Face recognition processing method, device, terminal and readable storage medium

Abstract

The present application is about a face recognition processing method, device, terminal and readable storage medium, belonging to the field of biometrics. The method includes: receiving a face recognition service request, the face recognition service request is set with a recognition accuracy requirement; obtaining a face recognition video stream in the face recognition process according to the face recognition service request; identifying occluders on the face recognition video stream; in response to the occlusion of the face recognition video stream by an occluder, determining a face recognition image from the face recognition video stream according to a selection strategy corresponding to the recognition accuracy requirement; sending the face recognition image to a server, and obtaining a first recognition result fed back by the server. When face recognition is required, the image used for face recognition is determined according to different recognition accuracy requirements, which solves the problem that the same face recognition processing method cannot be applied in different application scenarios, and can be flexibly adjusted according to the face recognition requirements of different scenarios, thereby improving the flexibility of face recognition processing.

Description

Technical Field

[0001] The present application relates to the field of biometrics, and in particular to a face recognition processing method, device, terminal and readable storage medium. Background Art

[0002] Facial recognition is a biometric technology that uses facial feature information to verify identity. Facial recognition technology is widely used in various fields. It is used in clocking in and out, taking intercity rail transit, and biometric payment.

[0003] In face recognition scenarios, it is common to see users with their faces obstructed. In related technologies, some application scenarios use recognition devices that can scan the face while wearing a mask. The processing method is to reduce the weight of mouth recognition during the recognition process.

[0004] This recognition device only optimizes the problem of mask occlusion, and the recognition device cannot be migrated and applied to other face recognition scenarios with different needs. It can only perform single face recognition with mouth occlusion. Summary of the invention

[0005] The embodiments of the present application provide a face recognition processing method, device, terminal and readable storage medium, which can improve the flexibility of face recognition processing. The technical solution is as follows:

[0006] On the one hand, a face recognition processing method is provided, the method comprising:

[0007] receiving a face recognition service request, wherein the face recognition service request is provided with a recognition accuracy requirement;

[0008] Obtaining a face recognition video stream during the face recognition process according to the face recognition service request;

[0009] Performing occlusion recognition on the face recognition video stream;

[0010] In response to the face recognition video stream being blocked by an obstruction, determining a face recognition image from the face recognition video stream using a selection strategy corresponding to the recognition accuracy requirement;

[0011] The face recognition image is sent to a server to obtain a first recognition result fed back by the server.

[0012] On the other hand, a face recognition processing device is provided, the device comprising:

[0013] A receiving module, used to receive a face recognition service request, wherein the face recognition service request is provided with a recognition accuracy requirement;

[0014] An acquisition module, used for acquiring a face recognition video stream in a face recognition process according to the face recognition service request;

[0015] A recognition module, used for performing occlusion recognition on the face recognition video stream;

[0016] A determination module, configured to determine a face recognition image from the face recognition video stream according to a selection strategy corresponding to the recognition accuracy requirement in response to the face recognition video stream being blocked by an obstruction;

[0017] The sending module is used to send the face recognition image to the server to obtain the first recognition result fed back by the server.

[0018] In an optional embodiment, the recognition accuracy requirement includes a first accuracy requirement and a second accuracy requirement;

[0019] The determining module comprises:

[0020] a determining unit, configured to determine the face recognition image from the face recognition video stream in a manner of ignoring occlusion of a preset part in response to the recognition accuracy requirement meeting a first accuracy requirement;

[0021] The determination unit is further configured to, in response to the recognition accuracy requirement meeting the second accuracy requirement, determine a video frame without occlusion from the face recognition video stream as the face recognition image.

[0022] In an optional embodiment, the determination module is further used to determine the occluded portion in the face image in the video frame in response to the presence of occlusion by the occluder in the face recognition video stream.

[0023] In an optional embodiment, the preset part occlusion includes at least one of mouth occlusion, forehead occlusion, eye occlusion, and ear occlusion, and the first accuracy requirement further corresponds to requirement information, and the requirement information includes that the preset part occlusion is allowed;

[0024] The determining unit is further configured to, in response to the requirement information that the mouth occlusion is allowed, determine a video frame in the face recognition video stream where the mouth occlusion is ignored as the face recognition image;

[0025] The determining unit is further configured to, in response to the requirement information that the forehead occlusion is allowed, determine a video frame in the face recognition video stream ignoring the forehead occlusion as the face recognition image;

[0026] The determining unit is further configured to, in response to the requirement information that the eye occlusion is allowed, determine a video frame in the face recognition video stream where the eye occlusion is ignored as the face recognition image;

[0027] The determination unit is further configured to determine, in response to the requirement information that the ear occlusion is allowed, a video frame in the face recognition video stream that ignores the ear occlusion as the face recognition image.

[0028] In an optional embodiment, the acquisition module is further used to acquire verification information in response to the first recognition result corresponding to unsuccessful face recognition, wherein the verification information corresponds to the current account information;

[0029] The sending module is further used to send the verification information to the server;

[0030] The receiving module is further used to receive a second recognition result fed back by the server according to the verification information.

[0031] In an optional embodiment, the device further comprises:

[0032] The display module is configured to display a prompt message in response to the second recognition result corresponding to unsuccessful face recognition, wherein the prompt message is used to prompt to eliminate the occlusion.

[0033] In an optional embodiment, the acquisition module is further used to perform quality analysis on the video frame in the face recognition video stream to obtain a quality analysis result corresponding to the video frame;

[0034] The determination module is further used to determine whether the face image in the video frame is blocked according to the quality analysis result.

[0035] In an optional embodiment, the sending module is also used to send the recognition accuracy requirement corresponding to the face recognition service request to the server, and the server is used to determine a target comparison strategy based on the recognition accuracy requirement, and the target comparison strategy is used to perform face recognition processing on the face recognition image; the target comparison strategy includes at least one of a comparison strategy that allows the existence of facial occlusion and a comparison strategy that prohibits the existence of facial occlusion.

[0036] On the other hand, a terminal is provided, comprising a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement a face recognition processing method as described in any of the above-mentioned embodiments of the present application.

[0037] On the other hand, a computer-readable storage medium is provided, characterized in that at least one instruction, at least one program, a code set or an instruction set is stored in the computer-readable storage medium, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by a processor to implement a face recognition processing method as described in any of the above-mentioned embodiments of the present application.

