Face recognition control method and effect monitoring method, related device and system

Abstract

One or more embodiments of the present specification provide a face recognition control method and its effect monitoring method, related devices and systems. On the target terminal side of the face recognition related service system, when the image acquisition page before the start of face recognition displays the current image captured by the camera, the face outline of the user in the current image is dynamically captured, and preset decorative elements are loaded in the page according to the face outline; or when the current image is sent to the target server for comparison, the image acquisition page is switched to the status prompt page, and the current image is stylized through the preset generation model, and the real face image of the user in the current image is processed into a non-real face image of a preset style, and displayed as a user avatar on the above-mentioned status prompt page. In addition, on the target server side, the preset monitoring parameters of the target terminal side are obtained, and the face recognition control process on the terminal side is optimized according to the preset monitoring parameters.

Description

[0001] This application is a divisional application of the invention application with application date of July 24, 2024, application number 202410994463.8, and invention name “Face recognition control method and effect monitoring method, related devices and systems”. Technical Field

[0002] One or more embodiments of the present specification relate to the field of information processing technology, and in particular, to a face recognition control method and effect monitoring method thereof, and related devices and systems. Background Art

[0003] As a means of identity authentication, face recognition is widely used in scenarios such as electronic payment, attendance check-in, and access control. In these actual scenarios, the terminal device used to collect face images is usually equipped with a display screen facing the authenticated user, namely, the C screen, which can synchronously display the image currently captured by the camera of the terminal device, so that the authenticated user can adaptively adjust the relative position between himself and the terminal device, so that his face image can be fully collected, thereby improving the efficiency and success rate of authentication.

[0004] Although identity authentication through facial recognition has the advantages of convenience and speed, there are still some users who are not very enthusiastic about using related facial recognition devices or services due to concerns about privacy data leakage, which affects the promotion of related devices or services. Summary of the invention

[0005] In order to increase users' interest in face recognition services and reduce users' concerns about privacy leakage during the face recognition process, one or more embodiments of this specification provide a face recognition control method and its effect monitoring method, related devices and systems.

[0006] In a first aspect, one or more embodiments of this specification provide a face recognition control method, including:

[0007] Synchronously displaying the current image captured by the camera of the target device on the display screen of the target terminal through an image capture page;

[0008] Dynamically capture the facial contour in the current image;

[0009] Preset decorative elements are loaded into the image acquisition page according to the facial contour.

[0010] In a possible implementation, the step of loading a preset decorative element in the image acquisition page according to the face contour includes:

[0011] The preset decorative element is adjusted according to the size and / or position of the facial contour.

[0012] In a possible implementation, the method further includes:

[0013] When receiving a face recognition instruction, inputting the current image into a preset generation model, and performing stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0014] Switching the image acquisition page to a status prompt page, and displaying the stylized image on the status prompt page;

[0015] Among them, the status prompt page also displays prompt information indicating the face recognition status.

[0016] In a possible implementation, the preset generation model satisfies at least one of the following:

[0017] The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model;

[0018] The preset generation model includes a binary classification model for controlling the quality of image generation;

[0019] The preset generation model is an end-to-end model obtained based on distillation technology.

[0020] In a second aspect, one or more embodiments of this specification provide a face recognition control method, including:

[0021] When receiving a face recognition instruction, inputting a current image collected by a camera of the target terminal into a preset generation model, and performing stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0022] Displaying the stylized image on a status prompt page;

[0023] The status prompt page also displays prompt information indicating the face recognition status; the status prompt page is displayed through the display screen of the target terminal.

[0024] In a possible implementation, the preset generation model satisfies at least one of the following:

[0025] The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model;

[0026] The preset generation model includes a binary classification model for controlling the quality of image generation;

[0027] The preset generation model is an end-to-end model obtained based on distillation technology.

[0028] In a third aspect, one or more embodiments of this specification further provide a method for monitoring face recognition control effect, including:

[0029] Determine the target terminal according to the preset selection rules;

[0030] Sending a preset generation model and / or a preset dynamic tracking model to the target terminal;

[0031] The preset selection rule includes a selection rule of at least one of the following dimensions: terminal performance, user label of the scene where the terminal is located;

[0032] The preset generation model is used to perform stylization processing on the current image collected by the target terminal;

[0033] The preset dynamic tracking model is used to dynamically capture the face contour in the current image, and load and display preset decorative elements according to the face contour.

[0034] In a possible implementation, the method further includes:

[0035] Acquire preset monitoring parameters on each of the target terminals; the preset monitoring parameters include the number of face recognition times;

[0036] The preset generation model and / or the preset dynamic tracking model are optimized according to the preset monitoring parameters.

[0037] In a fourth aspect, one or more embodiments of this specification provide a face recognition control device, including:

[0038] A first display unit, configured to synchronously display a current image captured by a camera of the target device on a display screen of the target terminal through an image capture page;

[0039] The dynamic tracking unit is used to dynamically capture the face contour in the current image; and load the preset decorative elements in the image acquisition page according to the face contour.

[0040] In a possible implementation, the dynamic tracking unit is used to load a preset decorative element in the image acquisition page according to the face contour, including:

[0041] The dynamic tracking unit is used to adjust the preset decorative element according to the size and / or position of the facial contour.

[0042] In a possible implementation manner, the device further includes:

[0043] A first stylization unit is configured to input the current image into a preset generation model upon receiving a face recognition instruction, and perform stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0044] A second display unit, used to switch the image acquisition page to a status prompt page, and display the stylized image on the status prompt page;

[0045] Among them, the status prompt page also displays prompt information indicating the face recognition status.

[0046] In a possible implementation, the preset generation model in the first stylization unit satisfies at least one of the following:

[0047] The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model;

[0048] The preset generation model includes a binary classification model for controlling the quality of image generation;

[0049] The preset generation model is an end-to-end model obtained based on distillation technology.

[0050] In a fifth aspect, one or more embodiments of this specification provide a face recognition control device, including:

[0051] A second stylization unit is used to input a current image collected by a camera of a target terminal into a preset generation model upon receiving a face recognition instruction, and perform stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0052] A third display unit, used to display the stylized image on a status prompt page;

[0053] The status prompt page also displays prompt information indicating the face recognition status; the status prompt page is displayed through the display screen of the target terminal.

[0054] In a possible implementation, the preset generation model in the second stylization unit satisfies at least one of the following:

[0055] The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model;

[0056] The preset generation model includes a binary classification model for controlling the quality of image generation;

[0057] The preset generation model is an end-to-end model obtained based on distillation technology.

[0058] In a sixth aspect, one or more embodiments of this specification provide a face recognition control effect monitoring device, including:

[0059] A selection unit, used to determine a target terminal according to a preset selection rule;

[0060] A sending unit, used for sending a preset generation model and / or a preset dynamic tracking model to the target terminal;

[0061] The preset selection rule includes at least one of the following selection rules: terminal performance, user label of the scene where the terminal is located;

[0062] The preset generation model is used to perform stylization processing on the current image collected by the target terminal;

[0063] The preset dynamic tracking model is used to dynamically capture the face contour in the current image, and load and display preset decorative elements according to the face contour.

