Model training method and device, request processing method and device, equipment and medium

Abstract

The invention provides a model training method and device, a request processing method and device, equipment and a medium, and the method comprises the steps: responding to a target transaction request operation initiated by a user, and calling a camera to collect a first video comprising a user face; performing first risk analysis on face information contained in an image foreground in the first video and environment information contained in an image background in the first video, and determining a first risk index value of the first video; according to the first risk index value of the first video, determining whether the first video passes verification; under the condition that the verification of the first video is passed, feedback information of a first face recognition verification mode adopted for the first transaction request before the first video is collected is acquired; wherein the feedback information is determined based on a first face recognition verification result of the first face recognition verification mode and a first risk recognition result of the first transaction, and the first transaction request is used for processing the first transaction; and training the mode prediction model according to the feedback information.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a model training method, a request processing method and apparatus, a device and a medium. Background Art

[0002] With the development of computer technology, face recognition technology has been widely used in many fields, providing convenience for identity authentication, payment security, etc. At present, the mainstream face recognition strategy is still based on static / dynamic face recognition based on established patterns. However, there are some users who regard bypassing face recognition to pass identity authentication as a major research direction. With the popularization of artificial intelligence image / video generation tools, this established pattern of static / dynamic face recognition technology has brought great security challenges.

[0003] In order to ensure the security and reliability of facial recognition technology, it is necessary to continuously strengthen technological research and development. Summary of the invention

[0004] The present invention provides a model training method, a request processing method, an apparatus, a device and a medium.

[0005] According to a first aspect of an embodiment of the present disclosure, a model training method is provided, the method comprising:

[0006] In response to a target transaction request operation initiated by a user, calling a camera to capture a first video including a face of the user;

[0007] Performing a first risk analysis on facial information contained in the foreground of the image in the first video and environmental information contained in the background of the image in the first video to determine a first risk index value of the first video;

[0008] Determining whether the first video passes verification according to a first risk indicator value of the first video;

[0009] When the first video verification passes, obtaining feedback information of a first face recognition verification mode adopted by a first transaction request before the first video is collected; wherein the feedback information is determined based on a first face recognition verification result of the first face recognition verification mode and a first risk identification result of the first transaction, and the first transaction request is used to process the first transaction;

[0010] The pattern prediction model is trained according to the feedback information.

[0011] According to a second aspect of an embodiment of the present disclosure, a request processing method is provided, the method comprising:

[0012] In response to a target transaction request operation initiated by a user, calling a camera to capture a first video including a face of the user;

[0013] If the first video verification passes, obtaining a first face recognition verification mode used for the first transaction request before the first video is captured;

[0014] Based on the first face recognition verification mode, using a pattern prediction model, determining a second face recognition verification mode for performing face recognition verification on the user;

[0015] Performing face recognition verification on the user using the second face recognition verification mode to obtain a target face recognition verification result;

[0016] According to the target face recognition verification result, the transaction processing corresponding to the target transaction request operation is executed.

[0017] According to a third aspect of an embodiment of the present disclosure, a model training device is provided, the device comprising:

[0018] A calling module, configured to call a camera to capture a first video including a user's face in response to a target transaction request operation initiated by the user;

[0019] An analysis module, configured to perform a first risk analysis on facial information contained in the foreground of the image in the first video and environmental information contained in the background of the image in the first video to determine a first risk index value of the first video;

[0020] A determination module, configured to determine whether the first video has passed verification according to a first risk indicator value of the first video;

[0021] an acquisition module, configured to acquire feedback information of a first face recognition verification mode adopted by a first transaction request before the first video is collected, if the first video verification passes; wherein the feedback information is determined based on a first face recognition verification result of the first face recognition verification mode and a first risk identification result of the first transaction, and the first transaction request is used to process the first transaction;

[0022] The training module is used to train the pattern prediction model according to the feedback information.

[0023] According to a fourth aspect of an embodiment of the present disclosure, a request processing device is provided, the device comprising:

[0024] A calling module, configured to call a camera to capture a first video including a user's face in response to a target transaction request operation initiated by the user;

[0025] an acquisition module, configured to acquire, when the first video verification passes, a first face recognition verification mode used for the first transaction request before the first video is collected;

[0026] A determination module, configured to determine, based on the first face recognition verification mode, a second face recognition verification mode for performing face recognition verification on the user using a pattern prediction model;

[0027] A verification module, configured to perform face recognition verification on the user using the second face recognition verification mode to obtain a target face recognition verification result;

[0028] A processing module is used to execute transaction processing corresponding to the target transaction request operation according to the target face recognition verification result.

[0029] According to the fifth aspect of the embodiments of the present disclosure, an electronic device is provided, comprising: a processor, and a memory communicatively connected to the processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory to implement the model training method disclosed in the embodiment of the first aspect of the present disclosure, or the request processing method disclosed in the embodiment of the second aspect of the present disclosure.

[0030] According to the fourth aspect of an embodiment of the present disclosure, a computer-readable medium is provided, in which computer execution instructions are stored. When the computer execution instructions are executed by a processor, they are used to implement the model training method disclosed in the embodiment of the first aspect of the present disclosure, or the request processing method disclosed in the embodiment of the second aspect of the present disclosure.

[0031] According to the fifth aspect of the embodiments of the present disclosure, the present disclosure provides a computer program product, including a computer program, which, when executed by a processor, implements the model training method disclosed in the embodiment of the first aspect of the present disclosure, or the request processing method disclosed in the embodiment of the second aspect of the present disclosure.