[0038] On the other hand, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes any of the face recognition processing methods described in the above embodiments.

[0039] The beneficial effects brought by the technical solution provided by the embodiment of the present application include at least:

[0040] When the terminal receives a face recognition service request, it determines the corresponding selection strategy according to the recognition accuracy requirement corresponding to the current face recognition service request, determines the face recognition image from the face recognition video stream according to the corresponding selection strategy, and then uploads the face recognition image to the server. The server completes the face recognition processing process, and finally the terminal obtains the recognition result fed back by the server. This solves the problem that the same face recognition processing method cannot be applied in different application scenarios. It can be flexibly adjusted according to the face recognition requirements of different scenarios, thereby improving the flexibility of face recognition processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 It is a schematic diagram of an implementation environment provided by an exemplary embodiment of the present application;

[0043] Figure 2 is a flowchart of a face recognition processing method provided by an exemplary embodiment of the present application;

[0044] Figure 3 is a schematic diagram of a face occlusion situation provided by an exemplary embodiment of the present application;

[0045] Figure 4 This is a secondary identification interface diagram provided by an exemplary embodiment of the present application;

[0046] Figure 5 is a secondary identification interface diagram provided by another exemplary embodiment of the present application;

[0047] Figure 6 is a flowchart of a face recognition processing method provided by another exemplary embodiment of the present application;

[0048] Figure 7 is a flowchart of a face recognition processing method provided by another exemplary embodiment of the present application;

[0049] Figure 8 is a flow chart of the interaction between a terminal and a server provided by an exemplary embodiment of the present application;

[0050] Fig. 9 is a block diagram of a face recognition processing device provided by an exemplary embodiment of the present application;

[0051] Fig.10 is a block diagram of a face recognition processing device provided by another exemplary embodiment of the present application;

[0052] Fig.11 It is a structural block diagram of a server provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0054] First, the application scenarios involved in the embodiments of the present application are introduced.

[0055] With the continuous development of biometric technology, the application of various biometric technologies is flooding our daily life, especially face recognition technology. In practical applications, such as the ID card recognition system in the access control system or attendance system, e-commerce or e-government applications, users can verify the unity of digital identity and real identity through the corresponding face recognition service. At present, the same device only supports fixed face recognition processing methods, that is, it cannot be migrated and used in multiple scenarios, such as the face recognition system in the ordinary access control system cannot be migrated and used in the face recognition ticket verification system of rail transit; or in the same application, different functional modules that require face recognition services all use the same fixed face recognition processing method, or multiple face recognition processing methods are designed according to different functional modules, and the various processing methods are not related to each other, and the correlation is not high, and the flexibility of migration and use of face recognition processing methods is not high.

[0056] The face recognition processing method in the embodiment of the present application can be applied in different application scenarios. For example, the corresponding device of the embodiment of the present application can be applied to the face recognition system in the ordinary access control system, and its recognition accuracy requirement is relatively low. When the terminal device captures the image for face recognition processing, it can ignore the occlusion of some parts, reduce the preferred requirements of the face recognition image, improve the efficiency of image acquisition, and facilitate users to quickly pass through the access control system; the device can also be migrated and applied to the face recognition ticket inspection system of rail transportation. During the processing, its recognition accuracy requirement is relatively high. When the terminal device captures the image for face recognition processing, it selects the image without occlusion of the face; it can also be migrated and applied to the face scanning security inspection system of subway transportation. Taking the current need to wear a mask when going out as an example, when the terminal device captures the image for face recognition processing, it can ignore the occlusion of the mouth, or in winter, pedestrians wear more hats, then ignore the occlusion of the forehead. The corresponding device of the embodiment of the present application can also be applied to different functional modules in the same application. Taking the payment application as an example, when logging in to an account, face recognition verification is used as an auxiliary verification method after entering the password. Its recognition accuracy requirement is relatively low. When capturing images for face recognition processing, a method of ignoring the occlusion of some parts can be used; and when using face recognition to verify identity and perform payment operations, when capturing images for face recognition processing, select images without facial occlusion. The face recognition processing method provided by the embodiment of the present application can be migrated and applied in different scenarios, which improves the flexibility of face recognition processing.

[0057] Secondly, the implementation environment of the embodiments of the present application is described.

[0058] For illustration, please refer to Figure 1 The implementation environment includes a terminal 110 and a server 120.

[0059] The terminal 110 may be an electronic device such as a mobile phone, a tablet computer, a multimedia player, a wearable device, a desktop computer, a biometric integrated machine, etc. Optionally, an application for face recognition is installed in the terminal 110, which may be implemented as an independent application, or as a small program in a host application, or as a functional module in an independent application. The terminal 110 is used to capture the face recognition video stream and pre-process it, and determine the face recognition image from the face recognition video stream according to the selection strategy corresponding to the recognition accuracy requirement, and the face recognition image is processed by the server 120.

[0060] The server 120 is used to provide face recognition services to the terminal 110. Optionally, the server 120 is a background server corresponding to the above-mentioned application for implementing face recognition. The server 120 can be a single server, or a server cluster composed of several servers, or a cloud computing service center. After receiving the face recognition image sent by the terminal 110, the server 120 processes it to obtain a first recognition result, and the server 120 returns the first recognition result to the terminal 110. When the first recognition result is a recognition failure, the server 120 receives the verification information returned by the terminal 110, and the server 120 performs re-verification based on the verification information to obtain a second recognition result, and feeds it back to the terminal 110.

[0061] The server 120 may establish a communication connection with the terminal 110 via a network. The network may be a wireless network or a wired network.

[0062] Please refer to Figure 2 , which shows a flow chart of a face recognition processing method shown in an embodiment of the present application. The method can be applied to Figure 1 In the terminal in the implementation environment shown. The method may include the following steps:

[0063] Step 201: Receive a face recognition service request.

[0064] The face recognition service request is a request generated when the terminal needs to call the face recognition service to authenticate the current operator before achieving the target function. For example, when the face recognition gate senses that someone needs to pass through, the face recognition terminal on the gate receives the face recognition service request; or, when the payment application in the mobile phone receives the user's payment operation for a certain order, it needs to call the face recognition service to verify the user's identity, and the face recognition service module in the application receives the face recognition service request.