[0064] In a possible implementation manner, the device further includes:

[0065] An acquisition unit, configured to acquire preset monitoring parameters on each of the target terminals; the preset monitoring parameters include the number of face recognition times;

[0066] An optimization unit is used to optimize the preset generation model and / or the preset dynamic tracking model according to the preset monitoring parameters.

[0067] In a seventh aspect, one or more embodiments of this specification further provide a service system based on face recognition, characterized in that it includes: a target service end and at least one target terminal;

[0068] The target terminal is used to perform face recognition control according to the method of the first aspect or the second aspect when interacting with the user;

[0069] The target server is used to monitor and optimize the face recognition control process of the target terminal according to the method of the third aspect mentioned above.

[0070] In an eighth aspect, one or more embodiments of the present specification further provide an electronic device comprising a memory and a processor; the memory is used to store a computer program product; the processor is used to execute the computer program product stored in the memory, and when the computer program product is executed, the method of the first aspect, the second aspect or the third aspect mentioned above is implemented.

[0071] In a ninth aspect, one or more embodiments of the present specification further provide a computer-readable storage medium, wherein the computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed, the method of the first aspect, the second aspect, or the third aspect described above is implemented.

[0072] In summary, one or more embodiments of this specification provide a face recognition control method and its effect monitoring method, related devices and systems. In view of the situation that some users refuse to use face recognition and related services due to concerns about privacy leakage in face recognition, face recognition control is performed on the target terminal that directly interacts with the user, so as to weaken the user's concerns about privacy leakage in each link of the face recognition chain, thereby bringing about the effect of customer acquisition and diversion; face recognition control effect monitoring is also performed on the target service end, and the actual effect of face recognition control on each target terminal is monitored and evaluated, and the preset generation model involved in the above face recognition control method can be optimized according to the evaluation results to improve its customer acquisition and diversion effect.

[0073] Among them, through the above-mentioned face recognition control method and device, it is possible to reduce the user's concerns about privacy leakage in each link of the face recognition full link, which is specifically reflected in the following aspects:

[0074] On the one hand, when the current image captured by the camera is displayed on the image acquisition page before face recognition starts, the user's face outline in the current image can be dynamically captured, and preset decorative elements can be loaded into the page according to the face outline, which not only increases the interest of the image acquisition page and helps to attract users, but also partially blocks the user's face image displayed on the page through the preset decorative elements, that is, weakens the user's concerns about privacy leakage before face recognition starts;

[0075] On the other hand, when the current image is sent to the target server for comparison, that is, face recognition, the image acquisition page is switched to the status prompt page, and the current image is stylized through a preset generation model, and the user's real face image in the current image is processed into a non-real face image of a preset style, and displayed as a user avatar on the above status prompt page, that is, the user's concerns about privacy leakage are weakened in the status prompt link during face recognition;

[0076] At the same time, the stylized image obtained by the above processing can also be used as the user avatar required to be displayed on the service result page, reducing the user's concerns about privacy leakage in the result display link after the face recognition and related services are completed.

[0077] In addition, the face recognition-based service system provided in this embodiment controls the face recognition process through the target terminal, weakens user privacy concerns, monitors and optimizes the control effect of the target terminal through the target service end, realizes automatic control and optimization of the entire service system, and continuously improves the customer acquisition and flow effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] In order to more clearly illustrate the technical solutions of one or more embodiments of the present specification, 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 one or more embodiments of the present specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0079] Figure 1 A flowchart of a face recognition control method provided for one or more embodiments of this specification;

[0080] Figure 2 A schematic diagram of an application scenario provided for one or more embodiments of this specification;

[0081] Figure 3 A schematic diagram of an application scenario provided for one or more embodiments of this specification;

[0082] Figure 4 A schematic diagram of a user interaction interface when the face recognition control method provided by one or more embodiments of this specification is applied to a face payment scenario;

[0083] Figure 5 A flowchart of another face recognition control method provided for one or more embodiments of this specification;

[0084] Figure 6 A flowchart of a method for monitoring the control effect of face recognition provided in one or more embodiments of this specification;

[0085] Figure 7 A structural block diagram of a face recognition control device provided for one or more embodiments of this specification;

[0086] Figure 8 A structural block diagram of another face recognition control device provided for one or more embodiments of this specification;

[0087] Fig. 9 A structural block diagram of another face recognition control device provided for one or more embodiments of this specification;

[0088] Fig.10A structural block diagram of a face recognition control effect monitoring device provided by one or more embodiments of this specification;

[0089] Fig.11 A structural block diagram of an electronic device provided for one or more embodiments of this specification. DETAILED DESCRIPTION

[0090] One or more embodiments of the present invention are further described in detail below through the accompanying drawings and examples. Through these descriptions, the features and advantages of one or more embodiments of the present invention will become clearer and more specific.

[0091] The word "exemplary" is used exclusively herein to mean "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise noted.

[0092] In addition, the technical features involved in different implementations of one or more embodiments of this specification described below can be combined with each other as long as they do not conflict with each other.

[0093] To facilitate understanding, the application scenarios of the technical solutions provided by one or more embodiments of this specification are first described below. Figure 2 is a schematic diagram of a face recognition system in the related art, Figure 3 A schematic diagram of an interface for interaction between a terminal of a face recognition system and a user in related technology.

[0094] like Figure 2 As shown, the face recognition system 90 includes a server 91 and at least one terminal 92; wherein the terminal 92 is used to collect the face image P of the user to be identified and upload it to the server 91 through the Internet of Things, and the server 91 is used to compare the face image P collected by the terminal 92 with the face image P0 of the pre-stored legal user to obtain the face recognition result, that is, to determine whether the user to be identified is a legal user of the system. In different application scenarios, different responses can be executed after obtaining the recognition result; for example, in the electronic payment scenario of a supermarket, after the face recognition is successful, the server can deduct the payment amount from the current user's account to complete the face payment, and can also return the recognition result and the payment result to the terminal through the Internet of Things in turn, so as to display it to the current user; in the access control scenario of an office building, the terminal can temporarily release the access control according to the face recognition success result returned by the server, allowing the current user to pass; in the attendance check-in scenario of a company, after the face recognition is successful, the server can mark the check-in status of the current user as a successful check-in.

[0095] The terminal 92 is usually equipped with modules such as a camera 921 and a display screen 922. As long as the user to be identified enters the field of view of the camera 921, the user to be identified can be captured; at the same time, the display screen 922 can synchronously display the image captured by the camera 921, so that the user to be identified can understand whether his face is completely in the field of view of the camera by observing the display screen 922, and can adaptively adjust his position to make the face image captured by the camera 921 higher in quality and improve the recognition success rate. Among them, because the display screen 922 is for individual users (Customer), it is usually called a C screen, which is different from the corporate users (Business) introduced by the above-mentioned supermarkets, office buildings, companies, etc.