[0032] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects:

[0033] In response to a target transaction request operation initiated by a user, a camera is called to capture a first video including the user's face; a first risk analysis is performed on the facial information contained in the foreground of the image in the first video and the environmental information contained in the background of the image in the first video to determine the first risk index value of the first video; based on the first risk index value of the first video, it is determined whether the first video has passed the verification; if the first video has passed the verification, feedback information of the first face recognition verification mode adopted for the first transaction request before the first video is collected is obtained; wherein the feedback information is determined based on the first face recognition verification result of the first face recognition verification mode and the first risk identification result of the first transaction, and the first transaction request is used to process the first transaction; based on the feedback information, a pattern prediction model is trained. Thus, the pattern prediction model can be automatically trained based on the feedback information, so that the trained pattern prediction model can learn and generalize a face recognition verification mode that adapts to the actual situation and risk level of the user.

[0034] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0036] Figure 1 is a flowchart of the model training method shown in the first embodiment of the present disclosure;

[0037] Figure 2 is a flow chart of a model training method shown in the second embodiment of the present disclosure;

[0038] Figure 3 is a flowchart of a model training method shown in the third embodiment of the present disclosure;

[0039] Figure 4 is a flowchart of a model training method shown in the fourth embodiment of the present disclosure;

[0040] Figure 5 is a flowchart of a request processing method shown in the fifth embodiment of the present disclosure;

[0041] Figure 6 is a structural schematic diagram of a model training device shown in the sixth embodiment of the present disclosure;

[0042] Figure 7 is a schematic diagram of the structure of a request processing device shown in the seventh embodiment of the present disclosure;

[0043] Figure 8It is a schematic structural diagram of an electronic device shown in an exemplary embodiment of the present disclosure.

[0044] The above drawings show clear embodiments of the present disclosure, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present disclosure in any way, but to illustrate the concepts of the present disclosure to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0045] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0046] It should be noted that the acquisition, storage, use, and processing of data in the technical solution disclosed in this disclosure are in compliance with the relevant provisions of national laws and regulations.

[0047] At present, in transaction scenarios, the main facial recognition strategy is still static / dynamic facial recognition in a fixed pattern. However, this fixed pattern facial recognition strategy is easily cracked by brute force and crashed, has limited risk identification capabilities, and relies on specific features, resulting in a high recognition failure rate. In addition, because the facial recognition strategies (or facial recognition patterns) used in high-risk scenarios and low-risk scenarios are similar, the computing resource utilization efficiency is low.

[0048] In response to at least one of the above problems, the present disclosure proposes a model training method, request processing method, device, equipment and medium.

[0049] The following describes the model training method, request processing method, device, equipment and medium of the embodiments of the present disclosure with reference to the accompanying drawings.

[0050] Figure 1 It is a flowchart of the model training method shown in the first embodiment of the present disclosure.

[0051] It should be noted that the executor of the model training method provided in the embodiment of the present disclosure is a model training device, which can be implemented by software and / or hardware. The model training device in this embodiment can be an electronic device, or can be configured in an electronic device.

[0052] The electronic device in the embodiment of the present disclosure may include a terminal device, a server, etc. For example, the terminal device may be a mobile phone, a tablet computer, a personal digital assistant, or other hardware device with various operating systems, a touch screen, and / or a display screen. The embodiment does not specifically limit the electronic device. In this example, the electronic device is a server as an example for exemplary description.

[0053] like Figure 1 As shown, the model training method may include the following steps S101-S105.

[0054] Step S101, in response to a target transaction request operation initiated by a user, calling a camera to capture a first video including the user's face.

[0055] The target transaction request operation may be, for example, a payment operation, a transfer operation, etc., which is not limited in the present disclosure.

[0056] It is understandable that after the user initiates a target transaction request operation, the execution subject of the present disclosure can call the camera and capture a first video including the user's face.

[0057] Optionally, after calling the camera to capture the video, the captured video can be detected to determine whether the captured video includes the user's face. If it is determined that the captured video includes the user's face, the captured video is determined to be the first video.

[0058] As an example, in order to accurately determine whether the captured video includes the user's face, when the captured video includes multiple image frames, for any image frame, DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is used to perform image segmentation to obtain multiple sub-images of the image frame; for any sub-image, a random forest algorithm is used to determine whether the sub-image is a face image; when there is a target sub-image of a face image among the multiple sub-images of the multiple image frames, it is determined that the captured video is the first video including the user's face.

[0059] Step S102, performing a first risk analysis on facial information contained in the foreground of the image in the first video and environmental information contained in the background of the image in the first video to determine a first risk index value of the first video.

[0060] Optionally, first, emotion recognition can be performed on the facial information of the user contained in the image foreground of the first video to obtain the user's emotional state; secondly, head recognition can be performed on the first video to obtain the number of heads; thirdly, the environmental information contained in the image background of the first video is analyzed to obtain the environmental category of the image background; finally, based on the user's emotional state, the number of heads and the environmental category, the first risk indicator value of the first video can be determined.

[0061] Among them, the user's emotional state can be a real or intuitive description of the user's emotional experience and mental state, such as basic emotions such as joy, anger, sadness, happiness, surprise, fear, love, and hate, or it can also be a positive, negative or neutral emotional tendency.

[0062] The environment categories may include, for example, a dark environment, a bright environment, an indoor environment, an outdoor environment, a clear environment, a blurry environment, etc., and the present disclosure does not impose any limitation on this.

