents and operations/steps, nor must they be performed in the order described. For example, some operations/steps can be decomposed, combined or partly combined, so the actual order of execution may be changed according to the actual situation.

It should be understood that the terminology used in the present disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the present disclosure. As used in this disclosure and the appended claims, the singular forms "a", "an" and "the" are intended to include plural referents unless the context clearly dictates otherwise.

Embodiments of the present disclosure provide a face recognition method, a face recognition device, and a storage medium. Wherein, the face recognition method can be applied to a mobile terminal, and the mobile terminal can be electronic equipment such as a mobile phone, a tablet computer, a notebook computer, a desktop computer, a personal digital assistant, and a wearable device.

Some embodiments of the present disclosure will be described in detail below in conjunction with the accompanying drawings. In the case of no conflict, the following embodiments and features in the embodiments can be combined with each other.

Please refer to FIG. 1 , which is a schematic flowchart of a face recognition method provided by an embodiment of the present disclosure.

As shown in FIG. 1 , the face recognition method includes steps S101 to S103.

Step S101, obtaining a plurality of similarity values in historical face recognition, wherein the similarity value is the similarity between the face image in the historical face recognition and the registered face image.

In one embodiment, firstly, the registered face images pre-registered in the face feature database are processed, and the face detection network is used to detect the face area, and then the extracted face area is input to the face recognition network for further processing. Feature extraction, get the registered face image features and save them. Every time face recognition is performed, the collected face images are extracted through the face detection network and face recognition network, and the face image features are matched with the registered face image features to calculate the similarity of the features. degree to get the similarity value of each face recognition.

In some embodiments, the face detection network can be a SSD target detection model (Single Shot MultiBox Detector, SSD), and the face recognition network can be a MobileFaceNet face recognition model, and of course the face detection network and the face recognition network can also be Other network models are not limited in this application.

It should be noted that the multiple similarity values only record the similarity values obtained by registered users through face recognition, and multiple similarity values may be obtained in each face recognition, and each face recognition only records one similarity value. degree value. In one embodiment, during the face recognition process based on the user's face image, if the calculated similarity value is greater than a preset threshold, the unlocking is successful; otherwise, the unlocking fails. During this process, if the user successfully unlocks through the face, record the similarity value when the unlock is successful. If the user cannot unlock the device through the face image, and there is no follow-up action, the current user is considered to be a non-registered user, and the similarity value of this face recognition will not be recorded. If the user fails to unlock through the face image, and then unlocks by entering a password or other authentication methods, the current user is considered to be a registered user, and the similarity value at the time of the last unlock failure is recorded.

In some embodiments, multiple similarity values can be obtained by acquiring historical face recognition data within a preset period, wherein the preset period includes: a preset duration, or a preset number of face recognition times , or the preset number of face recognition failures.

In some embodiments, the preset period is a preset time length, and each time the historical face recognition of the preset time length is passed, a plurality of similarity values obtained by face recognition within the preset time length are automatically obtained and analyzed And it is used for subsequently determining the adjustment strategy of the preset threshold.

For example, when the preset period is one week, after each week of historical face recognition, a plurality of similarity values obtained in the face recognition of the previous week are automatically obtained, analyzed and used for subsequent determination of the preset threshold adjustment strategy.

In some embodiments, the preset period is the preset number of times of face recognition, and when the number of times of face recognition reaches the preset number of times, a plurality of similarity values obtained by the preset face recognition are automatically obtained, and performed Analysis and adjustment strategies for subsequent determination of preset thresholds.

Exemplarily, when the preset period is 100 times of face recognition, when the number of face recognition reaches 100 times, the 100 similarity values obtained by these 100 times of face recognition are automatically obtained, analyzed and used for subsequent Identify adjustment strategies for preset thresholds.

In some embodiments, in order to better improve the user experience, the preset period can also be set to the preset number of face recognition failures. When the number of face recognition failures reaches the preset number of times, it will trigger the preset threshold Adjustment. The multiple similarity values obtained by face recognition within this period of time are automatically acquired, analyzed and used to subsequently determine an adjustment strategy for a preset threshold.

