CN112889062A

CN112889062A - Face recognition data processing method and device, mobile device and computer readable storage medium

Info

Publication number: CN112889062A
Application number: CN201880098811.6A
Authority: CN
Inventors: 梁俊豪
Original assignee: Guangdong Oppo Mobile Telecommunications Corp Ltd; Shenzhen Huantai Technology Co Ltd
Current assignee: Guangdong Oppo Mobile Telecommunications Corp Ltd; Shenzhen Huantai Technology Co Ltd
Priority date: 2018-12-07
Filing date: 2018-12-07
Publication date: 2021-06-01
Anticipated expiration: 2038-12-07
Also published as: WO2020113571A1; CN112889062B

Abstract

A face recognition data processing method comprises the following steps: the mobile equipment receives face recognition error feedback information, and adjusts a corresponding recognition control threshold value in a current face recognition algorithm according to the face recognition error feedback information; and carrying out face recognition according to the current face recognition algorithm after the recognition control threshold value is adjusted.

Description

Face recognition data processing method and device, mobile device and computer readable storage medium

Technical Field

The present application relates to the field of image recognition, and in particular, to a method and an apparatus for processing face recognition data, a mobile device, and a non-volatile computer-readable storage medium.

Background

With the development of image recognition technology, more and more mobile devices use face recognition to complete authentication, such as unlocking and payment verification, and convenience is brought to the life of people. The face recognition is influenced by the environment and the face state, such as light, hair shielding and the like, and the situation of false recognition may exist.

The traditional system does not process the situation of the false recognition, and the phenomenon of the false recognition is likely to occur when the same user uses the face recognition authentication in the same scene next time.

Disclosure of Invention

The embodiment of the application provides a face recognition data processing method and device, a mobile device and a nonvolatile computer readable storage medium, which can automatically adjust a face recognition algorithm control threshold value through feedback information, and improve the face recognition accuracy.

A face recognition data processing method comprises the following steps:

the mobile equipment receives face recognition error feedback information;

adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information; and

and carrying out face recognition according to the current face recognition algorithm after the recognition control threshold value is adjusted.

A face recognition data processing apparatus comprising:

the receiving module is used for receiving the face recognition error feedback information through the mobile equipment;

the adjusting module is used for adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information;

and the recognition module is used for carrying out face recognition according to the current face recognition algorithm after the recognition control threshold value is adjusted.

A mobile device comprising a memory and a processor, the memory having stored therein computer-readable instructions that, when executed by the processor, cause the processor to perform the steps of:

receiving face recognition error feedback information through the mobile equipment;

A non-transitory computer readable storage medium having stored thereon computer readable instructions which, when executed by a processor, cause the processor to perform the steps of:

According to the face recognition data processing method, the face recognition data processing device, the mobile equipment and the nonvolatile computer readable storage medium, the face recognition error feedback information is received through the mobile equipment, the corresponding recognition control threshold value in the current face recognition algorithm is adjusted according to the face recognition error feedback information, face recognition is carried out according to the current face recognition algorithm after the recognition control threshold value is adjusted, and when the face recognition is wrong, the face recognition algorithm control threshold value is automatically and intelligently adjusted through the face recognition feedback information, so that the phenomenon of next wrong recognition is reduced, and the face recognition accuracy is improved.

Drawings

For a better understanding of the description and/or illustration of embodiments and/or examples of those applications disclosed herein, reference may be made to one or more of the drawings. The additional details or examples used to describe the figures should not be considered as limiting the scope of any of the disclosed inventions, the presently described embodiments and/or examples, and the presently understood best mode of these applications.

Fig. 1 is a schematic diagram of an application environment of a face recognition data processing method in an embodiment.

FIG. 2 is a flow diagram of a method for face recognition data processing in one embodiment.

FIG. 3 is a diagram of a lock screen interface in one embodiment.

FIG. 4 is a diagram of a desktop after unlocking is successful, under an embodiment.

FIG. 5 is a flow diagram of a method for face recognition data processing in one embodiment.

FIG. 6 is a flow diagram of face recognition in one embodiment.

FIG. 7 is a flow diagram of adjusting recognition control thresholds in one embodiment.

Fig. 8 is a flowchart of a face recognition data processing method in a specific embodiment.

Fig. 9 is a block diagram showing the configuration of a face recognition data processing apparatus according to an embodiment.

Fig. 10 is a schematic internal structure diagram of a mobile device in one embodiment.

Fig. 11 is a block diagram of a partial structure of a mobile device related to the present embodiment in one embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

It will be understood that the terms "first," "second," and the like, as used in the embodiments of the present application, may be used herein to describe various elements, but the elements are not limited by these terms. These terms are only used to distinguish one element from another. For example, a first control may be referred to as a second control, both of which are controls, but which are not the same, without departing from the scope of the present application.

Fig. 1 is an application environment diagram of a face recognition data processing method in an embodiment. As shown in fig. 1, the application environment includes a mobile device 110. The mobile device 110 is provided with a camera which can collect a face image and perform face recognition on the collected face image, the mobile device 110 is provided with a feedback information receiving key which can be a physical key or a virtual key, the face recognition error feedback information is received through the feedback information receiving key, a corresponding recognition control threshold value in a current face recognition algorithm is adjusted according to the face recognition error feedback information, and the face recognition is performed according to the current face recognition algorithm after the recognition control threshold value is adjusted. The mobile device 110 may be a smartphone, a tablet, a wearable device, a personal digital assistant, and the like.

FIG. 2 is a flow diagram of a method for face recognition data processing in one embodiment. As shown in fig. 2, a method for processing face recognition data, which is described by taking the application to the mobile device in fig. 1 as an example, specifically includes:

in operation 202, the mobile device receives face recognition error feedback information.