[0065] In the embodiment of the present application, the face recognition service request is provided with a recognition accuracy requirement, which is the recognition accuracy of the current face recognition service corresponding to the target function requirement. Schematically, the recognition accuracy requirement can be determined according to the security requirement or response time requirement of the target function, wherein the recognition accuracy requirement is positively correlated with the security requirement, i.e., the higher the security requirement of the target function, the higher the recognition accuracy requirement, and the recognition accuracy requirement is positively correlated with the response time requirement, i.e., the longer the allowed response time of the target function, the higher the recognition accuracy requirement. For example, taking a payment application as an example, in this application, the security requirements for payment operations are relatively high, and the security requirements for login accounts are relatively low; taking payment operations as an example, the payment response speed requirements for "second kill" activities are high, and the allowed response time is short, while the payment response speed requirements for ordinary shopping are relatively low, and the allowed response time is long.

[0066] Step 202: Obtain a face recognition video stream during the face recognition process according to a face recognition service request.

[0067] After receiving the face recognition service request, the terminal needs to call the device's camera to obtain video data for face recognition. The video data can be a complete recognition video captured by the camera, or the video data can also be a video stream captured by the camera during the face recognition process.

[0068] After the terminal obtains the face recognition video stream, it needs to preprocess the video stream. Schematically, the preprocessing process includes decomposing the face recognition video stream frame by frame to obtain multiple video frames, and filtering out the video frames in which there are no face images or the face images are blurred; when there are more than one face images in the video frame, the preprocessing process also includes determining the target face image to be identified in the video frame, and filtering out the video frames with non-target face interference.

[0069] Step 203: perform occlusion recognition on the face recognition video stream.

[0070] In some application scenarios, the face image captured by the camera may be blocked. Please refer to Figure 3 As shown in image 301, the user may wear a hat, mask or other obstructions that may block facial features, which may affect face recognition. In one example, the types of obstructions can be roughly divided into two categories: head obstruction 302 and mouth obstruction 303, or both types of obstructions exist at the same time 304.

[0071] In an embodiment of the present application, a quality analysis is performed on video frames in a face recognition video stream to obtain quality analysis results corresponding to the video frames. Schematically, the quality analysis results include face image screening in the corresponding video frames, that is, screening out video frames with faces in the video frames; face registration, that is, locating the key point coordinates of the facial features in the video frames containing the faces, wherein the number of key points is a pre-set fixed value, schematically, it can be 5 points, 68 points, 90 points, etc.; face feature extraction, that is, converting the video frames containing the faces into a string of fixed-length values, wherein the input is the video frames and the key point coordinates, and the preset face feature extraction algorithm outputs a numerical string as a feature, which is provided to the subsequent steps.

[0072] According to the quality analysis results, it can be determined whether the face image in the video frame is blocked. Schematically, after performing face registration, that is, locating the key points of the face, the image is divided into multiple parts, and the feature space of each part is analyzed respectively, and it is determined whether the part is blocked according to the analysis results.

[0073] Step 204 , in response to the face recognition video stream being blocked by an occluder, a face recognition image is determined from the face recognition video stream using a selection strategy corresponding to the recognition accuracy requirement.

[0074] Optionally, the selection strategy includes whether to ignore the occlusion of a preset part to optimize the video frames in the face recognition video stream. Schematically, the video frames obtained after preprocessing include image A, image B, and image C, where the mouth in image A is occluded, the eyes in image B are occluded, and the forehead in image C is occluded. If the selection strategy corresponding to the current recognition accuracy requirement is to ignore the occlusion of the mouth for optimization, image A is determined to be a face recognition image and sent to the server by the terminal.

[0075] In an embodiment of the present application, the recognition accuracy requirement includes a first accuracy requirement and a second accuracy requirement, wherein the recognition accuracy corresponding to the first accuracy requirement is lower than the recognition accuracy corresponding to the second accuracy requirement.

[0076] After the terminal obtains the face recognition video stream and pre-processes it, it needs to perform a optimization operation on its video frame. The purpose of this optimization operation is to perform quality screening and find a suitable image for face recognition, which is mainly determined based on five features: occlusion range, blur range, illumination range, posture angle, and face integrity. Among them, when performing optimization for different recognition accuracy requirements, different optimization results can be achieved by changing the thresholds of different features. Optionally, the occlusion range threshold allowed in the selection strategy corresponding to the first accuracy requirement is greater than the occlusion range threshold allowed in the selection strategy corresponding to the second accuracy requirement. For example, see Table 1:

[0077] Table 1

[0078]

[0079]

[0080] Optionally, the selection strategy corresponding to the first accuracy requirement allows for occlusion of a preset part, while the selection strategy corresponding to the second accuracy requirement does not allow for any occlusion, for example, see Table 2:

[0081] Table 2

[0082]

[0083] Optionally, the face integrity threshold in the selection strategy corresponding to the first precision requirement is smaller than the face integrity threshold in the selection strategy corresponding to the second precision requirement. For example, 1 represents that there is a complete face in the image, and 0 represents that there is no face in the image. In the selection strategy corresponding to the first precision requirement, the face integrity threshold is 0.7, and in the selection strategy corresponding to the second precision requirement, the face integrity threshold is 1. That is, when the precision requirement corresponding to the face recognition service request is not high, even if the face image in the video captured by the camera for face recognition is incomplete, it can be directly recognized.

[0084] Optionally, the selection strategy includes using different facial feature recognition weights when calculating the overall facial quality score for different accuracy requirements during face recognition processing, wherein the overall facial quality score is used to evaluate the similarity between the current face recognition image and a pre-stored face recognition image for identity verification. Schematically, in the selection strategy corresponding to the first accuracy requirement, the corresponding weights of each facial feature are {mouth, 0; forehead, 0.25; ears 0.25; cheeks; 0.25; eyes; 0.25}, and in the selection strategy corresponding to the second accuracy requirement, the corresponding weights of each facial feature are {mouth, 0.2; forehead, 0.2; ears 0.2; cheeks; 0.2; eyes; 0.2}.

[0085] Step 205: Send the face recognition image to the server and obtain a first recognition result fed back by the server.