[0096] like Figure 3 As shown, taking the electronic payment scenario of face-swiping payment as an example, in the related technology, when the user enters the field of view of the camera 921, the display screen 922 can synchronously display the real image of the user collected by the camera through the image acquisition page 81; if the user agrees to use the face-swiping payment, the button 811 in the image acquisition page 81 can be clicked to trigger the terminal 92 to upload the facial image collected at this time to the server 91 for comparison. At the same time, the terminal 92 switches the display screen 922 from the image acquisition page 81 to the status prompt page 82. In the status prompt page 82, on the one hand, the thumbnail of the real facial image finally collected can be displayed in the image frame 821, and on the other hand, the facial recognition status and payment status of the current user can be displayed in text form, such as displaying status prompt texts such as "identifying" and "paying, please wait"; after the face-swiping payment is successful, the display screen 922 is switched to the payment result page 83 to show the user information such as the actual payment amount; of course, if the face-swiping payment fails, the payment result page 83 can display relevant information related to the failure.

[0097] like Figure 3 As shown in the image acquisition page 81, the status prompt page 82, and the payment result page 83, the relevant technology will display the user's real face image on each page of the entire link of face recognition. Although the face recognition system server will provide sufficient protection measures for the user's privacy data, visually, sensitive users will still worry about the leakage of their private information such as their photos, and thus refuse to use face recognition services, affecting the penetration rate of related equipment or technologies among the majority of users.

[0098] In view of this, in order to increase users' interest in face recognition services and reduce users' concerns about privacy leakage during the face recognition process, one or more embodiments of this specification provide a face recognition control method and device, which are described in detail below.

[0099] Figure 1This is a flow chart of a face recognition control method provided by an embodiment of this specification. The method can be applied to a target terminal of any target face recognition system to control the interface and content displayed on the display screen of the target terminal during the face recognition process. Figure 1 , the face recognition control method includes the following steps.

[0100] Step 102, synchronously displaying the current image captured by the camera of the target device on the display screen of the target terminal through an image capture page;

[0101] Step 104, dynamically capturing the face contour in the current image;

[0102] Step 106: load preset decorative elements into the image acquisition page according to the face contour.

[0103] Since the user in front of the target terminal can move, the current image captured by the camera of the target terminal at certain times may not contain the face of any user, or the position of the face image of the current user in the current image is not fixed. In view of this, this embodiment uses dynamic capture technology to capture the face outline in the current image, and if the face outline is successfully captured, the preset decorative element is loaded in the image acquisition page.

[0104] Figure 4 This is a schematic diagram of the effect of the face recognition control method provided in this embodiment when it is applied to the face payment scenario. Figure 4 Before the user agrees to start face recognition, this embodiment not only synchronously displays the current image captured by the camera through the image acquisition page 71, but also dynamically captures the face outline, and based on the captured face outline, loads preset decorative elements 712 such as "bow tie" and "copper coin" around the face outline. These preset decorative elements are equivalent to electronic stickers, thereby achieving instant interaction with the current user.

[0105] This embodiment loads the preset decorative elements in the image acquisition page before face recognition, that is, adding electronic stickers to the user's face image, which can not only enhance the fun of face recognition and increase the user's interest in face recognition related services, but also partially block the user's face in the image acquisition page, visually increase the difference between the overall image displayed in the image acquisition page and the user's real image, thereby weakening the user's concerns about privacy leakage in the face recognition process, increasing the probability of users using face recognition related services, and achieving customer acquisition and drainage of face recognition related services.

[0106] In a possible implementation, in step 106, loading a preset decorative element in the image acquisition page according to the face contour may specifically include:

[0107] The preset decorative element is adjusted according to the size and / or position of the facial contour.

[0108] In this embodiment, after successfully capturing the facial outline in the current image, it can also be dynamically followed, so that as the position or size of the facial outline in the current image changes, the position, size, etc. of the preset decorative elements are adaptively adjusted, thereby further increasing the fun and improving the user experience.

[0109] Of course, in other embodiments, the preset decorative elements loaded based on the face outline are not limited to Figure 4 As shown, it can also be different graphics such as wigs, glasses, red envelopes, etc., which can be configured according to the actual application scenarios and application requirements. Secondly, for the same target terminal, a variety of preset decorative elements can be pre-configured for it. Each time it needs to be loaded, one or more of them can be randomly selected for loading; in this way, the preset decorative elements loaded each time can be different, thereby further enhancing the fun of face recognition. In addition, the pre-configured preset decorative elements can be modified, replaced, added or deleted according to the application requirements of different periods. For example, during the New Year's Day, holiday-related elements such as "Happy New Year" art characters can be added as preset decorative elements, which further enhances the user's interest in face recognition-related services and helps to attract customers for face recognition-related services.

[0110] In a possible implementation, the face recognition control method may further include the following stylized image generating step and stylized image displaying step.

[0111] Stylized image generation steps:

[0112] When receiving a face recognition instruction, inputting the current image into a preset generation model, and performing stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0113] The above face recognition instruction can be manually triggered by the user or automatically triggered by relevant control conditions. Figure 4 As shown, when the user clicks button 711 in image acquisition page 71, the face recognition instruction can be manually triggered. In addition, some control conditions can be preset, and when the size, position, clarity, etc. of the face image in the current image collected meet these preset control conditions, the face recognition instruction is automatically generated.

[0114] When the target terminal detects the face recognition instruction, in addition to sending the current image P collected at this time to the corresponding target server, so that the target server can compare the features of the current image P with the stored standard images P0 of each legitimate user (or registered user), the current image P is also input into the preset generation model for stylization processing to obtain the corresponding stylized image.

[0115] The preset generation model, i.e., the trained artificial intelligence generated content (AIGC) model, can process the current image into an image of a preset style, such as a cartoon style, a hand-drawn style, a comic style, etc. In actual application scenarios, any model with a style image generation function can be obtained through fine-tuning training (i.e., fine tuning) and other model training methods.

[0116] Stylized image display steps:

[0117] Switching the image acquisition page to a status prompt page, and displaying the stylized image on the status prompt page;

[0118] Among them, the status prompt page also displays prompt information indicating the face recognition status.

[0119] In response to the above-mentioned face recognition instruction, the current image is uploaded to the target server for face recognition, and the content displayed on the display screen of the target terminal also needs to be switched from the image acquisition page to the status prompt page; the status prompt page displays prompt information indicating the face recognition status of the target server, as well as the stylized image obtained by processing the preset generation model.