[0063] As an example, after obtaining the user's emotional state, the number of people, and the environmental category, the user's emotional state, the number of people, and the environmental category can be input into the indicator prediction model to predict the first risk indicator value of the first video. Thus, relying on the indicator prediction model to predict the first risk indicator value of the first video can improve the accuracy and reliability of the first risk indicator value.

[0064] In this way, risk analysis of facial information contained in the image foreground of the first video and environmental information contained in the image background of the first video can be achieved, and the first risk index value of the first video can be effectively obtained.

[0065] Step S103, determining whether the first video has passed verification according to the first risk indicator value of the first video.

[0066] As an example, the first risk indicator value of the first video is compared with the first set threshold value, and whether the first video has passed the verification is determined based on the comparison result. For example, when the comparison result indicates that the first risk indicator value of the first video is greater than the first set threshold value, it is determined that the first video has passed the verification, otherwise, the verification has failed.

[0067] Step S104, when the first video verification passes, obtaining feedback information of the first face recognition verification mode adopted by the first transaction request before the first video is collected.

[0068] The feedback information may be determined based on the first face recognition verification result of the first face recognition verification mode and the first risk identification result of the first transaction, and the first transaction request may be used to process the first transaction.

[0069] The first transaction request occurs before the first video is captured.

[0070] Among them, the first face recognition verification mode is the face recognition verification mode generated when processing the first transaction request; among them, the face recognition verification mode may include a static image and / or a dynamic video including at least one action (such as turning the head, nodding, opening the mouth, blinking, smiling, etc.).

[0071] As an example, the first face recognition verification mode is "static image + turn head to the left + open mouth", or the first face recognition verification mode is "turn head to the right + static image + smile + open mouth + blink", etc. It should be noted that the above examples of the first face recognition verification mode are only exemplary.

[0072] Among them, the first face recognition verification result can be, for example, "no risk, verification passed", "risky, verification failed", etc., and the present disclosure does not impose any restrictions on this.

[0073] Among them, the first risk identification result can be, for example, "risky transaction", "risk-free transaction", "transaction with unidentifiable risk", or "abnormal transaction", "normal transaction", "transaction with unidentifiable risk", etc.

[0074] As an example, when the first face recognition verification result is "risk-free, verification passed" and the first risk identification result is "risk-free transaction", the feedback information may be "correct identification" or "correct verification"; when the first face recognition verification result is "risky, verification failed" and the first risk identification result is "risk-free transaction", the feedback information may be "identifying risk-free transactions as risky transactions" or "identifying normal transactions as risky transactions"; when the first face recognition result is "risk-free, verification passed" and the first risk identification result is "risky transaction", the feedback information may be "identifying risky transactions as risk-free transactions" or "identifying risky transactions as normal transactions".

[0075] It should be noted that the above examples of feedback information are merely illustrative, and in practical applications, there may be other examples, which are not limited in the present disclosure.

[0076] In the embodiment of the present disclosure, when the first video verification passes, feedback information of the first face recognition verification mode adopted by the first transaction request before the first video is collected can be obtained. Therefore, the subsequent process is executed only when the first video verification passes, instead of executing the model training process when the user initiates any transaction request operation, which can effectively avoid the waste of computing resources.

[0077] Step S105: training the pattern prediction model according to the feedback information.

[0078] Among them, the pattern prediction model can be used to predict the pattern used for face recognition verification, that is, predict the face recognition verification pattern.

[0079] The model training method of the disclosed embodiment, in response to the target transaction request operation initiated by the user, calls the camera to collect the first video including the user's face; performs a first risk analysis on the facial information contained in the image foreground of the first video and the environmental information contained in the image background of the first video to determine the first risk index value of the first video; determines whether the first video has passed the verification according to the first risk index value of the first video; if the first video has passed the verification, obtains feedback information of the first face recognition verification mode adopted for the first transaction request before the first video is collected; wherein the feedback information is determined based on the first face recognition verification result of the first face recognition verification mode and the first risk identification result of the first transaction, and the first transaction request is used to process the first transaction; and trains the pattern prediction model according to the feedback information. Thus, the pattern prediction model can be automatically trained based on the feedback information, so that the trained pattern prediction model can learn and generalize the face recognition verification mode that adapts to the actual situation and risk level of the user.

[0080] In order to clearly explain how to obtain the first risk identification result of the first transaction in the above embodiment of the present disclosure, the present disclosure also proposes a model training method.

[0081] Figure 2 It is a flowchart of the model training method shown in the second embodiment of the present disclosure.

[0082] like Figure 2 As shown, based on the previous embodiment of the present disclosure, the model training method may further include the following steps S201-S203.

[0083] Step S201, collecting transaction data generated during the first transaction request processing process.

[0084] Among them, transaction data may include but is not limited to channel, transaction code, ID number, bank card number, mobile phone number, transaction account number, device ID (Identity document), IP (Internet Protocol), session ID (Identity document), geographic location, transaction details, transaction time, etc.

[0085] Among them, it should be noted that the above-mentioned acquisition, storage, use, and processing of data are in compliance with the relevant provisions of relevant laws and regulations.

[0086] In the disclosed embodiment, the transaction data generated during the first transaction request processing may be collected. For example, the burying point technology may be used to collect the transaction data generated during the first transaction request processing.

[0087] Step S202: Based on the transaction data, a transaction risk identification model is used to perform a first risk identification on the first transaction to obtain a second risk identification result of the first transaction.