[0086] In the embodiment of the present application, the terminal sends the obtained face recognition image to the server, which is used to perform preset processing on the face recognition image and feed back the obtained recognition result to the terminal. Indicatively, the preset processing is liveness detection, where liveness detection is used to verify whether the user is a real live person, which can effectively resist common attack methods such as photos, face replacement, masks, occlusion, and screen reshoots, and help identify fraudulent behavior.

[0087] After liveness detection, a 1:1 comparison strategy or a 1:N comparison strategy is executed. The 1:1 comparison strategy is to compare two images to determine whether the person in the image is the same person to complete identity authentication; the 1:N comparison strategy is to compare one image with multiple images to determine whether there is an image in the multiple images that matches the image to be identified.

[0088] Schematically, after the liveness detection, the server executes a 1:1 matching strategy. Taking the payment application as an example, the server stores reference facial images corresponding to each account. The reference facial images are entered by the user when registering an account or completing the account information, and are used for matching in the facial recognition service. After receiving the facial recognition image sent by the terminal, the server obtains the account ID of the corresponding application logged in to the terminal, searches the database for the reference facial image based on the account ID, performs a 1:1 comparison between the facial recognition image and the reference facial image, obtains the first recognition result based on the similarity between the two, and returns the result to the terminal.

[0089] Indicatively, after the server performs liveness detection, it executes a 1:N comparison strategy. Taking the gate in the access control system as an example, the system server in the gate stores reference face images of all residents in the current building or community. The reference face images are entered in advance and will not be described here. The camera of the gate sends the face recognition image to the system server. The system server compares the face recognition image with each reference face image to obtain the similarity with each reference face image, and takes the highest similarity to determine whether the similarity reaches the preset threshold. If it does not reach, it means that the person currently being identified is not a resident of this community. If it reaches the threshold, it means that the person currently being identified is a resident of this community, and the first recognition result is returned to the terminal. Taking the e-government application as an example, the server stores a user face library, which contains reference face images of each registered user. After the terminal sends the face recognition image to the server, the server compares the face recognition image with each reference face image in the face library and outputs the first recognition result.

[0090] In an embodiment of the present application, the terminal sends the recognition accuracy requirement corresponding to the face recognition service request to the server, and the server is used to determine the target comparison strategy according to the recognition accuracy requirement, and the target comparison strategy is used to perform face recognition processing on the face recognition image uploaded by the terminal, wherein the target comparison strategy at least includes a face occlusion comparison strategy and a face occlusion comparison strategy. In an exemplary manner, the face occlusion comparison strategy includes allowing mouth occlusion, allowing forehead occlusion, allowing eye occlusion, allowing ear occlusion, etc.

[0091] In an embodiment of the present application, in response to the first recognition result corresponding to unsuccessful face recognition, verification information is obtained, the verification information corresponds to the current account information, and the verification information includes at least one of a mobile phone number, an ID number, an account binding email address, and an SMS verification code; the verification information is sent to a server; and a second recognition result fed back by the server based on the verification information is received. In one example, if the first recognition result fed back by the server is unsuccessful, the terminal prompts the user to enter a complete mobile phone number. After the server obtains the complete mobile phone number, it returns the mobile phone number to the server, which stores the account information and a reference face recognition image corresponding to the mobile phone number. The face recognition image is compared with the reference face recognition image to obtain a second recognition result, and the second recognition result is returned to the terminal. In response to the second recognition result corresponding to unsuccessful face recognition, a prompt message is displayed, and the prompt message is used to prompt the user to remove the occlusion. In one example, please refer to Figure 4 When the last recognition was unsuccessful, the application interface of the terminal 410 displays a prompt message 401 asking whether to perform recognition again. If the user selects the "Re-recognition" option, the face recognition interface is displayed again. If the user selects the "Scan code to pay" option, the face recognition operation is no longer performed. The face recognition video stream is obtained in the preset area 402, and a prompt message 403 for removing occlusion is displayed in its interface, such as "Please do not block your face."

[0092] In another example, see Figure 5 The last recognition was the first facial recognition service performed by the user using the account, that is, the server does not store a reference facial image corresponding to the account. When the user performs facial recognition again and the recognition is successful, the user is prompted whether to activate the interface for using facial recognition to complete the target function service. The interface includes activation information 501 and a confirmation control 502.

[0093] To summarize, the face recognition processing method provided in the embodiment of the present application determines the corresponding selection strategy according to the recognition accuracy requirement corresponding to the current face recognition service request when receiving a face recognition service request, determines the face recognition image from the face recognition video stream according to the corresponding selection strategy, and then uploads the face recognition image to the server, and the server completes the face recognition processing process. Finally, the terminal obtains the recognition result fed back by the server, which solves the problem that the same face recognition processing method cannot be applied in different application scenarios. It can be flexibly adjusted according to the face recognition requirements of different scenarios, thereby improving the flexibility of face recognition processing.

[0094] Please refer to Figure 6 , which shows a flow chart of a face recognition processing method shown in another embodiment of the present application. The method may include the following steps:

[0095] Step 601: Receive a face recognition service request.

[0096] In an embodiment of the present application, the face recognition processing method is applied to a payment application, wherein the face recognition service request is provided with a recognition accuracy requirement, which is the recognition accuracy of the target function requirement corresponding to the current face recognition service. In schematic form, the recognition accuracy requirement is determined according to the security requirements corresponding to the target function that the current payment application needs to execute. For example, the security requirement for face-swiping login is lower than the security requirement for payment operation, that is, face-swiping login corresponds to the first accuracy requirement, and payment operation corresponds to the second accuracy requirement.

[0097] Step 602: Obtain a face recognition video stream during the face recognition process according to a face recognition service request.

[0098] After receiving the face recognition service request generated by the payment application to complete the target function, the terminal calls the terminal's camera to start capturing the face recognition video stream during the face recognition process, and displays the captured video image in real time for the user on the application's interface.

[0099] Step 603: perform occlusion recognition on the face recognition video stream.

[0100] The terminal performs quality analysis on the video frames in the face recognition video stream to obtain quality analysis results corresponding to the video frames. According to the quality analysis results, it can be determined whether the face image in the video frame is blocked.

[0101] Step 6041, in response to the face recognition video stream not being blocked by any obstruction, a face recognition image is determined from the face recognition video stream according to a preset optimization logic.