[0120] like Figure 4 In the face-scanning payment scenario shown, after the user agrees to face-scanning payment and clicks button 711 in image acquisition page 71, the target terminal (specifically the payment terminal in this scenario) can detect the face recognition instruction and respond as follows: on the one hand, the current image captured by the camera at this time (including the face image of the current user) is sent to the target server (specifically the payment server in this scenario) for comparison to achieve face recognition; on the other hand, the current image is processed into a comic-style image through the preset generation model configured in the target terminal, that is, the real face image in the current image is processed into a comic-style face image; on the other hand, the image acquisition page 71 is switched to status prompt pages 72 and 72' in turn, and the status prompt pages 72 and 72' not only display prompt information indicating the face recognition status of the target server, such as "identifying" and "please wait during payment", but also display the above-mentioned comic-style face image.

[0121] Compared with displaying real face images in the status prompt page (such as Figure 3 As shown in the figure, by processing it into a non-real face image of a special style in this embodiment, the target terminal can be less exposed to the user's real face image during face recognition, the user's concerns about privacy leakage during the face recognition process can be reduced, the probability of users using face recognition related services can be increased, and customer acquisition and drainage of face recognition related services can be achieved.

[0122] In a possible implementation, the preset generation model may be an image generation model obtained by fine-tuning a Stable Diffusion model based on a Dream Booth model.

[0123] The above-mentioned Stable Diffusion model, i.e., the stable diffusion model, referred to as the SD model, is a large model for generating content from text to image. The images generated by it are diverse. Therefore, in order to obtain the images of a specific style required by this embodiment, the above-mentioned SD model can be fine-tuned and trained based on the fine-tuning training method such as the Dream Booth model, that is, additional learning and training are performed on the basis of the SD model, and information for controlling the style of the generated image is injected into the SD model; in this way, the model finally trained is used as the above-mentioned preset generation model, which can realize that after inputting a picture of any subject object, the stylized image corresponding to the subject object can be output. For example, in the process of fine-tuning training, cartoon / hand-painted / ... and other style description texts are added to the training samples, and then the preset generation model finally trained can generate corresponding cartoon / hand-painted / ... and other style face images for any input face image.

[0124] Of course, in other possible implementations, the Stable Diffusion model can also be fine-tuned and trained based on models such as LoRA (i.e., Low-Rank Adaptation) to obtain the above-mentioned preset generation model; or other types of AIGC models with specific style image generation functions, such as Midjourney, can also be used as the above-mentioned preset generation model. This embodiment will not be listed one by one.

[0125] In a possible implementation, the preset generation model may include a binary classification model for controlling image generation quality.

[0126] For an input real face image, the preset generation model may simultaneously generate multiple candidate images of different qualities and containing different detail features. This embodiment configures a binary classification model in the preset generation model and sets relevant thresholds, so that it is possible to determine whether each candidate image meets the threshold requirements and output the image that meets the threshold requirements as the required stylized image. The threshold may be a threshold for evaluating image resolution, a threshold for evaluating the degree of match between an image and a preset style, etc.

[0127] It can be seen that this embodiment configures a binary classification model in the preset generation model to ensure that the resolution of the finally generated stylized image and the degree of matching with the preset style are higher, and that the quality of the stylized image displayed in the above-mentioned status prompt page is controllable, which helps to achieve a balance between the optimization cost and optimization effect of the face recognition process.

[0128] In a possible implementation, the preset generation model may be an end-to-end model obtained based on distillation technology.

[0129] The image generation model directly obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model has a large size and consumes a lot of device performance during the process. In view of this, in this embodiment, it can be processed into an end-to-end small model based on the distillation technology, that is, the preset generation model, which can be an end-to-end model that has undergone fine-tuning training and distillation training successively. This can reduce the performance requirements of the preset generation model on the hardware device, and can be applied to more types of target terminals, thereby improving the image generation speed and improving the availability of the model in different scenarios and different terminals.

[0130] It can be seen that this embodiment configures the above-mentioned preset generation model in the target terminal, which can not only process the real face image in the current image P into a non-real face image in a cartoon, hand-painted or other style and display it, thereby reducing user privacy concerns, but also control the generation quality of the non-real face image, and can ensure that the preset generation model matches the performance of the target terminal through model distillation, thereby ensuring the usability of this embodiment in actual scenarios.

[0131] In one possible implementation, after the target terminal or target server executes the relevant response service based on the face recognition result, this embodiment can also switch the display interface of the target terminal from the above-mentioned status prompt page to the service result page. In addition to displaying the relevant response service results, the service result page can also include user avatar, user account and other information, where the user avatar can be the above-mentioned stylized image or other system default picture.

[0132] Reference Figure 4In the face payment scenario, the target server executes the corresponding payment response after successful face recognition, and can notify the target terminal after successful payment, thereby triggering the display screen of the target terminal to switch from the status prompt page 72 to the service result page 73 (i.e., the payment result page). The service result page 73 displays the user avatar (the aforementioned comic style image or the system default image or the image set by the user), the user account (i.e., "123****1111"), the payment amount "￥10.00", and other information, so that the user can understand the payment result. Among them, the user avatar in the service result page 73 is not the real face image of the user collected by the camera, but the system default image. Of course, the comic style image displayed in the status prompt page 72 can also be used, thereby weakening the user's privacy concerns during the service result display stage.

[0133] In summary, on the one hand, this embodiment can dynamically capture the face contour in the current image captured by the camera of the target terminal, and load the preset device element based on the face contour, so as to weaken the user's privacy concerns in the image acquisition link before face recognition; on the other hand, the generative artificial intelligence technology can be used to stylize the current image, so that the real face image therein is changed into a non-real face image of a preset style, and the processed stylized image is used to replace the original current image and displayed on the status prompt page, so as to weaken the user's privacy concerns in the status prompt link during face recognition; in addition, other pages that the target terminal needs to display, such as the service result page, if the user avatar needs to be loaded, can sample the above-mentioned stylized image or the system default image, reduce or even avoid the direct exposure of the display page of the target terminal to the user's face image, weaken the user's privacy concerns in the whole link before, during and after face recognition, enhance the user's willingness to use face recognition and its related services (such as face payment, face check-in, etc.), and bring customer acquisition and drainage effects to face recognition and its related services.

[0134] The foregoing embodiments provide optimization control methods in different ways at different stages of face recognition, so as to achieve the effect of reducing user privacy concerns; in different embodiments or application scenarios, a certain link in the entire face recognition link can be optimized as needed.

[0135] like Figure 5 As shown, one or more embodiments of the present specification also provide a face recognition control method, which may include the following steps.

[0136] Step 202, when receiving a face recognition instruction, inputting a current image collected by the camera of the target terminal into a preset generation model, and performing stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0137] Step 204, displaying the stylized image on a status prompt page;

[0138] The status prompt page also displays prompt information indicating the face recognition status; the status prompt page is displayed through the display screen of the target terminal.

[0139] In a possible implementation, the preset generation model satisfies at least one of the following:

[0140] The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model;

[0141] The preset generation model includes a binary classification model for controlling the quality of image generation;

[0142] The preset generation model is an end-to-end model obtained based on distillation technology.