The face recognition error feedback information is feedback information for describing a face recognition error, and the face recognition error comprises various types of errors, such as false acceptance recognition and false rejection recognition. The false acceptance identification means that the false face is identified as a preset face image after algorithm identification, if the preset face image recorded in the mobile device is the face image of the user a, when the acquired face image is the face image of the user B, the algorithm incorrectly identifies the face image of the user B as the face image of the user a, so that face verification is incorrectly successful. The false rejection identification means that the correct face is identified as a non-preset face image after algorithm identification, if the preset face image recorded in the mobile device is the face image of the user a, when the acquired face image is the face image of the user a, the algorithm incorrectly identifies the face image of the user a as the face image of another user, so that face verification is unsuccessful.

Specifically, the face recognition error feedback information is received through an operation acting on the mobile device, wherein the operation may be an operation directly or indirectly acting on the mobile device, and the operation directly acting on the mobile device may be an operation performed on a physical key or a virtual key of a screen. The operation indirectly acting on the mobile device may be a gesture operation, an eyeball operation, or the like. Different types of face recognition error feedback information can be fed back through different keys, or different types of face recognition error feedback information can be fed back through different gestures and eye movements. The system layer can be informed of the face recognition error feedback information through the UI interface.

And in operation 204, adjusting a corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information.

The recognition control threshold is a threshold used for controlling the face recognition accuracy, and different types of recognition control thresholds corresponding to different face recognition algorithms. As with 2D face recognition algorithms, recognition control thresholds may include alignment thresholds, prosthesis thresholds, and the like. In one embodiment, when the face recognition algorithm is a network model algorithm, the recognition control threshold may be a parameter of the network model. The recognition control threshold may be a recognition control threshold in a 2D face recognition algorithm or a 3D face recognition algorithm.

Aiming at different face recognition error feedback information, the recognition control threshold value needs to be adjusted through different adjusting modes, and a specific adjusting algorithm can be determined according to the content of the face recognition error feedback information and the meaning of the recognition control threshold value. The adjustment principle is that when the face recognition error feedback information is the wrong acceptance recognition feedback information, the face recognition algorithm needs to be stricter, and the face of other users is prevented from being recognized as the face of the preset user by mistake. When the face recognition error feedback information is the false rejection recognition feedback information, the face recognition algorithm needs to be more relaxed, and the situation that the face cannot be successfully verified due to the fact that the correct face of the preset user cannot be successfully recognized under the influence of the environment or other factors is avoided.

And operation 206, performing face recognition according to the current face recognition algorithm after the recognition control threshold is adjusted.

The system confirms the face recognition as the false recognition according to the face recognition error feedback information, so that the threshold value is adjusted to reduce the number of times of false acceptance recognition and/or false rejection recognition. The result of the face recognition can be used for operations needing face recognition verification, such as unlocking, payment, starting or closing of an application program, permission setting and the like, the face recognition is carried out according to the current face recognition algorithm after the recognition control threshold value is adjusted, the accuracy of the face recognition is improved, and therefore the validity of the permission of each operation carried out according to the face recognition result is guaranteed.

According to the face recognition data processing method, the face recognition error feedback information is received through the mobile equipment, the corresponding recognition control threshold value in the current face recognition algorithm is adjusted according to the face recognition error feedback information, the face recognition is carried out according to the current face recognition algorithm after the recognition control threshold value is adjusted, the face recognition algorithm can be dynamically and intelligently automatically adjusted when the face recognition is wrong, the next continuous error recognition can be avoided, and the accuracy of the face recognition is improved.

In one embodiment, the mobile device receives face recognition error feedback information, comprising: and receiving face recognition error feedback information through a control displayed on a screen of the mobile equipment, wherein different controls transmit different face recognition error feedback information.

Specifically, when the face recognition is wrong, the control on the screen of the mobile device receives the operation of the user, different controls correspond to different trigger events, for example, the trigger event corresponding to the first control is transmission error acceptance recognition feedback information, and the trigger event corresponding to the second control is transmission error rejection recognition feedback information. The display positions of the controls can be different, for example, the controls for transmitting the error rejection identification feedback information, namely the unlocking failure can be displayed on a screen locking interface, so that the controls can be operated to transmit the error rejection identification feedback information under the condition of unlocking failure. The control for transmitting the error receiving identification feedback information of the successful error unlocking can be displayed on an interface after the successful unlocking, so that the control can be rapidly operated under the condition of successful error unlocking to transmit the error receiving identification feedback information.

As shown in fig. 3, which is a schematic diagram of a first control displayed on a screen lock interface in an embodiment, the first control 208 is configured to transmit false recognition rejection feedback information to the face recognition algorithm module when the person fails to unlock the screen lock interface. As shown in fig. 4, which is a schematic diagram of a second control displayed on a desktop after successful unlocking in an embodiment, the second control 210 is used for transmitting false acceptance recognition feedback information to the face recognition algorithm module when the unlocking by a non-user is successful. In one embodiment, the second control is in a hidden state, and the display of the second control can be triggered by a pull-down operation on the desktop.

In the embodiment, different face recognition error feedback information is transmitted to the system through different controls displayed on the UI interface, and the face recognition error feedback information is fed back quickly through operation acting on the controls, so that convenience and timeliness of face recognition error feedback information feedback are improved.

In one embodiment, the face recognition error feedback information is received through a control displayed on a screen of the mobile device, and the method comprises the following steps: and responding to the operation of a first control displayed on the screen of the mobile equipment, generating error acceptance identification feedback information, taking the error acceptance identification feedback information as face identification error feedback information, and receiving the error acceptance identification feedback information transmitted by the first control.

The operation acting on the first control displayed on the screen of the mobile device can be any self-defined operation such as clicking, touch time exceeding a preset threshold value, sliding operation and the like. The first control can be a control specially used for feeding back the error acceptance identification feedback information, and can also be a control with other functions. When the first control has multiple functions, generating error acceptance identification feedback information through the triggering of a preset operation corresponding to the received face identification error feedback information. If the leftward sliding on the first control indicates that face false acceptance recognition occurs, generating false acceptance recognition feedback information when the first control recognizes the leftward sliding operation.