Exemplarily, when the preset period is 10 times of face recognition failures, when the number of face recognition failures reaches 10 times, it will automatically trigger the adjustment of the preset threshold, by obtaining multiple similar The degree value is analyzed and used to subsequently determine the adjustment strategy for the preset threshold.

Step S102 , determining an adjustment strategy for the preset threshold according to a comparison result between the plurality of similarity values and a preset threshold.

In one embodiment, according to the comparison results of multiple similarity values and the preset threshold value, the first ratio of the ratio used to indicate the success of face recognition and the second ratio of the ratio used to indicate the failure of face recognition are determined, and then according to The first ratio and the second ratio determine different adjustment strategies for the preset threshold.

Wherein, the method for determining the first ratio and the second ratio specifically includes counting the number of the plurality of similarity values greater than or equal to the preset threshold, and according to the number of the plurality of similarity values and the greater than Or the number equal to the preset threshold value determines the first ratio; counts the number of the multiple similarity values smaller than the preset threshold value, and according to the number of the multiple similarity values and the smaller than the The number of the preset thresholds determines the second ratio.

In some embodiments, when the first ratio, that is, the ratio of successful face recognition is greater than the first preset ratio, it means that the preset threshold is too conservative, and can be appropriately increased to ensure user safety. In this case, the adjustment strategy of the preset threshold is a first adjustment strategy, and the first adjustment strategy is to increase the preset threshold.

Exemplarily, when the first preset ratio is set to 95% and the first ratio is greater than 95%, the preset threshold may be appropriately increased, and the adjusted preset threshold may be used for subsequent face recognition.

In some embodiments, when the second ratio, that is, the ratio of face recognition failures is greater than the second preset ratio, it means that the preset threshold is too high, or the environment during this time period belongs to an abnormal scene, which can be appropriately Lower the threshold to improve user experience. At this time, the adjustment strategy of the preset threshold is the second adjustment strategy, and the second adjustment strategy is to lower the preset threshold.

Exemplarily, when the second preset ratio is set to 90% and the second ratio is greater than 90%, the preset threshold should be appropriately lowered, and the lowered preset threshold should be used for subsequent face recognition.

In an implementation manner, the adjustment range corresponding to the first adjustment strategy and the second adjustment strategy may be determined according to a difference between an average of multiple similarity values and a preset threshold. In one embodiment, the mean value of the similarity obtained by performing mean filtering on a plurality of similarity values is used as a reference value of the preset threshold. The adjusted preset threshold is determined according to the difference between the reference value of the preset threshold and the preset threshold before adjustment. The formula is characterized as:

S _new ＝S+0.1×(S _t -S)

Wherein, S _new is an adjusted preset threshold, S is a preset threshold before adjustment, and S _t is a reference value of the preset threshold.

In some embodiments, multiple similarity values can also be calculated by median average filtering, and the average value of the similarity calculated after removing the maximum and minimum values among the multiple similarity values can be used as a reference value for the preset threshold, In order to reduce the impact of extreme values on the mean similarity in abnormal scenarios.

In some embodiments, when the first ratio, that is, the ratio of face recognition success is greater than the third preset ratio and smaller than the first preset ratio, and the second ratio, that is, the ratio of face recognition failure is greater than the fourth preset ratio and is less than the second preset ratio, in this case, multiple similarity values fluctuate up and down the preset threshold, so it can be analyzed that the usage scenarios during this period of time are relatively complex, and there are both normal scenarios and some Abnormal scenes, such as frequent changes in the lighting conditions of the current scene, are more sensitive to the current preset threshold. In order to better improve user experience, the adjustment strategy of the preset threshold is the third adjustment strategy, the third adjustment strategy is to lower the preset threshold and the range of lowering the preset threshold is determined by the preset range Sure.

Exemplarily, if the third preset ratio is set to 40%, and the fourth preset ratio is set to 50%, when the first ratio, that is, the ratio of successful face recognition is greater than 40% and less than 95%, and the second When the second ratio, that is, the proportion of face recognition failures is greater than 50% and less than 80%, it indicates that the similarity value fluctuates around the preset threshold. In order to better adapt to various scenarios, the preset threshold should be lowered at this time.