[0143] above Figure 5 The face recognition control method shown realizes that in face recognition, the current image that is originally required to be displayed on the status prompt page is replaced with a stylized image through a preset generation model and displayed. Compared with directly displaying the current image containing the user's real face image, this embodiment can not only weaken the user's privacy concerns, but also improve the interest of the content displayed on the status prompt page, thereby increasing the user's willingness to continue using face recognition and related services.

[0144] Based on the same inventive concept, one or more embodiments of this specification also provide a method for monitoring face recognition control effects. Figure 6 Schematic diagram of the face recognition control effect monitoring method. Figure 6 The face recognition control effect monitoring method may include the following steps.

[0145] Step 302, determining the target terminal according to a preset selection rule;

[0146] Step 304, sending the preset generation model and / or the preset dynamic tracking model to the target terminal;

[0147] The preset generation model is used to perform stylization processing on the current image collected by the target terminal;

[0148] The preset dynamic tracking model is used to dynamically capture the face contour in the current image, and load and display preset decorative elements according to the face contour.

[0149] Previous Figure 1 , Figure 5The face recognition control method described in the corresponding embodiment, that is, the method for reducing user privacy concerns by optimizing different links of the face recognition full link, can be implemented with corresponding software models, that is, the preset generation model, the preset dynamic tracking model, etc. These software models can be stored in the target server of the face recognition system or the related service system. When necessary, the target server can send these software models to the corresponding target terminal, so that the target terminal can implement the above-mentioned by calling and running these software models. Figure 1 or Figure 5 The face recognition control method shown.

[0150] Since the operation of software requires the consumption of certain hardware resources, in order to ensure the normal operation of the above software model, the performance of the target terminal (such as memory size, data processing efficiency of the processor, etc.) needs to match it. Therefore, this embodiment uses the target server to execute the face recognition control effect monitoring method to screen the terminals that can execute the above face recognition control method. Specifically, the target server can select the appropriate target terminal through the preset selection rules and send the software models such as the preset generation model or the preset dynamic tracking model to it, so as to avoid the terminal with lower performance receiving the above software model but affecting the user experience due to the inability to operate normally, thereby ensuring the execution effect of the above face recognition control method from the source.

[0151] In addition, in fact, not all people have concerns about privacy leakage in the face recognition process. Therefore, the face recognition control method described in the previous embodiment can only be used for highly sensitive people who have concerns about privacy leakage in face recognition; the highly sensitive people can also be screened according to preset selection rules, that is, different regions, different fields, and different types of terminals may correspond to people with different sensitivities. Screening highly sensitive people means screening the target terminals that highly sensitive people may use.

[0152] In view of this, the above-mentioned preset selection rules for screening and determining the target terminal may include selection rules of one or more dimensions, specifically including but not limited to dimensions such as terminal performance and user tags of the scene where the terminal is located, and can be configured according to actual application requirements.

[0153] Specifically, in the terminal performance dimension, the terminal type or model, terminal scale (such as portable terminal or large fixed terminal, etc.), industry or scenario in which the terminal is used (such as cashier scenarios such as supermarkets and restaurants, company employee attendance or meeting sign-in scenarios, etc.) can be specifically limited; as long as the performance of the target terminal selected by the above limitations matches the operating requirements of the above-mentioned preset generation model and the preset dynamic tracking model.

[0154] In addition, in the user tag dimension, the user feature tags of users who use the face recognition service less frequently or sinking (silent churn) users can be analyzed in advance; since users who use the face recognition service less frequently or even do not use it for a long time (sinking users) can be considered to be highly sensitive users, that is, users who hope to be recalled and directed by the method described in this embodiment, these user feature tags can be used as a kind of selection rule to screen out user groups with the same characteristics, that is, highly sensitive people, and terminals that may be aimed at these highly sensitive people can be used as target terminals.

[0155] In this embodiment, according to actual application requirements, the selection rules of different dimensions can be combined in an “or” or “and” manner to form the above-mentioned preset selection rules.

[0156] It can be seen that compared with configuring the above-mentioned preset generation model and other software models on all terminals without selection, the face recognition control effect monitoring method provided in this embodiment selects the target terminals that can be configured with the above-mentioned software models from the dimensions of device performance, user tags, etc., which can not only ensure the normal operation of the above-mentioned software models on the target terminals, ensure the optimization effect of the face recognition process on the corresponding target terminals, bring about customer acquisition and drainage effects, but also save the maintenance costs of the face recognition system and related service systems.

[0157] In a possible implementation, the above-mentioned face recognition control effect monitoring method may further include:

[0158] Acquire preset monitoring parameters on each of the target terminals; the preset monitoring parameters include the number of face recognition times;

[0159] The preset generation model and / or the preset dynamic tracking model are optimized according to the preset monitoring parameters.

[0160] In this embodiment, after sending the software models such as the preset generation model to the selected target terminal, the target server also obtains preset monitoring parameters from each target terminal to evaluate and analyze the actual effect of the above model. Among them, the preset monitoring parameters may include the number of face recognition times N1, that is, the number of times the target terminal initiates face recognition (that is, the target terminal successfully captures the current image and sends it to the target server for face recognition); it may also include the number of times the target terminal initiates other authentication methods N2, the total number of times the target terminal initiates identity authentication Ns, etc., where Ns = N1 + N2.

[0161] The actual effect of the above-mentioned model is evaluated based on the above-mentioned preset monitoring parameters, which may specifically include: compared with before the configuration of the above-mentioned preset generation model and the preset dynamic tracking model, if the number of face recognition times N1 corresponding to a certain target terminal increases, or the proportion of face recognition times N1 / Ns increases, it means that the above-mentioned preset generation model and the preset dynamic tracking model have produced a certain customer acquisition and drainage effect, which can eliminate some users' concerns about privacy leakage; and the more N1 or N1 / Ns increases, the better the customer acquisition and drainage effect brought by the above-mentioned model.

[0162] Furthermore, in this embodiment, the above model can also be optimized according to the evaluation results obtained by the above preset monitoring parameters. Specifically, if the analysis based on the above preset monitoring parameters shows that no customer acquisition and drainage effect is produced (such as N1 or N1 / Ns does not increase), or the customer acquisition and drainage effect produced is limited (such as N1 or N1 / Ns increases less), then the optimization of the above preset generation model, preset dynamic tracking model and other models can be triggered, and then the optimized model can be resent to the corresponding target terminal to further increase the number of times users use face recognition on the target terminal and improve the customer acquisition and drainage effect of related services.

[0163] In other possible implementations, the preset monitoring parameters may also include other parameters that can be used to measure the actual effects of the software model, and the analysis based on these parameters is not limited to the changes in their own values.

[0164] In some implementations, the preset monitoring parameters may also include the number of low-activity users or sinking users (i.e., users who rarely use or have not used the face recognition service for a long time). Based on the change in this number before and after configuring the software model, it can be determined whether the application of the software model can transform low-activity users or sinking users into high-activity users, that is, whether the actual effect brought by the software model meets expectations, and further whether the software model needs to be optimized.