The first control corresponds to a preset feedback value, the preset feedback value represents that the identification feedback information is received in error, if the transmission value is 1, the identification feedback information is received in error, when the first control recognizes the preset operation, a feedback value with the value of 1 is generated, the feedback value is transmitted to a face recognition system, and the face recognition system judges the adjustment direction of the identification control threshold value according to the size of the feedback value.

In one embodiment, generating false acceptance identification feedback information in response to an operation on a first control displayed on a screen of the mobile device comprises: the mobile equipment acquires a current face image, when the current face image is not matched with a preset unlocking face image actually, the current face recognition algorithm recognizes the current face image as a preset unlocking face by mistake, and when the screen of the mobile equipment is triggered to be unlocked successfully, wrong receiving recognition feedback information is generated by responding to operation acting on the first control.

Specifically, the current face recognition algorithm recognizes the current face image as the preset unlocking face by mistake, and the screen of the mobile device is triggered to be unlocked successfully, which shows that the unlocking is mistakenly unlocked because the face recognition algorithm recognizes the faces of other users as the preset unlocking faces input by the mobile device by mistake, and if the faces of the users A are input, the mobile phone is unlocked by the user B. The user A can operate the first control, and the first control generates error acceptance identification feedback information according to the operation of the user A.

In this embodiment, when the unlocking is successful incorrectly, the incorrect reception identification feedback information can be fed back quickly through the first control. It can be understood that the application scenario may not be limited to unlocking, and if the application scenario is another application scenario, for example, payment is performed through a human face, an application is entered through a human face, and the like, the false acceptance identification feedback information may be quickly fed back corresponding to different controls matched with the application scenario.

In one embodiment, the receiving of the face recognition error feedback information through a control displayed on the screen of the mobile device comprises: and responding to the operation of a second control displayed on the screen of the mobile equipment, generating error rejection identification feedback information, taking the error rejection identification feedback information as face identification error feedback information, and receiving the error rejection identification feedback information transmitted by the second control.

Specifically, the operation acting on the second control displayed on the screen of the mobile device may be any user-defined operation such as clicking, touching for a time period exceeding a preset threshold, sliding, and the like. The second control can be a control specially used for feeding back the error rejection identification feedback information, and can also be a control with other functions. And when the second control has multiple functions, generating error rejection identification feedback information by triggering a preset operation corresponding to the received face identification error rejection identification feedback information. And if the second control slides to the right to indicate that face false rejection recognition occurs, generating false rejection recognition feedback information when the second control recognizes the operation of sliding to the right.

In one embodiment, the control for feeding back the error rejection identification feedback information and the error acceptance identification feedback information is the same control, and different types of information are fed back through different operations of the control. If sliding to the left, the error acceptance identification feedback information is fed back, and if sliding to the right, the error rejection identification feedback information is fed back. By integrating the controls, the generation of the controls can be saved, and resources are saved.

The second control corresponds to a preset feedback value, the preset feedback value represents the error rejection identification feedback information, if the transmission value is 2, the first control generates a feedback value with the value of 2 when recognizing the preset operation, the feedback value is transmitted to the face recognition system, and the face recognition system judges the adjustment direction of the recognition control threshold value according to the size of the feedback value.

In one embodiment, generating the false rejection identification feedback information in response to an operation acting on a second control displayed on the screen of the mobile device comprises: the mobile equipment acquires a current face image, and when the current face image is actually matched with a preset unlocking face image, the current face image is mistakenly identified as a non-preset unlocking face by a current face identification algorithm, so that the screen of the mobile equipment is unlocked unsuccessfully, and mistaken rejection identification feedback information is generated in response to the operation acting on the second control.

Specifically, the current face recognition algorithm identifies the current face image as a non-preset unlocking face by mistake, so that the screen unlocking of the mobile device fails, which indicates that the unlocking is failed due to the fact that the face recognition algorithm cannot identify the preset unlocking face input by the mobile device, and if the face input by the user a is input, the user a cannot unlock the mobile device. The user A can operate the second control, and the second control generates the feedback information of the false acceptance rejection according to the operation of the user A.

In this embodiment, when the correct face unlocking fails, the error rejection recognition feedback information can be fed back quickly through the second control. It can be understood that the application scenario may not be limited to unlocking, and if the application scenario is another application scenario, for example, payment is performed through a human face, an application is entered through a human face, and the like, the error rejection recognition feedback information may be quickly fed back corresponding to different controls matched with the application scenario.

In one embodiment, when the current face image is actually matched with the preset unlocking face image, the time length for recognizing the current face image as the preset unlocking face by the current face recognition algorithm is obtained, the target control is determined according to the time length, and the error rejection recognition feedback information is generated in response to the operation acting on the target control.

Specifically, when the current face image is actually matched with the preset unlocking face image, if the face of the user a is input, the current face image is the face of the user a, the current face recognition algorithm recognizes the non-preset unlocking face from the current face image within the first time period, which indicates that the current face recognition algorithm cannot recognize the face of the user a within the first time period, and recognizes the current face image as the preset unlocking face at the second time, the difference between the second time and the starting time is the time period used by the current face recognition algorithm to recognize the current face image as the preset unlocking face. The shorter the duration is, the more sensitive the current face recognition algorithm is, and the higher the recognition degree is. And determining target controls according to the recognition duration, wherein different target controls are used for indicating the current face recognition algorithm to adjust the recognition control threshold value in different amplitudes.

In the embodiment, the recognition control threshold is adjusted differently through the recognition duration, so that the accuracy of adjusting the recognition control threshold is further improved.

In one embodiment, identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; operation 204, comprising: and when the face recognition error feedback information is the error acceptance recognition feedback information, increasing the comparison threshold value and/or reducing the prosthesis threshold value.