[0088] Among them, the second risk identification result is, for example, "risky transaction", "risk-free transaction", "unable to identify risk", etc.

[0089] As an example, transaction features can be extracted from transaction data and input into a transaction risk identification model, thereby obtaining a second risk identification result of the first transaction output by the transaction risk identification model, thereby completing the risk identification of the first transaction.

[0090] Step S203: Generate a first risk identification result based on the second risk identification result.

[0091] As an example, the second risk identification result may be determined as the first risk identification result.

[0092] It is understandable that there may be a situation where the second risk identification result indicates that the risk cannot be identified or cannot be determined. In order to maintain the accuracy and reliability of the first risk identification result, in a possible implementation method of the present disclosure, when the second risk identification result indicates that the risk cannot be determined, a risk identification request can be generated based on the transaction data, and the risk identification request can be sent to the target terminal, so that the target terminal performs a second risk identification on the first transaction based on the transaction data, and obtains a third risk identification result of the first transaction; finally, the first risk identification result can be generated based on the third risk identification result.

[0093] Among them, the third risk identification result is, for example, "risky transaction", "risk-free transaction", "unable to identify risk", etc.

[0094] As an example, when the second risk identification result indicates that the risk cannot be determined, a risk identification request carrying the transaction data can be sent to the target terminal corresponding to the relevant risk control staff. Thus, the risk control staff can identify the risk of the first transaction based on the transaction data displayed by the target terminal, thereby obtaining the third risk identification result of the first transaction. Finally, the third risk identification result of the first transaction can be determined as the first risk identification result.

[0095] Thus, the risk of the transaction data of the first transaction is further identified through the target terminal, thereby improving the accuracy and flexibility of determining the first risk identification result.

[0096] The model training method of the embodiment of the present disclosure collects transaction data generated during the first transaction request processing; based on the transaction data, a transaction risk identification model is used to perform a first risk identification on the first transaction to obtain a second risk identification result of the first transaction; and a first risk identification result is generated according to the second risk identification result. Thus, the first risk identification result of the first transaction is effectively determined by relying on the transaction data generated during the first transaction request processing and by using the transaction risk identification model.

[0097] In order to clearly explain how to obtain the first face recognition verification result of the first face recognition verification mode in the above embodiment of the present disclosure, the present disclosure also proposes a model training method.

[0098] Figure 3 It is a flowchart of the model training method shown in the third embodiment of the present disclosure.

[0099] like Figure 3 As shown, based on any of the above embodiments of the present disclosure, the model training method may further include the following steps S301-S304.

[0100] Step S301, capturing a second video in a first face recognition verification mode.

[0101] Step S302: Perform a second risk analysis on the second video to obtain a second risk index value of the second video.

[0102] Among them, the higher the second risk index value is, the more it is proportional to the risk level of the second video, that is, the higher the second risk index value is, the higher the risk level of the second video is, and the lower the second risk index value is, the lower the risk level of the second video is.

[0103] In order to clearly explain how to perform the second risk analysis on the second video and obtain the second risk index value of the second video, the following steps may be used to determine the second risk index value of the second video:

[0104] Step S3021, extracting features of multiple dimensions from the second video.

[0105] In the embodiments of the present disclosure, features of multiple dimensions may include, for example, user biometric features and user behavioral features; wherein, user biometric features may include, for example, the position of facial features, face shape, angle, position and shape of teeth, etc., and the present disclosure does not impose any restrictions on this.

[0106] Step S3022: Input the features of any dimension into the corresponding prediction branch in the image risk identification model to obtain the sub-risk index value corresponding to the dimension.

[0107] As an example, assume that there are features in multiple dimensions, including user biometrics and user behavior features; the user biometrics are input into the corresponding prediction branch in the image risk identification model to obtain the sub-risk indicator value corresponding to the dimension. For example, the similarity between the user biometrics and the registered biometrics in the database can be calculated through the prediction branch corresponding to the dimension in the image risk identification model, and the similarity is used as the sub-risk indicator value of the prediction branch; the user behavior features are input into the corresponding prediction branch in the image risk identification model to obtain the sub-risk indicator value corresponding to the dimension. For example, the prediction branch corresponding to the dimension in the image risk identification model can be used to analyze and determine the user's action completion in the first face recognition verification mode based on the user behavior features, and the action completion is used as the sub-risk indicator value of the prediction branch.

[0108] Step S3023: Determine the second risk index value of the second video according to the sub-risk index values ​​corresponding to each dimension.

[0109] As an example, a weighted sum of the sub-risk index values ​​corresponding to multiple dimensions may be performed to obtain a second risk index value of the second video.

[0110] Step S303: Perform a first risk analysis on the login account associated with the first transaction request to determine an account risk index.

[0111] As a possible implementation method, target behavior data can be extracted from the corresponding historical transaction data of the user corresponding to the login account, and a first indicator value can be determined based on the target behavior data; a second indicator value corresponding to the basic information can be determined based on the basic information of the user corresponding to the login account; finally, the account risk index can be determined based on the first indicator value and the second indicator value.

[0112] The first indicator value may be used to indicate the behavior score of the target behavior data of the user corresponding to the login account. It should be noted that the higher the first indicator value, the higher the behavior score of the target behavior data of the user corresponding to the login account.

[0113] The target behavior data may include, but is not limited to, the number of transactions, the average transaction amount, the number of normal logins, the number of failed logins, the number of transaction interface accesses, etc.