[0165] In some implementations, based on the number of face recognition times N1 corresponding to each target terminal, the influencing factors causing the difference between N1 corresponding to different target terminals can also be analyzed, and these influencing factors can be used to provide a reference direction for model optimization. For example, if the image style generated by the preset generation model configured on the target terminals corresponding to different N1, or the type of preset decorative elements loaded by the preset dynamic tracking model, etc. are different, then the image style generated by the preset generation model and the type of preset decorative elements can be regarded as the influencing factors causing the difference between N1 corresponding to different target terminals. When optimizing the above model, the image style (or a style similar to it) or the type of preset decorative element (or a type similar to it) with a larger corresponding N1 value can be configured in the model to be optimized.

[0166] In addition, it should be noted that in different application scenarios, the above-mentioned preset monitoring parameters may be different. For example, in the electronic payment scenario, the above-mentioned preset monitoring parameters may include the number of face-swiping payments (corresponding to the above-mentioned N1), the number of code scanning payments (corresponding to the above-mentioned N2), etc. In the access control scenario, the above-mentioned preset monitoring parameters may include the number of times the access control is opened by face-swiping (corresponding to the above-mentioned N1), the number of times the access control is opened by swiping a card (corresponding to the above-mentioned N2), etc.; therefore, the required preset monitoring parameters can be configured according to the actual application scenario to accurately monitor the execution effect of the above-mentioned software model and the corresponding face recognition control method.

[0167] To summarize, this embodiment aims at the situation where some users refuse to use facial recognition and related services due to concerns about privacy leakage in facial recognition. By executing the above-mentioned facial recognition control method on the target terminal that directly interacts with the user, the user's concerns about privacy leakage are weakened in each link of the entire facial recognition chain, thereby bringing about customer acquisition and drainage effects. In addition, by executing the above-mentioned facial recognition control effect monitoring method on the target service end, the actual effect of executing the above-mentioned facial recognition control method on each target terminal is monitored and evaluated, and the preset generation model involved in the above-mentioned facial recognition control method can be optimized according to the evaluation results to improve its customer acquisition and drainage effects.

[0168] Among them, through the above-mentioned face recognition control method, it is possible to achieve the weakening of users' concerns about privacy leakage in each link of the whole chain of face recognition, which is specifically reflected in: on the one hand, when the current image collected by the camera is displayed on the image acquisition page before the start of face recognition, the face outline of the user in the current image can be dynamically captured, and preset decorative elements can be loaded in the page according to the face outline, which not only increases the fun of the image acquisition page and helps to attract users, but also can partially block the user's face image displayed on the page through the preset decorative elements, that is, weakening the user's concerns about privacy leakage in the link before the start of face recognition; on the other hand, it can also be used in The current image is sent to the target server for comparison. That is, when face recognition is performed, the image acquisition page is switched to the status prompt page, and the current image is stylized through a preset generation model. The real face image of the user in the current image is processed into a non-real face image of a preset style, and displayed as a user avatar on the above status prompt page, that is, the user's concerns about privacy leakage are weakened in the status prompt link during face recognition; at the same time, the stylized image obtained by the above processing can also be used as the user avatar to be displayed in the service result page, which weakens the user's concerns about privacy leakage in the result display link after the face recognition and related services are completed.

[0169] It can be understood that the above embodiments are only examples and can be modified in actual implementation. Those skilled in the art can understand that the modification methods of the above embodiments without creative labor fall within the protection scope of one or more embodiments of this specification and will not be repeated in the embodiments.

[0170] Based on the same inventive concept, one or more embodiments of the present specification also provide a face recognition control device. Since the principle of the problem solved by the device is similar to that of the aforementioned face recognition control method, the implementation of the face recognition control device can refer to the implementation of the aforementioned face recognition control method, and the repeated parts will not be repeated.

[0171] Figure 7 This is a structural block diagram of a face recognition control device provided by one or more embodiments of this specification. Figure 7 As shown, the face recognition control device 400 may include:

[0172] The first display unit 401 is used to synchronously display the current image captured by the camera of the target device on the display screen of the target terminal through the image capture page;

[0173] The dynamic tracking unit 402 is used to dynamically capture the face contour in the current image; and load the preset decorative element in the image acquisition page according to the face contour.

[0174] In a possible implementation, the dynamic tracking unit 402 is used to load a preset decorative element in the image acquisition page according to the face contour, which may specifically include:

[0175] The dynamic tracking unit 402 is used to adjust the preset decorative element according to the size and / or position of the face contour.

[0176] In one possible implementation, Figure 8 As shown, the face recognition control device 400 also includes:

[0177] The first stylization unit 403 is used to input the current image into a preset generation model when receiving a face recognition instruction, and perform stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0178] A second display unit 404 is used to switch the image acquisition page to a status prompt page, and display the stylized image on the status prompt page;

[0179] Among them, the status prompt page also displays prompt information indicating the face recognition status.

[0180] In a possible implementation, the preset generation model involved in the first stylization unit 403 satisfies at least one of the following:

[0181] The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model;

[0182] The preset generation model includes a binary classification model for controlling the quality of image generation;

[0183] The preset generation model is an end-to-end model obtained based on distillation technology.

[0184] In the above-mentioned face recognition control device 400, by configuring the first display unit 401 and the dynamic tracking unit 402, the fun of the image acquisition page can be increased, which helps to attract users. The user's face image displayed in the image acquisition page can also be partially blocked by preset decorative elements, that is, the user's concerns about privacy leakage are weakened before the face recognition starts. According to actual application requirements, the first stylization unit 403 and the second display unit 404 can also be configured in the above-mentioned face recognition control device 400, so as to further weaken the user's concerns about privacy leakage in the state prompt link of face recognition in progress; in this way, the above-mentioned face recognition control device 400 can achieve the weakening of the user's concerns about privacy leakage in each link of the whole link of face recognition, bringing about the effect of attracting customers.

[0185] Fig. 9 This is a structural block diagram of another face recognition control device provided in one or more embodiments of this specification. Fig. 9 As shown, the face recognition control device 600 may include:

[0186] The second stylization unit 601 is used to input the current image collected by the camera of the target terminal into a preset generation model when receiving the face recognition instruction, and perform stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0187] The third display unit 602 is used to display the stylized image on a status prompt page;

[0188] The status prompt page also displays prompt information indicating the face recognition status; the status prompt page is displayed through the display screen of the target terminal.

[0189] In a possible implementation, the preset generation model involved in the second stylization unit 601 satisfies at least one of the following:

[0190] The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model;

[0191] The preset generation model includes a binary classification model for controlling the quality of image generation;

[0192] The preset generation model is an end-to-end model obtained based on distillation technology.

[0193] According to actual application requirements, the above-mentioned face recognition control device 600 can only be configured with the second stylization unit 601 and the third display unit 602, so as to reduce the user's concerns about privacy leakage in the face recognition status prompt link. Of course, if it is also necessary to reduce the user's concerns about privacy leakage in the image acquisition link before the face recognition starts, the above-mentioned face recognition control device 400 can be used.