[0102] The terminal determines that there is no occlusion in the face images in each video frame based on the face recognition video stream, and then determines the face recognition images in these video frames according to the preset preferential logic, and the preferential condition corresponding to the preset preferential logic is higher.

[0103] Step 6042, in response to the presence of an occlusion in the face recognition video stream, determining the occluded portion in the face image in the video frame.

[0104] When there is occlusion in the facial image in the video frame, the terminal needs to identify the occluded part in the video frame and determine whether there is occlusion in each preset part, for example, determine whether the terminal has at least one of mouth occlusion, forehead occlusion, eye occlusion or ear occlusion.

[0105] Step 6051, in response to the recognition accuracy requirement meeting the first accuracy requirement, determining the face recognition image from the face recognition video stream in a manner of ignoring occlusion of a preset part.

[0106] The terminal determines that there is occlusion in the face image in the video frame and has determined the occluded part, and then determines the corresponding selection strategy according to the recognition accuracy requirement corresponding to the current face recognition service request.

[0107] In an exemplary manner, the first accuracy requirement further corresponds to requirement information, which includes that occlusion of a preset part is allowed. In response to the requirement information that mouth occlusion is allowed, a video frame that ignores mouth occlusion in the face recognition video stream is determined as a face recognition image; in response to the requirement information that forehead occlusion is allowed, a video frame that ignores forehead occlusion in the face recognition video stream is determined as a face recognition image; in response to the requirement information that eye occlusion is allowed, a video frame that ignores eye occlusion in the face recognition video stream is determined as a face recognition image; in response to the requirement information that ear occlusion is allowed, a video frame that ignores ear occlusion in the face recognition video stream is determined as a face recognition image.

[0108] Step 6052, in response to the recognition accuracy requirement meeting the second accuracy requirement, determining a video frame without occlusion from the face recognition video stream as a face recognition image.

[0109] When the recognition accuracy requirement corresponding to the face recognition service request is the second accuracy requirement, which is higher than the first accuracy requirement, the terminal selects a video frame without occlusion in the face recognition video stream, and determines a face recognition image according to a preset preferred logic.

[0110] When all face images in the video frames of the face recognition video stream are blocked, a prompt message is displayed and the face recognition video stream is re-acquired. The prompt message is used to prompt the user to remove the blockage, for example, "Please do not block your face" is displayed on the interface.

[0111] Step 606: Send the face recognition image to the server and obtain the first recognition result fed back by the server.

[0112] The terminal sends the obtained face recognition image to the server, and the server performs face recognition processing services. The terminal only needs to obtain the first recognition result fed back by the server.

[0113] To summarize, the face recognition processing method provided in the embodiment of the present application determines a corresponding selection strategy according to the recognition accuracy requirement corresponding to the current face recognition service request when receiving a face recognition service request, and determines a face recognition image from the face recognition video stream according to the corresponding selection strategy, wherein, for requests with high recognition accuracy requirements, video frames without occlusion are selected from the face recognition video stream as face recognition images, and for requests with low recognition accuracy requirements, the occlusion of preset parts can be ignored to select face recognition images, and then the face recognition image is uploaded to the server, and the server completes the face recognition processing process, and finally the terminal obtains the recognition result fed back by the server, which solves the problem that the same face recognition processing method cannot be applied in different application scenarios, and can be flexibly adjusted according to the face recognition requirements of different scenarios, thereby improving the flexibility of face recognition processing.

[0114] Please refer to Figure 7 , which shows a flowchart of a face recognition processing method shown in another embodiment of the present application. The method can be applied to Figure 1 In the server in the implementation environment shown. The method may include the following steps:

[0115] Step 701: receiving the recognition accuracy requirement and the face recognition image sent by the terminal, and determining the target comparison strategy according to the recognition accuracy requirement.

[0116] In an embodiment of the present application, a target comparison strategy is used to perform face recognition processing on a face recognition image. Optionally, the target comparison strategy may be a 1:1 comparison strategy, i.e., the server compares the face recognition image with a reference face image, where the face reference image corresponds to the user's current login account and is entered by the user when registering an account or completing account information. Optionally, the target comparison strategy may be a 1:N comparison strategy, i.e., the server compares the face recognition image with N reference face images in a face library, where the face images in the face library are stored or entered in advance.

[0117] The target comparison strategy at least includes a comparison strategy that allows the existence of facial occlusion and a comparison strategy that prohibits the existence of facial occlusion. Among them, the comparison strategy that allows the existence of facial occlusion also includes a comparison strategy that allows the existence of mouth occlusion, a comparison strategy that allows the existence of forehead occlusion, a comparison strategy that allows the existence of ear occlusion, and a comparison strategy that allows the existence of eye occlusion.

[0118] Step 702, obtaining the quality score of each characteristic part according to the face recognition image.

[0119] The server performs feature analysis on the face recognition image to obtain the quality score of each feature part, which corresponds to the similarity between the face recognition image and the reference face image. Schematically, the server extracts the face features of the face recognition image to obtain a feature value of a fixed length, which has the ability to characterize the characteristics of the face, and then inputs the feature value into the face comparison algorithm to output the similarity of the features between the face recognition image and the reference face image, that is, the quality score.

[0120] Step 7031, in response to the recognition accuracy requirement meeting the first accuracy requirement, obtaining a first recognition result according to a facial occlusion comparison strategy.

[0121] The server determines the target comparison strategy according to the recognition accuracy requirement corresponding to the face recognition service uploaded by the terminal. When the recognition accuracy requirement meets the first accuracy requirement, the first recognition result is obtained according to the comparison strategy that allows facial occlusion. The first recognition result is obtained according to the overall recognition quality score, wherein the overall recognition quality score is obtained by weighted calculation of the quality scores corresponding to each special diagnosis part.