[0114] The second indicator value may be used to indicate the value score of the user corresponding to the login account.

[0115] Step S304: determining a first face recognition verification result of a first face recognition verification mode according to the second risk indicator value and / or the account risk index.

[0116] As an example, first, the second risk indicator value and the account risk index can be weighted and summed to obtain a third risk indicator value. Secondly, the third risk indicator value and the second set threshold are compared. Finally, based on the comparison result, the first face recognition verification result of the first face recognition verification mode is determined.

[0117] As another example, the second risk indicator value is compared with the second set threshold value, and the first face recognition verification result of the first face recognition verification mode is determined according to the comparison result.

[0118] It is understandable that there may be a situation where the third risk indicator value indicates that the user has a high risk level in this transaction or the risk cannot be clearly determined. At this time, in order to prevent verification errors as much as possible and improve the fault tolerance rate, in a possible implementation method of the present embodiment, based on the second risk indicator value and / or the account risk index, it is determined whether facial recognition verification is needed again; if it is determined that facial recognition verification is needed again, facial recognition verification is performed again using the second facial recognition verification mode to obtain a second facial recognition verification result of the second facial recognition verification mode; based on the second facial recognition verification result of the second facial recognition verification mode, the first facial recognition verification result of the first facial recognition verification mode is generated.

[0119] As an example, the second risk indicator value and the account risk index can be weighted and summed to obtain a third risk indicator value; when the third risk indicator value is greater than the third set threshold, or the third risk indicator value is less than the fourth set threshold, it is determined that face recognition verification is needed again, and the second face recognition verification mode can be determined based on the first face recognition verification mode and the third risk indicator value through a reinforcement learning algorithm (such as a Monte Carlo reinforcement learning algorithm), and face recognition verification can be performed again using the second face recognition verification mode to obtain a second face recognition verification result of the second face recognition verification mode; the second face recognition verification result of the second face recognition verification mode is determined as the first face recognition verification result of the first face recognition verification mode.

[0120] Therefore, on the one hand, the first face recognition verification result of the first face recognition verification mode can be effectively determined, and on the other hand, the fault tolerance rate and the accuracy and reliability of the face recognition verification result can be improved.

[0121] The model training method of the disclosed embodiment collects a second video in the first face recognition verification mode; performs a second risk analysis on the second video to obtain a second risk index value of the second video; performs a first risk analysis on the login account associated with the first transaction request to determine the account risk index; and determines the first face recognition verification result of the first face recognition verification mode according to the second risk index value and / or the account risk index. Thus, the face recognition verification result obtained is more accurate and reliable by combining the risk analysis of the video and the risk analysis of the login account.

[0122] In order to clearly illustrate how the pattern prediction model is trained based on feedback information in any embodiment of the present disclosure, the present disclosure also proposes a model training method.

[0123] Figure 4 It is a flowchart of the model training method shown in the fourth embodiment of the present disclosure.

[0124] like Figure 4 As shown, the model training method may also include the following steps S401-S406.

[0125] Step S401, in response to a target transaction request operation initiated by a user, calling a camera to capture a first video including the user's face.

[0126] Step S402, performing a first risk analysis on facial information contained in the foreground of the image in the first video and environmental information contained in the background of the image in the first video to determine a first risk index value of the first video.

[0127] Step S403: determining whether the first video has passed verification according to the first risk indicator value of the first video.

[0128] Step S404, when the first video verification passes, obtaining feedback information of the first face recognition verification mode adopted by the first transaction request before the first video is collected.

[0129] It should be noted that the execution process of step S401 to step S404 can refer to the execution process of any embodiment of the present disclosure, and will not be repeated here.

[0130] Step S405: determining a corresponding reward coefficient according to the feedback information.

[0131] As an example, when the feedback information indicates that the verification is accurate, the reward coefficient is determined to be an upward setting coefficient; when the feedback information indicates that the risky first transaction is identified as a normal transaction, the reward coefficient is determined to be a first downward setting coefficient; when the feedback information indicates that the normal first transaction is identified as a risky transaction, the reward coefficient is determined to be a second downward setting coefficient; when the feedback information indicates that the static face recognition verification mode is used additionally, the reward coefficient is determined to be a third downward setting coefficient; when the feedback information indicates that the dynamic face recognition verification mode is used additionally, the reward coefficient is determined to be a fourth downward setting coefficient.

[0132] Among them, the upward setting coefficient, the first downward setting coefficient, the second downward setting coefficient, the third downward setting coefficient, and the fourth downward setting coefficient can all be pre-set, and the present disclosure does not limit the values ​​of the upward setting coefficient, the first downward setting coefficient, the second downward setting coefficient, the third downward setting coefficient, and the fourth downward setting coefficient.

[0133] As an example, when the feedback information indicates that the verification is accurate, the reward coefficient is determined to be +10; when the feedback information indicates that the risky first transaction is identified as a normal transaction, the reward coefficient is determined to be -10; when the feedback information indicates that the normal first transaction is identified as a risky transaction, the reward coefficient is determined to be -5; when the feedback information indicates that the static face recognition verification mode is used additionally, the reward coefficient is determined to be -1; when the feedback information indicates that the dynamic face recognition verification mode is used additionally, the reward coefficient is determined to be -2.

[0134] Therefore, corresponding reward coefficients can be set and determined for different feedback information.

[0135] Step S406: fine-tune the pattern prediction model based on the reward coefficient.