[0194] In order to monitor the control effect of the above-mentioned face recognition control device 400 or 600, one or more embodiments of this specification also provide a face recognition control effect monitoring device. Fig.10 This is a structural block diagram of the face recognition control effect monitoring device. Fig.10 The face recognition control effect monitoring device 1100 specifically includes:

[0195] A selection unit 1101 is used to determine a target terminal according to a preset selection rule;

[0196] A sending unit 1102, configured to send a preset generation model and / or a preset dynamic tracking model to the target terminal;

[0197] The preset selection rule includes at least one of the following selection rules: terminal performance, user label of the scene where the terminal is located;

[0198] The preset generation model is used to perform stylization processing on the current image collected by the target terminal;

[0199] The preset dynamic tracking model is used to dynamically capture the face contour in the current image, and load and display preset decorative elements according to the face contour.

[0200] In a possible implementation, the face recognition control effect monitoring device 1100 may further include:

[0201] An acquisition unit, configured to acquire preset monitoring parameters on each of the target terminals; the preset monitoring parameters include the number of face recognition times;

[0202] An optimization unit is used to optimize the preset generation model and / or the preset dynamic tracking model according to the preset monitoring parameters.

[0203] Based on the same inventive concept, one or more embodiments of this specification also provide a service system based on face recognition, which may include: a target service end and at least one target terminal; the target service end and each target terminal may be connected as follows Figure 2 As shown, communication connection is carried out through the Internet of Things.

[0204] Among them, the target terminal in the service system can interact with the user to complete the user services related to face recognition provided by the system. In different service systems, the user services completed by the target terminal are also different; for example, the above-mentioned service system based on face recognition can be an electronic payment system, and the user services completed by the above-mentioned target terminal can include face payment, etc.; for another example, the above-mentioned service system based on face recognition can be a sign-in system, and the user services completed by the above-mentioned target terminal can include face sign-in, etc. In this embodiment, when the target terminal interacts with the user, face recognition control can be performed according to the face recognition control method described in any of the previous embodiments, which reduces the user's concerns about privacy leakage during face recognition.

[0205] The target server in the service system can monitor and optimize the face recognition control process of the target terminal according to the face recognition control effect monitoring method described in any of the above embodiments.

[0206] The face recognition-based service system provided in this embodiment controls the face recognition process through the target terminal, reduces user privacy concerns, monitors and optimizes the control effect of the target terminal through the target service end, realizes automatic control and optimization of the entire service system, and continuously improves the customer acquisition and flow effect.

[0207] See also Fig.11 , Fig.11 This is a structural block diagram of an electronic device provided in one or more embodiments of this specification. Fig.11 As shown, the electronic device 500 may include a processor 501 and a memory 502; the memory 502 may be coupled to the processor 501. It is worth noting that the Fig.11 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0208] In a possible implementation, the functions of the face recognition control device 400 may be integrated into the processor 501. The processor 501 may be configured to perform the following operations:

[0209] Synchronously displaying the current image captured by the camera of the target device on the display screen of the target terminal through an image capture page;

[0210] Dynamically capture the facial contour in the current image;

[0211] Preset decorative elements are loaded into the image acquisition page according to the facial contour.

[0212] In another possible implementation, the face recognition control device 400 may be configured separately from the processor 501 . For example, the face recognition control device 400 may be configured as a chip connected to the processor 501 , and face recognition control is implemented through the control of the processor 501 .

[0213] In a possible implementation, the functions of the face recognition control device 600 may be integrated into the processor 501. The processor 501 may be configured to perform the following operations:

[0214] When receiving a face recognition instruction, inputting a current image collected by a camera of the target terminal into a preset generation model, and performing stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image;

[0215] Displaying the stylized image on a status prompt page;

[0216] The status prompt page also displays prompt information indicating the face recognition status; the status prompt page is displayed through the display screen of the target terminal.

[0217] In another possible implementation, the face recognition control device 600 may be configured separately from the processor 501 . For example, the face recognition control device 600 may be configured as a chip connected to the processor 501 , and face recognition control may be implemented through the control of the processor 501 .

[0218] In a possible implementation, the functions of the face recognition control effect monitoring device 1100 may be integrated into the processor 501. The processor 501 may be configured to perform the following operations:

[0219] Determine the target terminal according to the preset selection rules;

[0220] Sending a preset generation model and / or a preset dynamic tracking model to the target terminal;

[0221] The preset selection rule includes at least one of the following selection rules: terminal performance, user label of the scene where the terminal is located;

[0222] The preset generation model is used to perform stylization processing on the current image collected by the target terminal;

[0223] The preset dynamic tracking model is used to dynamically capture the face contour in the current image, and load and display preset decorative elements according to the face contour.

[0224] In another possible implementation, the face recognition control effect monitoring device 1100 can be configured separately from the processor 501. For example, the face recognition control effect monitoring device 1100 can be configured as a chip connected to the processor 501, and face recognition control effect monitoring is implemented through the control of the processor 501.

[0225] In addition, in some optional implementations, the electronic device 500 may also include: a communication module, an input unit, an audio processor, a display, a power supply, etc. It is worth noting that the electronic device 500 does not necessarily have to include Fig.11 In addition, the electronic device 500 may also include Fig.11 For components not shown, reference may be made to the prior art.

[0226] In some optional implementations, the processor 501 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor device and / or logic device, which receives input and controls the operation of various components of the electronic device 500.

[0227] The memory 502 may be, for example, a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The memory 502 may store the information related to the face recognition control device 400 or 600 or the face recognition control effect monitoring device 1100, and may also store a program for executing the related information. The processor 501 may execute the program stored in the memory 502 to implement information storage or processing.

[0228] The input unit can provide input to the processor 501. The input unit is, for example, a key or a touch input device. The power supply can be used to provide power to the electronic device 500. The display can be used to display display objects such as images and text. The display can be, for example, an LCD display, but is not limited thereto.

[0229] The memory 502 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 502 may also be some other type of device. The memory 502 includes a buffer memory (sometimes referred to as a buffer). The memory 502 may include an application / function storage unit for storing application programs and function programs or processes for executing the operation of the electronic device 500 through the processor 501.

[0230] The memory 502 may also include a data storage unit for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit of the memory 502 may include various drivers for the communication function of the computer device and / or for executing other functions of the computer device (such as a messaging application, a contact book application, etc.).

[0231] The communication module is a transmitter / receiver that sends and receives signals via an antenna. The communication module (transmitter / receiver) is coupled to the processor 501 to provide input signals and receive output signals, which may be the same as in a conventional mobile communication terminal.

[0232] Based on different communication technologies, multiple communication modules may be provided in the same computer device, such as a cellular network module, a Bluetooth module and / or a wireless local area network module, etc. The communication module (transmitter / receiver) is also coupled to a speaker and a microphone via an audio processor to provide an audio output via the speaker and receive an audio input from the microphone, thereby realizing a common telecommunication function. The audio processor may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor is also coupled to the processor 501, so that recording can be performed on the machine through the microphone, and the sound stored on the machine can be played through the speaker.