[0122] Optionally, the first accuracy requirement also corresponds to requirement information, which includes allowing occlusion of a preset part. Illustratively, when the first recognition result is obtained by calculating the overall recognition quality score, in response to the requirement information that mouth occlusion is allowed, the weight of the mouth feature quality score is reduced when calculating the overall recognition quality score; in response to the requirement information that forehead occlusion is allowed, the weight of the forehead feature quality score is reduced when calculating the overall recognition quality score; in response to the requirement information that eye occlusion is allowed, the weight of the eye feature quality score is reduced when calculating the overall recognition quality score; in response to the requirement information that ear occlusion is allowed, the weight of the ear feature quality score is reduced when calculating the overall recognition quality score. Optionally, when obtaining the first recognition result by calculating the overall recognition quality score, the first recognition result is obtained with different overall quality score thresholds, wherein the overall quality score threshold corresponding to the first precision requirement is lower than the overall quality score threshold corresponding to the second precision requirement, that is, reducing the impact of occlusion on the overall quality score. For example, the sorting quality score threshold corresponding to the first precision requirement is 80 points (100 points represents that the face recognition image is exactly the same as the reference face image), and the overall quality score threshold corresponding to the second precision requirement is 95 points.

[0123] Step 7032, in response to the recognition accuracy requirement meeting the second accuracy requirement, obtaining a first recognition result according to a facial occlusion comparison strategy that prohibits the presence of facial occlusion.

[0124] When the recognition accuracy requirement meets the second accuracy requirement, a comparison strategy that prohibits facial occlusion is used during the comparison operation. Since the face recognition image obtained by the terminal according to the second accuracy requirement does not have occlusion, the corresponding weights of each feature are the same when calculating the overall recognition quality score.

[0125] Step 704: Feedback the first recognition result to the terminal.

[0126] The server feeds back the obtained first recognition result to the terminal, such as "recognition success / failure". The terminal displays a corresponding display interface according to the first recognition result.

[0127] In an embodiment of the present application, in response to the first recognition result corresponding to the unsuccessful face recognition, the terminal obtains verification information, which corresponds to the current account information, and the verification information includes at least one of a mobile phone number, an ID card number, an account binding email address, and a text message verification code, and then sends the verification information to the server. After the server receives the verification information, it finds one or more reference face images corresponding to the verification information in the database, compares the reference face image with the face recognition image, and obtains a second recognition result. In one example, the terminal obtains the first recognition result fed back by the server as unsuccessful, then the terminal prompts the user to enter the complete mobile phone number, and after the server obtains the complete mobile phone number, the mobile phone number is returned to the server, and the server stores the account information and reference face recognition image corresponding to the mobile phone number, compares the face recognition image with the reference face recognition image, obtains the second recognition result, and returns the second recognition result to the terminal. If the second recognition result is still unsuccessful face recognition, the terminal displays a prompt message, which is used to prompt to eliminate the occlusion.

[0128] Please refer to Figure 8 , which shows the interaction process between the terminal and the server.

[0129] After receiving the face recognition service request, the terminal starts the face scanning process 801; the terminal obtains the face recognition video stream 802, that is, calls the camera to capture the facial image of the current user; then performs frame-by-frame quality judgment 803, which includes filtering out video frames without facial images, filtering out video frames with blurred facial features, etc.; the terminal records the current recognition 804, that is, the terminal records that the current recognition process is the first recognition within the preset time period. If the current recognition process is not the first recognition within the preset time period, it directly jumps to the preferred ignoring mouth occlusion judgment 808.

[0130] When the current recognition process is the first recognition within a preset time period, the terminal determines whether there is any occlusion for each video frame 805. If there is no occlusion, the preset optimization logic 807 is used to perform optimization operations on the video frame, that is, the face image in the video frame is not allowed to have occlusion; if there is occlusion, it is determined whether the occlusion is allowed to pass 806. If the occlusion is not allowed, the preset optimization logic 807 is still used to perform optimization operations on the video frame; if the occlusion is allowed to pass, the optimization is used to ignore the occlusion judgment of the preset part 808, that is, in the face image of the video frame, if there is occlusion in the preset part, the face recognition image can also be obtained through the optimization operation; the terminal uploads the face recognition image to the server 809.

[0131] The server determines whether the face recognition image is blocked 810, and marks whether there is occlusion. If there is no occlusion, it performs a normal threshold quality score judgment 811; if there is occlusion, the server records the current recognition as mask recognition 812; then a low threshold quality score judgment is performed on the face recognition image 813; the face recognition image after the quality score judgment is performed liveness detection 814; when there is no occlusion mark after the above judgment process 810, the conventional 1:N strategy 815 is used to compare the face recognition images.

[0132] When there is an occlusion mark after the above judgment process 810, it is determined whether mask recognition is turned on 816. If mask recognition is not turned on, the conventional 1:N strategy 815 is used to perform face recognition image comparison. If mask recognition is turned on, a dynamic threshold quality score judgment 817 is performed, that is, the threshold for allowing occlusion of the preset part is increased for quality score judgment; when the dynamic threshold quality score judgment is passed, the server receives the last four digits of the mobile phone number sent by the terminal and uses the 1:N strategy 819, wherein the last four digits of the mobile phone number are entered by the user, and the server searches for the corresponding N reference face images stored according to the last four digits of the data, and performs a comparison operation; when the dynamic threshold quality score judgment is not passed, the server receives the complete mobile phone number sent by the terminal and uses the 1:1 strategy 818, that is, the server searches for the corresponding reference face image according to the complete mobile phone number entered by the user, and performs a 1:1 comparison operation; finally, the server sends the recognition result to the terminal 820.

[0133] To summarize, the face recognition processing method provided in the embodiment of the present application determines the corresponding selection strategy according to the recognition accuracy requirement corresponding to the current face recognition service request when receiving a face recognition service request, determines the face recognition image from the face recognition video stream according to the corresponding selection strategy, and then uploads the face recognition image to the server, and the server completes the face recognition processing process. Finally, the terminal obtains the recognition result fed back by the server, which solves the problem that the same face recognition processing method cannot be applied in different application scenarios. It can be flexibly adjusted according to the face recognition requirements of different scenarios, thereby improving the flexibility of face recognition processing.

[0134] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0135] Please refer to Fig. 9 , which shows a block diagram of a face recognition device provided by an embodiment of the present application. The device has the function of implementing the above method example, and the function can be implemented by hardware or by hardware executing corresponding software. The device may include:

[0136] A receiving module 910 is used to receive a face recognition service request, wherein the face recognition service request is provided with a recognition accuracy requirement;

[0137] An acquisition module 920 is used to acquire a face recognition video stream in a face recognition process according to the face recognition service request;

[0138] The recognition module 930 is used to perform occlusion recognition on the face recognition video stream;

[0139] A determination module 940 is configured to determine a face recognition image from the face recognition video stream according to a selection strategy corresponding to the recognition accuracy requirement in response to the face recognition video stream being blocked by an obstruction;

[0140] The sending module 950 is used to send the face recognition image to the server to obtain the first recognition result fed back by the server.