[0136] As a possible implementation method, the original cumulative reward score before fine-tuning the pattern prediction model is obtained; the original cumulative reward score and the reward coefficient are added to obtain the target cumulative reward score; thus, the pattern prediction model is fine-tuned based on the target cumulative reward score.

[0137] Therefore, with the help of the reward coefficient, the pattern prediction model is fine-tuned so that the fine-tuned pattern prediction model can better adapt to the needs of actual application scenarios, and the generated face recognition verification pattern can be better suitable for the specific circumstances of the transaction users and not easily cracked by brute force.

[0138] Any of the above embodiments is a method for training a model, and the present disclosure also provides a method for applying the model.

[0139] Figure 5 It is a flowchart of a request processing method shown in the fifth embodiment of the present disclosure.

[0140] like Figure 5 As shown, the request processing method may include the following steps S501-S505.

[0141] Step S501, in response to a target transaction request operation initiated by a user, calling a camera to capture a first video including the user's face.

[0142] Step S502, when the first video verification passes, obtain the first face recognition verification mode used for the first transaction request before the first video is captured.

[0143] Optionally, a first risk analysis is performed on facial information contained in the foreground of the image in the first video and environmental information contained in the background of the image in the first video to determine a first risk index value of the first video; based on the first risk index value of the first video, it is determined whether the first video has passed the verification.

[0144] Step S503: Based on the first face recognition verification mode, a pattern prediction model is used to determine a second face recognition verification mode for performing face recognition verification on the user.

[0145] Optionally, feedback information of the first face recognition verification mode can be obtained; the corresponding reward coefficient can be determined according to the feedback information; the reward coefficient corresponding to the first face recognition verification mode, the first face recognition verification mode, and the first risk indicator value can be input into the model prediction model to obtain the second face recognition verification mode output by the model. Thus, the generated second face recognition verification mode is adapted to the user's transaction situation and the risks existing in the transaction scenario, so that the second face recognition verification mode generated in a high-risk scenario can be more complex, and the second face recognition verification mode generated in a low-risk scenario can be simpler.

[0146] Step S504: perform face recognition verification on the user using the second face recognition verification mode to obtain a target face recognition verification result.

[0147] Among them, the target face recognition verification result is, for example, "no risk, verification passed", "risky, verification failed", etc., and this disclosure does not impose any restrictions on this.

[0148] Step S505: Execute transaction processing corresponding to the target transaction request operation according to the target face recognition verification result.

[0149] As an example, assuming that the target face recognition verification result is "no risk, verification passed", the transaction processing corresponding to the target transaction request operation is executed; on the contrary, assuming that the target face recognition verification result is "risky, verification failed", the transaction processing corresponding to the target transaction request operation is refused.

[0150] The request processing method of the disclosed embodiment, in response to the target transaction request operation initiated by the user, calls the camera to collect the first video including the user's face; in the case where the first video verification passes, obtains the first face recognition verification mode used for the first transaction request before the first video is collected; based on the first face recognition verification mode, adopts the pattern prediction model to determine the second face recognition verification mode for face recognition verification of the user; adopts the second face recognition verification mode to perform face recognition verification on the user to obtain the target face recognition verification result; according to the target face recognition verification result, executes the transaction processing corresponding to the target transaction request operation. Thus, the face recognition verification mode suitable for the user can be intelligently determined, effectively avoiding the interference and influence of the face recognition process caused by the pre-made video, and due to the real-time and timeliness of the face recognition verification mode generation, it can effectively avoid the situation of being cracked by force; and the face recognition verification mode used in high-risk scenarios and low-risk scenarios is different, in the low-risk scenario, a simpler face recognition verification mode can be used, and in the high-risk scenario, a more complex face recognition verification mode can be used, on the one hand, the accuracy of face recognition verification is improved, and on the other hand, the waste of computing resources can be effectively avoided.

[0151] With the above Figures 1 to 4 Corresponding to the model training method provided in the embodiment, the present disclosure also provides a model training device. Figures 1 to 4 The model training method provided in the embodiment corresponds to the embodiment, so the implementation method of the model training method is also applicable to the model training device provided in the embodiment of the present disclosure, and will not be described in detail in the embodiment of the present disclosure.

[0152] Figure 6 It is a structural diagram of the model training device shown in the sixth embodiment of the present disclosure.

[0153] Reference Figure 6 The model training device 600 may include: a calling module 601, an analyzing module 602, a determining module 603, an acquiring module 604 and a training module 605.

[0154] The calling module 601 is used to call a camera to capture a first video including the user's face in response to a target transaction request operation initiated by the user.

[0155] The analysis module 602 is used to perform a first risk analysis on the facial information contained in the foreground of the image in the first video and the environmental information contained in the background of the image in the first video to determine a first risk index value of the first video.

[0156] The determination module 603 is used to determine whether the first video has passed the verification according to the first risk indicator value of the first video.

[0157] The acquisition module 604 is used to obtain feedback information of the first face recognition verification mode adopted by the first transaction request before the first video is collected when the first video verification passes; wherein the feedback information is determined based on the first face recognition verification result of the first face recognition verification mode and the first risk identification result of the first transaction, and the first transaction request is used to process the first transaction.

[0158] The training module 605 is used to train the pattern prediction model according to the feedback information.