Specifically, the 2D face recognition algorithm mainly includes two thresholds, a comparison threshold and a prosthesis threshold. And comparing threshold values to control similarity thresholds of the input face and the face trying to be verified, wherein the lower the threshold value is, the easier the verification is successful, and if the face is verified to be unlocked, the lower the threshold value is, the easier the unlocking is. When the face recognition error feedback information is the error acceptance recognition feedback information, the current face recognition algorithm is easy to recognize other faces as the preset faces, and the similarity threshold is low, so that the comparison threshold needs to be improved. The comparison threshold is used as a threshold for determining the degree of authenticity of the face attempting to unlock, and the higher the threshold is, the easier it is to identify the false face as a true face. When the face recognition error feedback information is the error acceptance recognition feedback information, the current face recognition algorithm is easy to recognize the false face as the real face, and the threshold for the true degree is higher, so that the threshold of the false is required to be reduced. It will be appreciated that the increase in comparison threshold and the decrease in prosthesis threshold may be adjusted alternatively or both, and the magnitude of the adjustment may be matched or customized to the current face recognition algorithm.

In one embodiment, identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; operation 204, comprising: and when the face recognition error feedback information is the error rejection recognition feedback information, reducing the comparison threshold and/or improving the prosthesis threshold.

Specifically, when the face recognition error feedback information is the false rejection recognition feedback information, it is indicated that the face recognition algorithm is easy to fail to correctly input the face recognition, for example, the face of the user a is input by the mobile device, but when the user a is affected by illumination change, hair shielding, face angle deflection and other reasons, even if the user a is a face of the user a, the face cannot be recognized, and it is indicated that the similarity threshold is higher, and the comparison threshold needs to be reduced. When the face recognition error feedback information is the error rejection recognition feedback information, it indicates that the current face recognition algorithm is easy to be recognized as a false face by error even if the face is a real face, and the threshold for indicating the true degree is low, and the threshold of the false needs to be increased. It will be appreciated that the reduction of the comparison threshold and the increase of the prosthesis threshold may be adjusted alternatively or both, and the magnitude of the adjustment may be matched or customized to the current face recognition algorithm.

In one embodiment, prior to operation 204, as shown in fig. 5, further comprising:

in operation 302, a current recognition scene corresponding to the current face recognition is obtained, and a feedback recognition scene corresponding to the feedback information of the face recognition error is obtained.

Operation 304, determine when the current recognition scene matches the feedback recognition scene, if yes, go to operation 204, if not, do not adjust.

Specifically, the background pattern, the illuminance, the brightness, the shielding degree, the distance degree, the face definition and the like can be used as judgment factors of different scenes, a current scene judgment factor corresponding to the current face recognition is obtained by analyzing the currently acquired image, the current scene judgment factor is compared with a feedback scene judgment factor corresponding to the feedback recognition scene, whether the current recognition scene is matched with the feedback recognition scene or not is judged, and a specific matching algorithm can be defined. In one embodiment, when the similarity of more than a preset number of judgment factors exceeds a preset threshold, the current recognition scene is judged to be matched with the feedback recognition scene.

The image of the current camera frame and the image of the feedback identification scene frame can be directly judged and identified to be the same or similar scene through an image analysis algorithm, and if the images are the same or similar scene, the current identification scene is matched with the feedback identification scene.

In the embodiment, the recognition control threshold value is adjusted only when the current recognition scene is matched with the feedback recognition scene, so that the accuracy of adjusting the recognition control threshold value is further improved, and the wrong adjustment is avoided.

In one embodiment, as shown in fig. 6, the method further comprises:

and operation 402, obtaining a feedback identification scene corresponding to the face recognition error feedback information, and establishing a matching relationship between the feedback identification scene and the adjusted recognition control threshold.

Specifically, different scenes corresponding to the identification error feedback information fed back by the user are obtained, and the identification control threshold adjusted in the different scenes is obtained, so that the matching relationship between the different feedback identification scenes and the adjusted identification control threshold is established. For example, when face recognition is performed in both the scene a and the scene B, face recognition error feedback information is received, so that the scene a and the scene B are different feedback recognition scenes. The method comprises the steps of firstly adjusting an identification control threshold value under a scene A to obtain an adjusted target first identification control threshold value, secondly adjusting the identification control threshold value under a scene B to obtain an adjusted target second identification control threshold value, respectively establishing a matching relation between the scene A and the target first identification control threshold value, and establishing a matching relation between the scene B and the target second identification control threshold value.

In operation 404, a current recognition scene corresponding to the current face recognition is obtained, and a target feedback recognition scene matching the current recognition scene is determined.

Specifically, a current recognition scene corresponding to the current face recognition is obtained through analysis of a current collected image, and is compared with each feedback recognition scene in the matching relation to obtain a target feedback recognition scene matched with the current recognition scene, wherein the target feedback recognition scene is the same as or similar to the current recognition scene, and a specific matching algorithm can be customized.

And operation 406, acquiring a target identification control threshold corresponding to the target feedback identification scene according to the matching relationship.

Specifically, the recognition control threshold corresponding to each feedback recognition scene is the adjusted threshold matched with the scene, so that different recognition control thresholds are guaranteed to be used in different scenes, and the accuracy of face recognition in different scenes is improved. By acquiring the target identification control threshold corresponding to the target feedback identification scene, the target feedback identification scene is a scene matched with the current identification scene, so that the target identification control threshold is more suitable for the current identification scene.

In operation 408, face recognition is performed according to the current face recognition algorithm corresponding to the target recognition control threshold.

Specifically, the current face recognition algorithm is obtained according to the target recognition control threshold value, face recognition is carried out, the fact that the control threshold value which is adjusted before and matched with the current recognition scene is used in the current recognition scene is guaranteed, and therefore accuracy of face recognition in the current recognition scene is improved.