[0141] In an alternative embodiment, please refer to Fig.10 , the recognition accuracy requirement includes a first accuracy requirement and a second accuracy requirement;

[0142] The determination module 940 includes:

[0143] A determination unit 941 is configured to determine the face recognition image from the face recognition video stream in a manner of ignoring occlusion of a preset part in response to the recognition accuracy requirement meeting the first accuracy requirement;

[0144] The determining unit 941 is further configured to, in response to the recognition accuracy requirement meeting the second accuracy requirement, determine a video frame without occlusion from the face recognition video stream as the face recognition image.

[0145] In an optional embodiment, the determination module 940 is further configured to determine an occluded portion in the face image in the video frame in response to the face recognition video stream being occluded by the occluder.

[0146] In an optional embodiment, the preset part occlusion includes at least one of mouth occlusion, forehead occlusion, eye occlusion, and ear occlusion, and the first accuracy requirement further corresponds to requirement information, and the requirement information includes that the preset part occlusion is allowed;

[0147] The determining unit 941 is further configured to, in response to the requirement information that the mouth occlusion is allowed, determine a video frame in the face recognition video stream where the mouth occlusion is ignored as the face recognition image;

[0148] The determining unit 941 is further configured to, in response to the requirement information that the forehead occlusion is allowed, determine a video frame in the face recognition video stream where the forehead occlusion is ignored as the face recognition image;

[0149] The determining unit 941 is further configured to, in response to the requirement information that the eye occlusion is allowed, determine a video frame in the face recognition video stream where the eye occlusion is ignored as the face recognition image;

[0150] The determining unit 941 is further configured to determine, in response to the requirement information that the ear occlusion is allowed, a video frame in the face recognition video stream that ignores the ear occlusion as the face recognition image.

[0151] In an optional embodiment, the acquisition module 920 is further configured to acquire verification information in response to the first recognition result corresponding to unsuccessful face recognition, wherein the verification information corresponds to the current account information;

[0152] The sending module 950 is further used to send the verification information to the server;

[0153] The receiving module 910 is further configured to receive a second recognition result fed back by the server according to the verification information.

[0154] In an alternative embodiment, please refer to Fig.10 , the device further comprises:

[0155] The display module 960 is configured to display a prompt message in response to the second recognition result corresponding to unsuccessful face recognition, wherein the prompt message is used to prompt to remove the occlusion.

[0156] In an optional embodiment, the acquisition module 920 is further used to perform quality analysis on the video frame in the face recognition video stream to obtain a quality analysis result corresponding to the video frame;

[0157] The determination module 940 is further configured to determine whether the face image in the video frame is blocked according to the quality analysis result.

[0158] In an optional embodiment, the sending module 950 is also used to send the recognition accuracy requirement corresponding to the face recognition service request to the server, and the server is used to determine a target comparison strategy based on the recognition accuracy requirement, and the target comparison strategy is used to perform face recognition processing on the face recognition image; the target comparison strategy includes at least one of a comparison strategy that allows the existence of facial occlusion and a comparison strategy that prohibits the existence of facial occlusion.

[0159] To sum up, the device provided by the embodiment of the present application, when receiving a face recognition service request, determines a corresponding selection strategy according to the recognition accuracy requirement corresponding to the current face recognition service request, determines a face recognition image from the face recognition video stream according to the corresponding selection strategy, and then uploads the face recognition image to the server, and the server completes the face recognition processing process. Finally, the terminal obtains the recognition result fed back by the server, which solves the problem that the same face recognition processing method cannot be applied in different application scenarios, and can be flexibly adjusted according to the face recognition requirements of different scenarios, thereby improving the flexibility of face recognition processing.

[0160] Fig.11 The following is a schematic diagram showing the structure of a server provided by an exemplary embodiment of the present application. Specifically:

[0161] The server 1100 includes a central processing unit (CPU) 1101, a system memory 1104 including a random access memory (RAM) 1102 and a read only memory (ROM) 1103, and a system bus 1105 connecting the system memory 1104 and the central processing unit 1101. The server 1100 also includes a mass storage device 1106 for storing an operating system 1113, application programs 1114, and other program modules 1115.

[0162] The mass storage device 1106 is connected to the central processing unit 1101 through a mass storage controller (not shown) connected to the system bus 1105. The mass storage device 1106 and its associated computer-readable media provide non-volatile storage for the server 1100. That is, the mass storage device 1106 may include a computer-readable medium (not shown) such as a hard disk or a compact disc read only memory (CD-ROM) drive.

[0163] Without loss of generality, computer-readable media may include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable media implemented by any method or technology for storing information such as computer-readable instructions, data structures, program modules or other data. Computer storage media include RAM, ROM, Erasable Programmable Read Only Memory (EPROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other solid-state storage technologies, CD-ROM, Digital Versatile Disc (DVD) or other optical storage, cassettes, tapes, disk storage or other magnetic storage devices. Of course, those skilled in the art will appreciate that computer storage media are not limited to the above. The above-mentioned system memory 1104 and mass storage device 1106 may be collectively referred to as memory.

[0164] According to various embodiments of the present application, the server 1100 can also be connected to a remote computer on the network through a network such as the Internet. That is, the server 1100 can be connected to the network 1112 through the network interface unit 1111 connected to the system bus 1105, or the network interface unit 1111 can be used to connect to other types of networks or remote computer systems (not shown).

[0165] The above-mentioned memory also includes one or more programs, and the one or more programs are stored in the memory and configured to be executed by the CPU.

[0166] The embodiment of the present application also provides a computer device, which includes a processor and a memory, wherein at least one instruction, at least one program, code set or instruction set is stored in the memory, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the biometric identification method provided by the above-mentioned method embodiments. Optionally, the computer device can be a terminal or a server.

[0167] An embodiment of the present application also provides a computer-readable storage medium, on which is stored at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the biometric recognition method provided by the above-mentioned method embodiments.

[0168] The embodiments of the present application also provide a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs any of the biometric identification methods described in the above embodiments.