It should be noted that when the similarity value fluctuates above and below the preset threshold, if the adjustment range of the preset threshold is determined according to the average value of multiple similarities, since the multiple similarity values are greater than or equal to the preset threshold and less than the preset The distribution of the thresholds is relatively close. At this time, the reference value of the preset threshold obtained according to the average value of multiple similarities may be relatively close to the original threshold before adjustment. Therefore, in this case, the preset range is used as the adjustment range of the preset threshold, and the specific value of the preset range can be determined according to experimental data and actual usage scenarios of face recognition, which is not limited in the present application.

It should be noted that the first preset ratio, the second preset ratio, the third preset ratio, and the fourth preset ratio can be predetermined according to experimental data and specific application scenarios of face recognition. Not limited.

Step S103. Adjust the preset threshold according to the determined adjustment strategy, and use the adjusted preset threshold for face recognition comparison.

In one embodiment, after the adjustment strategy of the preset threshold is determined according to multiple similarity values of historical face recognition, the preset threshold is adjusted based on the strategy of increasing or decreasing and the adjustment range included in the adjustment strategy, and then The adjusted preset threshold is used for subsequent face recognition until the next adjustment of the preset threshold is triggered.

Due to the wide range of application scenarios of face recognition, in some application scenarios, the threshold of face recognition can be associated with the corresponding user of the terminal through the terminal device of the registered user, and the preset threshold of a registered user can be adjusted separately , such as face recognition access control attendance system, face recognition anti-theft door. In other application scenarios, the threshold of face recognition is common to all registered users, and does not support individual adjustment of the preset threshold of a certain registered user, such as the face recognition intelligent access control system in residential quarters. Therefore, the adjustment of the preset threshold can be divided into separate statistical analysis of multiple similarity values from the same registered user or different registered users according to the actual application scenario, and adjust the threshold according to the analysis results of multiple similarity values. The preset threshold for the same registered user or the preset threshold for different registered users.

In some embodiments, when the multiple similarity values are from the same registered user, the determined adjustment policy is used to adjust the preset threshold for face recognition performed by the same registered user.

Exemplarily, in the face recognition access control time attendance system, the access control time attendance system is associated with each registered user through the user's terminal device. When the face recognition method is enabled, the initial value of the preset threshold of all registered users is the same, and then According to the comparison result between the similarity value of each registered user in the historical face recognition and the preset threshold value, the preset threshold value of a certain registered user is individually adjusted.

In some embodiments, when the multiple similarity values are from different registered users, the determined adjustment strategy is used to adjust the preset thresholds for face recognition performed by the different registered users.

Exemplarily, the face recognition intelligent access control system in a residential area may not be associated with each registered user in the residential area through a device, according to the similarity of all registered users who use the face recognition intelligent access control system within this period The degree value is used to determine the adjustment strategy of the preset threshold, and the adjusted preset threshold is also used for face recognition by all registered users registered in the access control system.

In order to make the purpose, technical solution and advantages of the present disclosure clearer, the embodiments of the present disclosure will be described in detail below in conjunction with the accompanying drawings. It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined arbitrarily with each other.

Embodiment one

This embodiment discloses a face recognition method, which is applied to an intelligent access control system in a remote control scene, and the method includes the following steps.

Step S1: Process the registered face images that are pre-entered into the face feature database, first use the SSD face detection network to detect face regions, and then input the extracted face image regions to the MobileFaceNet face recognition network for feature extraction , get the registered face image features and save them.

Step S2: When the user uses the face image to unlock the access control, record the similarity value when the user unlocks the access control each time. In one embodiment, each time face recognition is performed, the user's face image is collected through the front camera, and features of the collected face image are extracted. Extracting features is also to perform face detection through the SSD face detection network first, and then input the detected face image into the MobileFaceNet face recognition network to extract face image features, and perform the extracted face image features and registered face image features Match, calculate the similarity value. If the calculated similarity value is greater than or equal to the preset threshold, the access control can be successfully unlocked, otherwise, the unlocking fails.