In one embodiment, as shown in FIG. 7, operation 204 comprises:

in operation 204a, an initial recognition control threshold corresponding to the current face recognition algorithm is obtained.

Specifically, the initial recognition control threshold is a recognition control threshold that is not adjusted, and for example, the recognition control threshold corresponding to the initial model corresponding to the current face recognition algorithm may be a recognition control threshold that is customized.

In operation 204b, an adjustment range of the recognition control threshold is obtained, and the corresponding recognition control threshold in the current face recognition algorithm is adjusted according to the face recognition error feedback information, so that the difference between the adjusted current recognition control threshold and the initial recognition control threshold is within the adjustment range of the recognition control threshold.

Specifically, the recognition control threshold adjustment range controls the recognition control threshold within a certain range, for example, in an embodiment, if the initial recognition control threshold is 10, the recognition control threshold adjustment range is [ -10-10], which indicates that the recognition control threshold can only be adjusted by 10 magnitude up or down, and if the recognition control threshold is beyond the recognition control threshold adjustment range after adjustment, which indicates that the adjustment is excessive. Because the threshold adjustment needs a certain range, the threshold adjustment cannot be too wide, otherwise, the false acceptance rate and the false rejection rate are seriously affected, the false acceptance rate is the probability of successful verification after the false face is subjected to algorithm identification, and the false rejection rate is the probability of failure verification after the correct face is subjected to algorithm identification. The adjustment of the threshold value is allowed to have slight fluctuation, but needs to be restored under appropriate conditions, because the error acceptance rate and the error rejection rate of the default threshold value belong to a relatively balanced state from the viewpoint of big data, and breaking the balance for a long time can introduce new problems under other scenes.

In the embodiment, the difference between the adjusted current recognition control threshold and the initial recognition control threshold is within the recognition control threshold adjustment range, so that the stability of face recognition is ensured.

As shown in fig. 8, the following describes the face recognition data processing method in detail with reference to a specific example. Firstly, suppose that the mobile device displays different feedback controls on different interfaces, the mobile device is in a screen locking state, a first control is displayed on an interface corresponding to the screen locking state, and a character 'can't be unlocked by oneself? And the first control is used for prompting the user to operate to feed back error rejection identification feedback information when the user cannot unlock the lock.

In operation 502, the mobile device receives feedback error rejection identification feedback information through the first control, and the user notifies the system layer of the error rejection identification feedback information through the UI interface. For example, if the face of the user a is entered and the user a cannot unlock the system, the first control receives the click operation of the user a, and the first control transmits the error rejection identification feedback information to the face identification algorithm module of the system layer.

In operation 504, when the system layer receives the false rejection identification feedback information, the comparison threshold is decreased, the prosthesis threshold is increased, and it is ensured that the adjusted comparison threshold and the adjusted prosthesis threshold are within the corresponding adjustment ranges, respectively.

In operation 506, when the user B successfully unlocks, a second control is displayed on the interface where the unlocking is successful, and a character "is not personally unlocked? And when the user is not personally unlocked, the user is prompted to operate the first control to feed back the error acceptance identification feedback information. The mobile device receives feedback error through the second control, receives recognition feedback information, for example, a face of the user A is input, a face of the user B passes verification, and the screen is unlocked successfully.

In operation 508, when the system layer receives the false acceptance identification feedback information, the comparison threshold is increased, the prosthesis threshold is decreased, and it is ensured that the adjusted comparison threshold and the adjusted prosthesis threshold are within the corresponding adjustment ranges, respectively.

In the embodiment, after the face is recognized by mistake, the recognition control threshold value can be adjusted by feeding back the defects of the algorithm to the system through the operation of the control, the threshold value is automatically adjusted through feedback information, the face unlocking is used by the same user in the same scene next time, the phenomenon of false recognition can be reduced, and the accuracy of face recognition is improved. And the software version does not need to be updated in the whole adjusting process, so that the development cost is reduced.

It should be understood that although the steps in the flowcharts of fig. 2, 5-8 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2, 5-8 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

The embodiment of the application also provides the mobile equipment. The mobile device includes a memory and a processor, the memory having stored therein computer-readable instructions that, when executed by the processor, cause the processor to perform the steps of: receiving face recognition error feedback information through the mobile equipment; adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information; and carrying out face recognition according to the current face recognition algorithm after the recognition control threshold value is adjusted.

In one embodiment, the mobile device receives face recognition error feedback information, comprising: and receiving face recognition error feedback information through a control displayed on the screen of the mobile equipment, wherein different controls transmit different face recognition error feedback information.

In one embodiment, the face recognition error feedback information is received through a control displayed on a screen of the mobile device, and the method comprises the following steps: responding to the operation of a first control displayed on a screen of the mobile equipment, generating error acceptance identification feedback information, and taking the error acceptance identification feedback information as face identification error feedback information; and receiving the error acceptance identification feedback information transmitted by the first control.

In one embodiment, generating the false acceptance identification feedback information in response to an operation acting on a first control displayed on a screen of the mobile device comprises: the mobile equipment acquires a current face image; and when the current face image is not matched with the preset unlocking face image actually, the current face recognition algorithm wrongly recognizes the current face image as the preset unlocking face and triggers the screen of the mobile equipment to be unlocked successfully, and wrongly-received recognition feedback information is generated in response to the operation acting on the first control.

In one embodiment, the face recognition error feedback information is received through a control displayed on a screen of the mobile device, and the method comprises the following steps: responding to the operation of a second control displayed on the screen of the mobile equipment, generating error rejection identification feedback information, and taking the error rejection identification feedback information as face identification error feedback information; and receiving the error rejection identification feedback information transmitted by the second control.