[0159] In a possible implementation of the embodiment of the present disclosure, the acquisition module 604 is used to: collect transaction data generated by the first transaction request processing process; based on the transaction data, use a transaction risk identification model to perform a first risk identification on the first transaction to obtain a second risk identification result of the first transaction; and generate a first risk identification result based on the second risk identification result.

[0160] In a possible implementation of the embodiment of the present disclosure, the acquisition module 604 is also used to: in response to the second risk identification result indicating that the risk cannot be determined, generate a risk identification request based on the transaction data; send the risk identification request to the target terminal, so that the target terminal performs a second risk identification on the first transaction based on the transaction data, and obtains a third risk identification result of the first transaction; and generate a first risk identification result based on the third risk identification result.

[0161] In a possible implementation of the embodiment of the present disclosure, the acquisition module 604 is used to: collect a second video in the first face recognition verification mode; perform a second risk analysis on the second video to obtain a second risk index value of the second video; perform a first risk analysis on the login account associated with the first transaction request to determine the account risk index; determine the first face recognition verification result of the first face recognition verification mode based on the second risk index value and / or the account risk index.

[0162] In a possible implementation of the embodiment of the present disclosure, the acquisition module 604 is also used to: extract features of multiple dimensions from the second video; input the features of any dimension into the corresponding prediction branch in the image risk identification model to obtain the sub-risk indicator value corresponding to the dimension; and determine the second risk indicator value of the second video based on the sub-risk indicator values ​​corresponding to each dimension.

[0163] In a possible implementation of the embodiment of the present disclosure, the acquisition module 604 is also used to: determine whether face recognition verification is needed again based on the second risk indicator value and / or the account risk index; when it is determined that face recognition verification is needed again, perform face recognition verification again using the second face recognition verification mode to obtain a second face recognition verification result of the second face recognition verification mode; and generate the first face recognition verification result of the first face recognition verification mode based on the second face recognition verification result of the second face recognition verification mode.

[0164] In a possible implementation of the embodiment of the present disclosure, the training module 605 is used to: determine a corresponding reward coefficient according to feedback information; and fine-tune the pattern prediction model based on the reward coefficient.

[0165] In a possible implementation of the embodiment of the present disclosure, the training module 605 is used to: in response to the feedback information indicating that the verification is accurate, determine the reward coefficient as an upward setting coefficient; in response to the feedback information indicating that the risky first transaction is identified as a normal transaction, determine the reward coefficient as a first downward setting coefficient; in response to the feedback information indicating that the normal first transaction is identified as a risky transaction, determine the reward coefficient as a second downward setting coefficient; in response to the feedback information indicating that the static face recognition verification mode is used additionally, determine the reward coefficient as a third downward setting coefficient; in response to the feedback information indicating that the dynamic face recognition verification mode is used additionally, determine the reward coefficient as a fourth downward setting coefficient.

[0166] In a possible implementation of the embodiment of the present disclosure, the analysis module 602 is used to: perform emotion recognition on the facial information of the user contained in the image foreground of the first video to obtain the user's emotional state; perform head recognition on the first video to obtain the number of heads; analyze the environmental information contained in the image background of the first video to obtain the environmental category of the image background; and generate a first risk indicator value for the first video based on the user's emotional state, the number of heads and the environmental category.

[0167] The model training device of the disclosed embodiment, in response to the target transaction request operation initiated by the user, calls the camera to collect the first video including the user's face; performs a first risk analysis on the facial information contained in the foreground of the image in the first video and the environmental information contained in the background of the image in the first video, and determines the first risk index value of the first video; determines whether the first video has passed the verification according to the first risk index value of the first video; if the first video has passed the verification, obtains feedback information of the first face recognition verification mode adopted for the first transaction request before the first video is collected; wherein the feedback information is determined based on the first face recognition verification result of the first face recognition verification mode and the first risk identification result of the first transaction, and the first transaction request is used to process the first transaction; and trains the pattern prediction model according to the feedback information. Thus, the pattern prediction model can be automatically trained based on the feedback information, so that the trained pattern prediction model can learn and generalize the face recognition verification mode that adapts to the actual situation and risk level of the user.

[0168] With the above Figure 5 Corresponding to the request processing method provided in the embodiment, the present disclosure also provides a request processing device. Figure 5 The request processing method provided in the embodiment corresponds to the request processing method, so the implementation of the request processing method is also applicable to the request processing device provided in the embodiment of the present disclosure, and will not be described in detail in the embodiment of the present disclosure.

[0169] Figure 7 It is a structural diagram of a request processing device shown in the seventh embodiment of the present disclosure.

[0170] Reference Figure 7 The request processing device 700 may include: a calling module 701, an obtaining module 702, a determining module 703, a verifying module 704 and a processing module 705.

[0171] The calling module 701 is used to call a camera to capture a first video including the user's face in response to a target transaction request operation initiated by the user.

[0172] The acquisition module 702 is used to acquire the first face recognition verification mode used for the first transaction request before the first video is captured, if the first video verification passes.

[0173] The determination module 703 is used to determine a second face recognition verification mode for performing face recognition verification on the user based on the first face recognition verification mode by using a pattern prediction model.

[0174] The verification module 704 is used to perform face recognition verification on the user using the second face recognition verification mode to obtain a target face recognition verification result.

[0175] The processing module 705 is used to execute the transaction processing corresponding to the target transaction request operation according to the target face recognition verification result.