[0169] Optionally, the computer readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a solid state drive (SSD), or an optical disk. Among them, the random access memory may include a resistance random access memory (ReRAM) and a dynamic random access memory (DRAM). The serial numbers of the above embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments.

[0170] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0171] The above description is only an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A face recognition processing method, It is characterized in that The method comprises: receiving a face recognition service request, wherein the face recognition service request is provided with a recognition accuracy requirement, wherein the recognition accuracy requirement is a recognition accuracy required by a target function corresponding to the face recognition service, wherein the target function includes at least one of a payment function and an account login function; Obtaining a face recognition video stream during the face recognition process according to the face recognition service request; Performing occlusion recognition on the face recognition video stream; In response to the face recognition video stream being blocked by an obstruction, when the recognition accuracy requirement meets the first accuracy requirement, a face recognition image is determined from the face recognition video stream in a manner of ignoring the blockage of a preset part, where the preset part blockage includes at least one of mouth blockage, forehead blockage, eye blockage, and ear blockage; when the recognition accuracy requirement meets the second accuracy requirement, a video frame without blockage is determined from the face recognition video stream as the face recognition image, and the first accuracy requirement is lower than the second accuracy requirement; The face recognition image is sent to a server to obtain a first recognition result fed back by the server.

2. The method according to claim 1, It is characterized in that After the face recognition video stream is subjected to occlusion recognition, the method further includes: In response to the face recognition video stream being blocked by the blocking object, an blocked portion in the face image in the video frame is determined.

3. The method according to claim 2, It is characterized in that The first accuracy requirement also corresponds to requirement information, and the requirement information includes allowing occlusion of a preset part; The step of determining a face recognition image from the face recognition video stream in a manner of ignoring occlusion of a preset part includes: In response to the requirement information that the mouth occlusion is allowed, determining a video frame in the face recognition video stream that ignores the mouth occlusion as the face recognition image; In response to the requirement information that the forehead occlusion is allowed, determining a video frame in the face recognition video stream that ignores the forehead occlusion as the face recognition image; In response to the requirement information that the eye occlusion is allowed, determining a video frame in which the eye occlusion is ignored in the face recognition video stream as the face recognition image; In response to the requirement information that the ear occlusion is allowed, a video frame in which the ear occlusion is ignored in the face recognition video stream is determined as the face recognition image.

4. The method according to any one of claims 1 to 3, It is characterized in that After obtaining the first recognition result fed back by the server, the method further includes: In response to the first recognition result corresponding to unsuccessful face recognition, obtaining verification information, the verification information corresponding to the current account information; Sending the verification information to the server; Receive a second recognition result fed back by the server according to the verification information.

5. The method according to claim 4, It is characterized in that After receiving the second recognition result fed back by the server according to the verification information, the method further includes: In response to the second recognition result corresponding to unsuccessful face recognition, a prompt message is displayed, where the prompt message is used to prompt to remove the occlusion.

6. The method according to any one of claims 1 to 3, It is characterized in that The performing occlusion recognition on the face recognition video stream includes: Performing quality analysis on the video frames in the face recognition video stream to obtain quality analysis results corresponding to the video frames; Determine whether there is occlusion in the face image in the video frame according to the quality analysis result.

7. The method according to claim 1, It is characterized in that Before obtaining the first recognition result fed back by the server, the method further includes: The recognition accuracy requirement corresponding to the face recognition service request is sent to the server, and the server is used to determine a target comparison strategy based on the recognition accuracy requirement, and the target comparison strategy is used to perform face recognition processing on the face recognition image; the target comparison strategy includes at least one of a comparison strategy that allows the existence of facial occlusion and a comparison strategy that prohibits the existence of facial occlusion.

8. A face recognition processing device, It is characterized in that The device comprises: A receiving module, configured to receive a face recognition service request, wherein the face recognition service request is provided with a recognition accuracy requirement, wherein the recognition accuracy requirement is a recognition accuracy required by a target function corresponding to the face recognition service, wherein the target function includes at least one of a payment function and an account login function; An acquisition module, used for acquiring a face recognition video stream in a face recognition process according to the face recognition service request; A recognition module, used for performing occlusion recognition on the face recognition video stream; a determination module, configured to, in response to the face recognition video stream being blocked by an obstruction, determine a face recognition image from the face recognition video stream by ignoring a preset part blockage when the recognition accuracy requirement meets a first accuracy requirement, wherein the preset part blockage includes at least one of mouth blockage, forehead blockage, eye blockage, and ear blockage; and determine a video frame without blockage from the face recognition video stream as the face recognition image when the recognition accuracy requirement meets a second accuracy requirement, wherein the first accuracy requirement is lower than the second accuracy requirement; The sending module is used to send the face recognition image to the server to obtain the first recognition result fed back by the server.

9. The device according to claim 8, It is characterized in that The determination module is further configured to determine an occluded portion in the face image in the video frame in response to the presence of an occlusion by the occluder in the face recognition video stream.

10. The device according to claim 9, It is characterized in that The first accuracy requirement also corresponds to requirement information, and the requirement information includes allowing occlusion of a preset part; The determining unit is further configured to, in response to the requirement information that the mouth occlusion is allowed, determine a video frame in the face recognition video stream where the mouth occlusion is ignored as the face recognition image; The determining unit is further configured to, in response to the requirement information that the forehead occlusion is allowed, determine a video frame in the face recognition video stream where the forehead occlusion is ignored as the face recognition image; The determining unit is further configured to, in response to the requirement information that the eye occlusion is allowed, determine a video frame in the face recognition video stream where the eye occlusion is ignored as the face recognition image; The determination unit is further configured to determine, in response to the requirement information that the ear occlusion is allowed, a video frame in the face recognition video stream that ignores the ear occlusion as the face recognition image.

11. A terminal, It is characterized in that The terminal includes a processor and a memory, wherein at least one program is stored in the memory, and the at least one program is loaded and executed by the processor to implement the face recognition processing method according to any one of claims 1 to 7.

12. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores at least one program, and the at least one program is loaded and executed by the processor to implement the face recognition processing method according to any one of claims 1 to 7.

13. A computer program product, It is characterized in that It includes computer instructions, which, when executed by a processor, implement the face recognition processing method as described in any one of claims 1 to 7.

Images