In one embodiment, generating the false rejection identification feedback information in response to an operation acting on a second control displayed on the screen of the mobile device comprises: acquiring a current face image through mobile equipment; and when the current face image is actually matched with the preset unlocking face image, the current face recognition algorithm wrongly recognizes the current face image as a non-preset unlocking face, so that the screen of the mobile equipment is failed to be unlocked, and the wrongly refused recognition feedback information is generated in response to the operation acting on the second control.

In one embodiment, identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information, wherein the method comprises the following steps: and when the face recognition error feedback information is the error acceptance recognition feedback information, increasing the comparison threshold value and/or reducing the prosthesis threshold value.

In one embodiment, identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information, wherein the method comprises the following steps: the mobile equipment acquires a current face image; and when the current face image is not matched with the preset unlocking face image actually, the current face recognition algorithm wrongly recognizes the current face image as the preset unlocking face and triggers the screen of the mobile equipment to be unlocked successfully, and wrongly-received recognition feedback information is generated in response to the operation acting on the first control.

In one embodiment, the receiving of the face recognition error feedback information through a control displayed on the screen of the mobile device comprises: responding to the operation of a second control displayed on the screen of the mobile equipment, generating error rejection identification feedback information, and taking the error rejection identification feedback information as face identification error feedback information; and receiving the error rejection identification feedback information transmitted by the second control.

In one embodiment, generating the false rejection identification feedback information in response to an operation acting on a second control displayed on the screen of the mobile device comprises: acquiring a current face image through mobile equipment; and when the current face image is actually matched with a preset unlocking face image, the current face recognition algorithm wrongly recognizes the current face image as a non-preset unlocking face, so that the screen of the mobile equipment is failed to be unlocked, and the mistaken rejection recognition feedback information is generated in response to the operation acting on the second control.

In one embodiment, identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information, wherein the method comprises the following steps: and when the face recognition error feedback information is the error rejection recognition feedback information, reducing the comparison threshold and/or improving the prosthesis threshold.

In one embodiment, the processor performs the steps of: acquiring a current recognition scene corresponding to the current face recognition, and acquiring a feedback recognition scene corresponding to the face recognition error feedback information; and when the current recognition scene is matched with the feedback recognition scene, adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information.

In one embodiment, the processor performs the steps of: acquiring a feedback identification scene corresponding to the face identification error feedback information, and establishing a matching relation between the feedback identification scene and the adjusted identification control threshold; acquiring a current recognition scene corresponding to current face recognition, and determining a target feedback recognition scene matched with the current recognition scene; acquiring a target identification control threshold corresponding to a target feedback identification scene according to the matching relation; and carrying out face recognition according to the current face recognition algorithm corresponding to the target recognition control threshold.

In one embodiment, adjusting the corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information includes: acquiring an initial recognition control threshold corresponding to a current face recognition algorithm; acquiring an adjustment range of an identification control threshold; and adjusting the corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information, so that the difference between the adjusted current recognition control threshold and the initial recognition control threshold is in the recognition control threshold adjustment range.

Fig. 7 is a block diagram showing a configuration of a face recognition data processing apparatus according to an embodiment. As shown in fig. 5, a face recognition data processing apparatus includes a receiving module 602, an adjusting module 604, and a recognition module 606. Wherein:

the receiving module 602 is configured to receive the face recognition error feedback information through the mobile device.

The adjusting module 604 is configured to adjust a corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information.

The recognition module 606 is configured to perform face recognition according to the current face recognition algorithm after the recognition control threshold is adjusted.

In one embodiment, the receiving module 602 is further configured to receive the face recognition error feedback information through a control displayed on a screen of the mobile device, where different controls deliver different face recognition error feedback information.

In one embodiment, the receiving module 602 is further configured to generate error acceptance recognition feedback information in response to an operation performed on a first control displayed on a screen of the mobile device, use the error acceptance recognition feedback information as face recognition error feedback information, and receive the error acceptance recognition feedback information delivered by the first control.

In an embodiment, the receiving module 602 is further configured to obtain a current face image through the mobile device, and when the current face image is not actually matched with the preset unlocking face image, the current face recognition algorithm erroneously recognizes the current face image as a preset unlocking face, and triggers a screen of the mobile device to be successfully unlocked, generate false acceptance recognition feedback information in response to an operation performed on the first control.

In one embodiment, the receiving module 602 is further configured to generate error rejection recognition feedback information in response to an operation performed on a second control displayed on a screen of the mobile device, and use the error rejection recognition feedback information as the face recognition error feedback information; and receiving the error rejection identification feedback information transmitted by the second control.

In an embodiment, the receiving module 602 is further configured to obtain a current face image through the mobile device, and when the current face image is actually matched with the preset unlocking face image, the current face recognition algorithm erroneously recognizes the current face image as a non-preset unlocking face, so that the screen of the mobile device fails to be unlocked, and generates the false rejection recognition feedback information in response to an operation performed on the second control.

In one embodiment, identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; the adjusting module 604 is further configured to increase the comparison threshold and/or decrease the prosthesis threshold when the face recognition error feedback information is the false acceptance recognition feedback information.

In one embodiment, identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; the adjusting module 604 is further configured to decrease the comparison threshold and/or increase the prosthesis threshold when the face recognition error feedback information is the false rejection recognition feedback information.

In one embodiment, the apparatus further comprises:

and a scene matching module, configured to obtain a current recognition scene corresponding to current face recognition, obtain a feedback recognition scene corresponding to feedback information of face recognition errors, and enter the adjusting module 604 when the current recognition scene is matched with the feedback recognition scene.

In one embodiment, the apparatus further comprises:

and the matching relation module is used for acquiring a feedback identification scene corresponding to the face identification error feedback information and establishing the matching relation between the feedback identification scene and the adjusted identification control threshold value.

And the scene adjusting module is used for acquiring a current recognition scene corresponding to the current face recognition, determining a target feedback recognition scene matched with the current recognition scene, and acquiring a target recognition control threshold corresponding to the target feedback recognition scene according to the matching relation.