[0233] One or more embodiments of this specification also provide a computer-readable storage medium capable of implementing all the steps in the face recognition control method or face recognition control effect monitoring method in the above embodiments, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, all the steps in the face recognition control method or face recognition control effect monitoring method in the above embodiments are implemented. The specific steps can be found in the above embodiments, and will not be repeated here.

[0234] Although one or more embodiments of this specification provide method operation steps as described in the embodiments or flowcharts, more or fewer operation steps may be included based on conventional or non-creative labor. The order of steps listed in the embodiments is only one way of executing the order of many steps and does not represent the only execution order. When the actual device or client product is executed, it can be executed in the order of the method shown in the embodiments or the drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment).

[0235] It should be understood by those skilled in the art that the embodiments of this specification may be provided as methods, devices (systems) or computer program products. Therefore, the embodiments of this specification may take the form of complete hardware embodiments, complete software embodiments or embodiments combining software and hardware. Moreover, one or more embodiments of this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0236] One or more embodiments of the present specification are described with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to one or more embodiments of the present specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0237] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0238] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1The steps for the functions specified in one or more boxes.

[0239] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0240] In this article, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. For those of ordinary skill in the art, the specific meaning of the above terms in one or more embodiments of this specification can be understood according to the specific circumstances.

[0241] It should be noted that, in the absence of conflict, one or more embodiments of this specification and the features in the embodiments may be combined with each other. One or more embodiments of this specification are not limited to any single aspect, nor to any single embodiment, nor to any combination and / or replacement of these aspects and / or embodiments. Moreover, each aspect and / or embodiment of one or more embodiments of this specification may be used alone or in combination with one or more other aspects and / or embodiments thereof.

[0242] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of one or more embodiments of this specification, rather than to limit them. Although one or more embodiments of this specification have been described in detail with reference to the aforementioned embodiments, a person of ordinary skill in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some or all of the technical features may be replaced by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of one or more embodiments of this specification, and they should all be included in the scope of the claims and the specification of one or more embodiments of this specification.

[0243] One or more embodiments of this specification are described above in conjunction with optional implementation methods, but these implementation methods are only exemplary and serve only as an illustration. On this basis, multiple replacements and improvements can be made to one or more embodiments of this specification, all of which fall within the scope of protection of one or more embodiments of this specification.

Claims

1. A face recognition control method, characterized in that: include: Synchronously displaying a current image captured by a camera of the target terminal on a display screen of the target terminal through an image capture page; When receiving a face recognition instruction, inputting the current image into a preset generation model, and performing stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image; Switching the image acquisition page to a status prompt page, and displaying the stylized image on the status prompt page; Among them, the status prompt page also displays prompt information indicating the face recognition status.

2. The method according to claim 1, characterized in that The preset generation model satisfies at least one of the following: The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model; The preset generation model includes a binary classification model for controlling the quality of image generation; The preset generation model is an end-to-end model obtained based on distillation technology.

3. A face recognition control method, characterized in that: include: When receiving a face recognition instruction, inputting a current image collected by a camera of the target terminal into a preset generation model, and performing stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image; Displaying the stylized image on a status prompt page; Among them, the status prompt page also displays prompt information indicating the face recognition status.

4. The method according to claim 3, characterized in that The preset generation model satisfies at least one of the following: The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model; The preset generation model includes a binary classification model for controlling the quality of image generation; The preset generation model is an end-to-end model obtained based on distillation technology.

5. A method for monitoring face recognition control effect, characterized in that: include: Determine the target terminal according to the preset selection rules; Sending the preset generation model to the target terminal; The preset generation model is used to perform stylized processing on the current image collected by the target terminal to obtain a stylized image corresponding to the current image, so that the target terminal switches the image collection page to the status prompt page, and displays the stylized image on the status prompt page; wherein, the status prompt page also displays prompt information indicating the face recognition status.

6. The method according to claim 5, characterized in that Also includes: Acquiring preset monitoring parameters on each of the target terminals; The preset monitoring parameters include the number of face recognition times; The preset generation model is optimized according to the preset monitoring parameters.

7. A face recognition control device, characterized in that: include: A first display unit, configured to synchronously display a current image captured by a camera of the target terminal on a display screen of the target terminal through an image capture page; A first stylization unit is configured to input the current image into a preset generation model upon receiving a face recognition instruction, and perform stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image; A second display unit, used to switch the image acquisition page to a status prompt page, and display the stylized image on the status prompt page; Among them, the status prompt page also displays prompt information indicating the face recognition status.

8. The device according to claim 7, characterized in that The preset generation model satisfies at least one of the following: The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model; The preset generation model includes a binary classification model for controlling the quality of image generation; The preset generation model is an end-to-end model obtained based on distillation technology.

9. A face recognition control device, characterized in that: include: A second stylization unit is used to input a current image collected by a camera of a target terminal into a preset generation model upon receiving a face recognition instruction, and perform stylization processing on the current image through the preset generation model to obtain a stylized image corresponding to the current image; A third display unit, used to display the stylized image on a status prompt page; Among them, the status prompt page also displays prompt information indicating the face recognition status.

10. The device according to claim 9, characterized in that The preset generation model satisfies at least one of the following: The preset generation model is an image generation model obtained by fine-tuning the Stable Diffusion model based on the Dream Booth model; The preset generation model includes a binary classification model for controlling the quality of image generation; The preset generation model is an end-to-end model obtained based on distillation technology.

11. A face recognition control effect monitoring device, characterized in that: include: A selection unit, used to determine a target terminal according to a preset selection rule; A sending unit, used for sending the preset generation model to the target terminal; The preset generation model is used to perform stylized processing on the current image collected by the target terminal to obtain a stylized image corresponding to the current image, so that the target terminal switches the image collection page to the status prompt page, and displays the stylized image on the status prompt page; wherein, the status prompt page also displays prompt information indicating the face recognition status.

12. The device according to claim 11, characterized in that Also includes: An acquisition unit, used for acquiring preset monitoring parameters on each of the target terminals; The preset monitoring parameters include the number of face recognition times; An optimization unit is used to optimize the preset generation model according to the preset monitoring parameters.

13. A service system based on face recognition, characterized in that: include: A target server and at least one target terminal; The target terminal is used to perform face recognition control according to any one of the methods of claims 1 to 4 when interacting with a user; The target server is used to monitor and optimize the effect of the face recognition control process of the target terminal according to the method described in any one of claims 5 to 6.

14. An electronic device, characterized in that: The electronic device comprises: a memory for storing a computer program product; A processor is used to execute the computer program product stored in the memory, and when the computer program product is executed, the method described in any one of claims 1 to 6 is implemented.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, which, when executed, implement the method described in any one of claims 1 to 6.

Images