During this process, if the user successfully unlocks the access control through the face image, record the similarity value when the unlock is successful; The face similarity value is not recorded. If the user enters a password to unlock after failing to unlock through the face image, the current user is considered to be a registered user, and the current environment belongs to an abnormal scene, and the similarity value at the time of the last unlocking failure is recorded.

Step S3: According to the preset cycle parameters set by the system, the similarity value in the current cycle is counted. In this embodiment, the preset cycle is set to unlock 100 times through face recognition. When the user unlocks the access control 100 times, the statistical analysis of the similarity value and the adjustment of the preset threshold are automatically triggered.

Step S4: According to the comparison result of the statistical similarity value and the preset threshold, determine different adjustment strategies, as follows:

For example, 96 times out of 100 are unlocked successfully, indicating that the preset threshold before adjustment is conservative for the current user, and the threshold can be increased to improve the security of the access control system.

For example, 82 unlocking failures out of 100 times indicate frequent unlocking failures in the current period, indicating that the user is frequently in abnormal situations. In order to improve user experience, the preset threshold needs to be lowered. The adjustment range is determined based on the mean value calculated by the median average filtering method for the similarity values obtained from these 100 face recognitions, and the average similarity value obtained after removing the maximum and minimum values is used as a reference value for the preset threshold, and according to The reference value of the preset threshold adjusts the preset threshold.

For example, 45 times out of 100 are unlocked successfully and 55 times failed, indicating that the face similarity value fluctuates above and below the threshold in the current period. To improve the user experience, the face recognition method can be adapted to different environments, the preset threshold can be lowered, and the adjustment range can be set to 0.5.

Step S5: Use the adjusted preset threshold for subsequent face recognition comparisons until the next preset threshold adjustment is triggered.

In the embodiments of the present disclosure, by analyzing multiple similarity values of a user in a historical face recognition process, an adjustment strategy for a preset threshold is determined and the preset threshold is dynamically adjusted. When the similarity between the face image to be recognized and the registered face image in a period of time is very high, the preset threshold is increased to increase the anti-prosthesis attack capability. When the similarity between the face image to be recognized and the registered face image fluctuates above and below the threshold for a period of time, resulting in a poor unlocking experience for the user, the preset threshold is appropriately lowered to reduce the false rejection rate for users while ensuring safety . Without changing the original face recognition model, the face recognition method can be adapted to different environments, and the user experience is improved.

Please refer to FIG. 2 . FIG. 2 is a schematic structural block diagram of a face recognition device provided by an embodiment of the present disclosure.

As shown in FIG. 2, the face recognition device 300 includes a processor 301 and a memory 302, and the processor 301 and the memory 302 are connected through a bus 303, such as an I2C (Inter-integrated Circuit) bus.

In one embodiment, the processor 301 is used to provide computing and control capabilities to support the operation of the entire face recognition device. The processor 301 can be a central processing unit (Central Processing Unit, CPU), and the processor 301 can also be other general processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC) ), Field-Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Wherein, the general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

In one embodiment, the memory 302 may be a Flash chip, a read-only memory (ROM, Read-Only Memory) disk, an optical disk, a U disk, or a mobile hard disk.

Those skilled in the art can understand that the structure shown in Figure 2 is only a block diagram of a partial structure related to the embodiment of the present disclosure, and does not constitute a limitation on the face recognition device to which the embodiment of the present disclosure is applied. , a particular server may include more or fewer components than shown in the figures, or combine certain components, or have a different arrangement of components.

Wherein, the processor is configured to run a computer program stored in the memory, and implement any one of the face recognition methods provided by the embodiments of the present disclosure when executing the computer program.

In one embodiment, the processor is configured to run a computer program stored in a memory, and implement the following steps when executing the computer program: obtain a plurality of similarity values in historical face recognition, wherein the similarity The degree value is the similarity between the face image in the historical face recognition and the registered face image; determine the adjustment strategy of the preset threshold according to the comparison results of the multiple similarity values and the preset threshold; according to the determined The adjustment strategy adjusts the preset threshold, and uses the adjusted preset threshold for face recognition comparison.