The recognition module 606 is further configured to perform face recognition according to a current face recognition algorithm corresponding to the target recognition control threshold.

In one embodiment, the adjusting module 604 is further configured to obtain an initial recognition control threshold corresponding to the current face recognition algorithm; acquiring an adjustment range of an identification control threshold; and adjusting the corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information, so that the difference between the adjusted current recognition control threshold and the initial recognition control threshold is within the recognition control threshold adjustment range.

The implementation of each module in the face recognition data processing apparatus provided in the embodiments of the present application may be in the form of computer readable instructions. The computer readable instructions may be executed on a terminal or a server. Program modules of the computer readable instructions may be stored on the memory of the terminal or the server. The computer readable instructions, when executed by a processor, perform the steps of the method described in the embodiments of the present application.

Fig. 10 is a schematic internal structure diagram of a mobile device in one embodiment. As shown in fig. 10, the mobile device includes a processor and a memory connected by a system bus. Wherein, the processor is used for providing calculation and control capability and supporting the operation of the whole electronic equipment. The memory may include a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer readable instructions. The computer readable instructions can be executed by a processor to implement a face recognition data processing method provided in the following embodiments. The internal memory provides a cached execution environment for the operating system computer-readable instructions in the non-volatile storage medium. The mobile device may be a mobile phone, a tablet computer, or a personal digital assistant or a wearable device, etc.

The embodiment of the application also provides the mobile equipment. The mobile device may be any terminal device including a mobile phone, a tablet computer, a PDA (Personal Digital Assistant), a Point of Sales (POS), a vehicle-mounted computer, a wearable device, and the like, and the mobile device is taken as the mobile phone as an example.

The embodiment of the application also provides the mobile equipment. The mobile device includes an Image Processing circuit, which may be implemented using hardware and/or software components, and may include various Processing units defining an ISP (Image Signal Processing) pipeline. FIG. 11 is a schematic diagram of an image processing circuit in one embodiment. As shown in fig. 11, for convenience of explanation, only aspects of the image processing technology related to the embodiments of the present application are shown.

As shown in fig. 11, the image processing circuit includes an ISP processor 740 and a control logic 750. The image data captured by the imaging device 710 is first processed by the ISP processor 740, and the ISP processor 740 analyzes the image data to capture image statistics that may be used to determine and/or control one or more parameters of the imaging device 710. The imaging device 710 may include a camera having one or more lenses 712 and an image sensor 714. The image sensor 714 may include an array of color filters (e.g., Bayer filters), and the image sensor 714 may acquire light intensity and wavelength information captured with each imaging pixel of the image sensor 714 and provide a set of raw image data that may be processed by the ISP processor 740. The sensor 720 (e.g., a gyroscope) may provide parameters of the acquired image processing (e.g., anti-shake parameters) to the ISP processor 740 based on the type of sensor 720 interface. The sensor 720 interface may utilize a SMIA (Standard Mobile Imaging Architecture) interface, other serial or parallel camera interfaces, or a combination of the above.

In addition, image sensor 714 may also send raw image data to sensor 720, sensor 720 may provide raw image data to ISP processor 740 based on the type of sensor 720 interface, or sensor 720 may store raw image data in image memory 730.

ISP processor 740 processes the raw image data pixel by pixel in a variety of formats. For example, each image pixel may have a bit depth of 8, 10, 12, or 14 bits, and ISP processor 740 may perform one or more image processing operations on the raw image data, collecting statistical information about the image data. Wherein the image processing operations may be performed with the same or different bit depth precision.

ISP processor 740 may also receive image data from image memory 730. For example, sensor 720 interface sends raw image data to image memory 730, and the raw image data in image memory 730 is then provided to ISP processor 740 for processing. The image Memory 730 may be a portion of a Memory device, a storage device, or a separate dedicated Memory within an electronic device, and may include a DMA (Direct Memory Access) feature.

ISP processor 740 may perform one or more image processing operations, such as temporal filtering, upon receiving raw image data from image sensor 714 interface or from sensor 720 interface or from image memory 730. The processed image data may be sent to image memory 730 for additional processing before being displayed. ISP processor 740 receives processed data from image memory 730 and performs image data processing on the processed data in the raw domain and in the RGB and YCbCr color spaces. The image data processed by ISP processor 740 may be output to display 770 for viewing by a user and/or further processed by a Graphics Processing Unit (GPU). Further, the output of ISP processor 740 may also be sent to image memory 730 and display 770 may read image data from image memory 730. In one embodiment, image memory 730 may be configured to implement one or more frame buffers. In addition, the output of the ISP processor 740 may be transmitted to the encoder/decoder 760 for encoding/decoding image data. The encoded image data may be saved and decompressed before being displayed on the display 770 device. The encoder/decoder 760 may be implemented by a CPU or GPU or coprocessor.

The statistical data determined by ISP processor 740 may be sent to control logic 750 unit. For example, the statistical data may include image sensor 714 statistics such as auto-exposure, auto-white balance, auto-focus, flicker detection, black level compensation, lens 712 shading correction, and the like. Control logic 750 may include a processor and/or microcontroller that executes one or more routines (e.g., firmware) that may determine control parameters of imaging device 710 and control parameters of ISP processor 740 based on the received statistical data. For example, the control parameters of imaging device 710 may include sensor 720 control parameters (e.g., gain, integration time for exposure control, anti-shake parameters, etc.), camera flash control parameters, lens 712 control parameters (e.g., focal length for focusing or zooming), or a combination of these parameters. The ISP control parameters may include gain levels and color correction matrices for automatic white balance and color adjustment (e.g., during RGB processing), as well as lens 712 shading correction parameters.

In an embodiment of the present application, the ISP processor 740 included in the mobile device implements the steps of the face recognition data processing method when executing the computer readable instructions stored on the memory.

The embodiment of the application also provides a computer readable storage medium. One or more non-transitory computer-readable storage media containing computer-readable instructions that, when executed by a processor, perform the steps of: receiving face recognition error feedback information through the mobile equipment; adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information; and carrying out face recognition according to the current face recognition algorithm after the recognition control threshold value is adjusted.

In one embodiment, the computer readable instructions are executable by a processor to: acquiring a current recognition scene corresponding to the current face recognition, and acquiring a feedback recognition scene corresponding to the face recognition error feedback information; and when the current recognition scene is matched with the feedback recognition scene, adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information.

In one embodiment, the computer readable instructions are executable by a processor to: acquiring a feedback identification scene corresponding to the face identification error feedback information, and establishing a matching relation between the feedback identification scene and the adjusted identification control threshold; acquiring a current recognition scene corresponding to current face recognition, and determining a target feedback recognition scene matched with the current recognition scene; acquiring a target identification control threshold corresponding to a target feedback identification scene according to the matching relation; and carrying out face recognition according to the current face recognition algorithm corresponding to the target recognition control threshold.

The embodiment of the application also provides a computer readable instruction product. A computer readable instruction product containing instructions which, when run on a computer, cause the computer to perform a method of face recognition data processing.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by computer readable instructions to instruct associated hardware, and that the programs can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. As used herein, any reference to memory, storage, database or other medium may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM), which acts as external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), synchronous Link (Synchlink) DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and bus dynamic RAM (RDRAM).

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

A face recognition data processing method is characterized by comprising the following steps:

the mobile equipment receives face recognition error feedback information;

adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information; and

and carrying out face recognition according to the current face recognition algorithm after the recognition control threshold value is adjusted.
The method of claim 1, wherein the mobile device receives face recognition error feedback information, comprising:

and receiving face recognition error feedback information through a control displayed on the screen of the mobile equipment, wherein different controls transmit different face recognition error feedback information.
The method of claim 2, wherein receiving the face recognition error feedback information through a control displayed on the screen of the mobile device comprises:

responding to the operation of a first control displayed on the screen of the mobile equipment, generating error acceptance identification feedback information, and taking the error acceptance identification feedback information as the face identification error feedback information;

and receiving the error acceptance identification feedback information transmitted by the first control.
The method of claim 3, wherein generating false acceptance identification feedback information in response to the operation of the first control displayed on the screen of the mobile device comprises:

the mobile equipment acquires a current face image;

and when the current face image is not matched with a preset unlocking face image actually, the current face recognition algorithm wrongly recognizes the current face image as a preset unlocking face, and when the screen of the mobile equipment is successfully unlocked, the wrong receiving recognition feedback information is generated in response to the operation acting on the first control.
The method of claim 2, wherein receiving the face recognition error feedback information through a control displayed on the screen of the mobile device comprises:

responding to the operation of a second control displayed on the screen of the mobile equipment, generating error rejection identification feedback information, and taking the error rejection identification feedback information as the face identification error feedback information;

and receiving the error rejection identification feedback information transmitted by the second control.
The method of claim 5, wherein generating error rejection identification feedback information in response to an operation of a second control displayed on the screen of the mobile device comprises:

the mobile equipment acquires a current face image;

and when the current face image is actually matched with a preset unlocking face image, the current face recognition algorithm wrongly recognizes the current face image as a non-preset unlocking face, so that the screen of the mobile equipment is failed to be unlocked, and the mistaken rejection recognition feedback information is generated in response to the operation acting on the second control.
The method of claim 1, wherein the identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; the adjusting of the corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information comprises:

and when the face recognition error feedback information is the error acceptance recognition feedback information, increasing the comparison threshold value and/or reducing the prosthesis threshold value.
The method of claim 1, wherein the identifying a control threshold comprises at least one of an alignment threshold, a prosthesis threshold; the adjusting of the corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information comprises:

and when the face recognition error feedback information is the error rejection recognition feedback information, reducing the comparison threshold and/or improving the prosthesis threshold.
The method according to claim 1, wherein before adjusting the corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information, the method further comprises:

acquiring a current recognition scene corresponding to current face recognition, and acquiring a feedback recognition scene corresponding to the face recognition error feedback information;

and when the current recognition scene is matched with the feedback recognition scene, adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information.
The method of claim 1, further comprising:

acquiring a feedback identification scene corresponding to the face identification error feedback information, and establishing a matching relation between the feedback identification scene and the adjusted identification control threshold;

acquiring a current recognition scene corresponding to current face recognition, and determining a target feedback recognition scene matched with the current recognition scene;

acquiring a target identification control threshold corresponding to the target feedback identification scene according to the matching relation;

and carrying out face recognition according to the current face recognition algorithm corresponding to the target recognition control threshold.
The method of claim 1, wherein the adjusting the corresponding recognition control threshold in the current face recognition algorithm according to the face recognition error feedback information comprises:

acquiring an initial recognition control threshold corresponding to the current face recognition algorithm;

acquiring an adjustment range of an identification control threshold;

and adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information, so that the difference between the adjusted current recognition control threshold value and the initial recognition control threshold value is within the recognition control threshold value adjustment range.
A face recognition data processing apparatus, comprising:

the receiving module is used for receiving the face recognition error feedback information through the mobile equipment;

the adjusting module is used for adjusting a corresponding recognition control threshold value in the current face recognition algorithm according to the face recognition error feedback information;

and the recognition module is used for carrying out face recognition according to the current face recognition algorithm after the recognition control threshold value is adjusted.
A mobile device comprising a memory and a processor, the memory having stored therein computer-readable instructions that, when executed by the processor, cause the processor to perform the steps of the method of any of claims 1 to 11.
A computer readable storage medium having computer readable instructions stored thereon, which, when executed by a processor, cause the processor to perform the steps of the method of any one of claims 1 to 11.