In an embodiment, when the processor realizes the adjustment strategy of determining the preset threshold according to the comparison result of the multiple similarity values and the preset threshold, it is configured to: determine the first threshold according to the comparison result. A ratio and a second ratio, wherein the first ratio is used to represent the ratio of successful face recognition, and the second ratio is used to represent the ratio of failed face recognition. counting the number of the multiple similarity values greater than or equal to the preset threshold, and determining the first Proportion; counting the number of the multiple similarity values smaller than the preset threshold, and determining the second ratio according to the number of the multiple similarity values and the number smaller than the preset threshold; An adjustment strategy for the preset threshold is determined according to the first ratio and the second ratio.

In an embodiment, when the processor realizes the adjustment strategy of determining different preset thresholds according to the first ratio and the second ratio, it is configured to realize: when the first ratio is greater than a first preset ratio, the adjustment strategy of the preset threshold is the first adjustment strategy, and the first adjustment strategy is to increase the preset threshold; when the second ratio is greater than the second preset ratio, the preset The threshold adjustment strategy is a second adjustment strategy, and the second adjustment strategy is to lower the preset threshold.

In an embodiment, when the processor implements the face recognition method, it is configured to implement: the adjustment range corresponding to the first adjustment strategy and the second adjustment strategy is based on the mean value and the value of the plurality of similarity values The difference between the preset thresholds is determined.

In an embodiment, when the processor realizes the adjustment strategy of determining different preset thresholds according to the first ratio and the second ratio, it is configured to realize: when the first ratio is greater than a third preset When the ratio is smaller than the first preset ratio, and the second ratio is larger than the fourth preset ratio and smaller than the second preset ratio, the adjustment strategy of the preset threshold is the third adjustment strategy, the first The third adjustment strategy is to lower the preset threshold, and the range of lowering the preset threshold is determined by the preset range.

In one embodiment, when the processor implements the face recognition method, it is used to realize that: the multiple similarity values are from the same registered user or different registered users; when the multiple similarity values are from When the same registered user, the determined adjustment strategy is used to adjust the preset threshold for face recognition of the same registered user; when the multiple similarity values are from different registered users, the determined The adjustment strategy is used to adjust the preset threshold for face recognition of the different registered users.

In an embodiment, when implementing the face recognition method, the processor is configured to realize: when recording multiple similarity values in the historical face recognition, when the face image in the historical face recognition is consistent with When the similarity value of the registered face image is less than the preset threshold, further confirm whether the user corresponding to the face image is a registered user; when confirming that the user corresponding to the face image is a registered user, record the last The similarity value calculated when a matching is unsuccessful; when it is confirmed that the user corresponding to the face image is a non-registered user, the similarity value calculated based on the face image is not recorded.

In an embodiment, when implementing the face recognition method, the processor is configured to realize: the historical face recognition is historical face recognition within a preset period, and the preset period includes: preset The duration, the preset number of times of face recognition or the preset number of times of failed face recognition.

It should be noted that those skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of the face recognition device described above can refer to the corresponding process in the aforementioned embodiment of the face recognition method. This will not be repeated here.

An embodiment of the present disclosure also provides a storage medium for computer-readable storage, the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the following: The steps of any face recognition method provided in the description of the embodiments of the present disclosure.

Wherein, the storage medium may be an internal storage unit of the face recognition device described in the foregoing embodiments, such as a hard disk or a memory of the face recognition device. The storage medium can also be an external storage device of the face recognition device, such as a plug-in hard disk equipped on the face recognition device, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, flash memory card (Flash Card), etc.

Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, the functional modules/units in the system, and the device can be implemented as software, firmware, hardware, and an appropriate combination thereof. In hardware embodiments, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be composed of several physical components. Components cooperate to execute. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit . Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). As known to those of ordinary skill in the art, the term computer storage media includes both volatile and nonvolatile media implemented in any method or technology for storage of information, such as computer readable instructions, data structures, program modules, or other data. permanent, removable and non-removable media. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cartridges, tape, magnetic disk storage or other magnetic storage devices, or can Any other medium used to store desired information and which can be accessed by a computer. In addition, as is well known to those of ordinary skill in the art, communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media .

It should be understood that the term "and/or" used in the present disclosure and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations. It should be noted that, as used herein, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or system comprising a set of elements includes not only those elements, It also includes other elements not expressly listed, or elements inherent in the process, method, article, or system. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article or system comprising that element.

The serial numbers of the above-mentioned embodiments of the present disclosure are for description only, and do not represent the advantages and disadvantages of the embodiments. The above are only specific embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any person skilled in the art can easily think of various equivalents within the technical scope of the present disclosure. Modifications or replacements should be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be determined by the protection scope of the claims.

Claims

A face recognition method, comprising:

Obtaining a plurality of similarity values in historical face recognition, wherein the similarity value is the similarity between a face image in historical face recognition and a registered face image;

determining an adjustment strategy for the preset threshold according to a comparison result between the plurality of similarity values and a preset threshold;

The preset threshold is adjusted according to the determined adjustment policy, and the adjusted preset threshold is used for face recognition comparison.
The face recognition method according to claim 1, wherein the adjustment strategy for determining the preset threshold based on the comparison result of the plurality of similarity values and the preset threshold includes:

Determine a first ratio and a second ratio according to the comparison result, wherein the first ratio is used to represent the ratio of successful face recognition, and the second ratio is used to represent the ratio of failed face recognition;

counting the number of the multiple similarity values greater than or equal to the preset threshold, and determining the first Proportion;

Counting the number of the multiple similarity values smaller than the preset threshold, and determining the second ratio according to the number of the multiple similarity values and the number smaller than the preset threshold;

An adjustment strategy for the preset threshold is determined according to the first ratio and the second ratio.
The face recognition method according to claim 2, wherein determining the adjustment strategy of the preset threshold according to the first ratio and the second ratio includes:

When the first ratio is greater than the first preset ratio, the adjustment strategy of the preset threshold is a first adjustment strategy, and the first adjustment strategy is to increase the preset threshold;

When the second ratio is greater than the second preset ratio, the adjustment strategy of the preset threshold is a second adjustment strategy, and the second adjustment strategy is to lower the preset threshold.
The face recognition method according to claim 3, wherein the adjustment range corresponding to the first adjustment strategy and the second adjustment strategy is determined according to the difference between the average of the multiple similarity values and the preset threshold .
The face recognition method according to claim 2, wherein the adjustment strategy for determining the preset threshold according to the first ratio and the second ratio includes:

When the first ratio is greater than the third preset ratio and smaller than the first preset ratio, and the second ratio is larger than the fourth preset ratio and smaller than the second preset ratio, the preset threshold The adjustment strategy is a third adjustment strategy, the third adjustment strategy is to lower the preset threshold and the range of lowering the preset threshold is determined by the preset range.
The face recognition method according to claim 1, wherein the multiple similarity values are from the same registered user or different registered users;

When the multiple similarity values come from the same registered user, the determined adjustment strategy is used to adjust the preset threshold for face recognition of the same registered user;

When the multiple similarity values are from different registered users, the determined adjustment strategy is used to adjust the preset thresholds for face recognition performed by the different registered users.
The face recognition method according to claim 1, further comprising:

When recording multiple similarity values in the historical face recognition, when the similarity value between the face image in the historical face recognition and the registered face image is less than the preset threshold, further confirm the Whether the user corresponding to the face image is a registered user;

When confirming that the user corresponding to the face image is a registered user, record the similarity value calculated when the last matching is unsuccessful;

When it is confirmed that the user corresponding to the face image is a non-registered user, the similarity value calculated based on the face image is not recorded.
The face recognition method according to any one of claims 1 to 7, wherein the historical face recognition is historical face recognition within a preset period, and the preset period includes: preset duration, preset The number of face recognition times or the preset number of face recognition failures.
A face recognition device, wherein the face recognition device includes a processor, a memory, a computer program stored on the memory and executable by the processor, and a computer program for implementing the processor and the memory A data bus connecting and communicating between them, wherein when the computer program is executed by the processor, the steps of the face recognition method according to any one of claims 1 to 8 are realized.
A storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement claims 1 to 8 The steps of any one of the face recognition methods.