[0176] The request processing device of the disclosed embodiment, in response to the target transaction request operation initiated by the user, calls the camera to collect the first video including the user's face; when the first video verification passes, obtains the first face recognition verification mode used for the first transaction request before the first video is collected; based on the first face recognition verification mode, adopts the pattern prediction model to determine the second face recognition verification mode for face recognition verification of the user; adopts the second face recognition verification mode to perform face recognition verification on the user to obtain the target face recognition verification result; according to the target face recognition verification result, executes the transaction processing corresponding to the target transaction request operation. Thus, the face recognition verification mode suitable for the user can be intelligently determined, effectively avoiding the interference and influence of the face recognition process caused by the pre-made video, and due to the real-time and timeliness of the face recognition verification mode generation, it can effectively avoid the situation of being cracked by force; and the face recognition verification mode used in high-risk scenarios and low-risk scenarios is different, in the low-risk scenario, a simpler face recognition verification mode can be used, and in the high-risk scenario, a more complex face recognition verification mode can be used, on the one hand, the accuracy of face recognition verification is improved, and on the other hand, the waste of computing resources can be effectively avoided.

[0177] In an exemplary embodiment, an electronic device is also provided.

[0178] Among them, electronic equipment includes:

[0179] processor;

[0180] a memory for storing processor-executable instructions;

[0181] Among them, the processor is configured to execute instructions to implement the model training method proposed in the aforementioned first aspect embodiment, or the request processing method proposed in the second aspect embodiment.

[0182] As an example, Figure 8 is a schematic diagram of the structure of an electronic device 800 shown in an exemplary embodiment of the present disclosure, such as Figure 8 As shown, the electronic device 800 may further include:

[0183] The memory 810 and the processor 820, a bus 830 connecting different components (including the memory 810 and the processor 820), the memory 810 stores a computer program, and when the processor 820 executes the program, the model training method described in the embodiment of the first aspect of the present disclosure, or the request processing method proposed in the embodiment of the second aspect is implemented.

[0184] Bus 830 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnect (PCI) bus.

[0185] The electronic device 800 typically includes a variety of electronic device readable media. These media can be any available media that can be accessed by the electronic device 800, including volatile and non-volatile media, removable and non-removable media.

[0186] The memory 810 may also include computer system readable media in the form of volatile memory, such as random access memory (RAM) 840 and / or cache memory 850. The server 800 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 860 may be used to read and write non-removable, non-volatile magnetic media ( Figure 8 not shown, usually called a "hard drive"). Although Figure 8 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 830 via one or more data medium interfaces. The memory 810 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the various embodiments of the present disclosure.

[0187] A program / utility 880 having a set (at least one) of program modules 870 may be stored, for example, in the memory 810, such program modules 870 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 870 generally perform the functions and / or methods of the embodiments described in the present disclosure.

[0188] The electronic device 800 may also communicate with one or more external devices 890 (e.g., keyboard, pointing device, display 891, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 800, and / or may communicate with any device that enables the electronic device 800 to communicate with one or more other computing devices (e.g., network card, modem, etc.). Such communication may be performed via an input / output (I / O) interface 892. Furthermore, the electronic device 800 may also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN) and / or public network, such as the Internet) via a network adapter 893. As shown, the network adapter 893 communicates with other modules of the electronic device 800 via a bus 830. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0189] The processor 820 executes various functional applications and data processing by running the programs stored in the memory 810 .

[0190] It should be noted that the implementation process and technical principles of the electronic device of this embodiment refer to the aforementioned explanation of the model training method of the first aspect embodiment of the present disclosure or the request processing method proposed in the second aspect embodiment, and will not be repeated here.

[0191] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory including instructions, and the above instructions can be executed by a processor of an electronic device to complete the method proposed in any of the above embodiments. Optionally, the computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0192] In an exemplary embodiment, a computer program product is also provided, including a computer program / instruction, wherein the computer program / instruction implements the method proposed in any of the above embodiments when executed by a processor.

[0193] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0194] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A model training method, characterized in that: The method comprises: In response to a target transaction request operation initiated by a user, calling a camera to capture a first video including a face of the user; Performing a first risk analysis on facial information contained in the foreground of the image in the first video and environmental information contained in the background of the image in the first video to determine a first risk index value of the first video; Determining whether the first video passes verification according to a first risk indicator value of the first video; When the first video verification passes, obtaining feedback information of a first face recognition verification mode adopted by a first transaction request before the first video is collected; wherein the feedback information is determined based on a first face recognition verification result of the first face recognition verification mode and a first risk identification result of the first transaction, and the first transaction request is used to process the first transaction; The pattern prediction model is trained according to the feedback information.

2. The method according to claim 1, characterized in that The obtaining of the first risk identification result of the first transaction includes: collecting transaction data generated during the first transaction request processing; Based on the transaction data, using a transaction risk identification model, performing a first risk identification on the first transaction to obtain a second risk identification result of the first transaction; The first risk identification result is generated according to the second risk identification result.

3. The method according to claim 2, characterized in that The method further comprises: In response to the second risk identification result indicating that the risk cannot be determined, generating a risk identification request based on the transaction data; sending the risk identification request to a target terminal, so that the target terminal performs a second risk identification on the first transaction based on the transaction data, and obtains a third risk identification result of the first transaction; The first risk identification result is generated according to the third risk identification result.

4. The method according to claim 1, characterized in that: The obtaining of the first face recognition verification result includes: Capturing a second video in the first face recognition verification mode; Performing a second risk analysis on the second video to obtain a second risk index value of the second video; Performing a first risk analysis on the login account associated with the first transaction request to determine an account risk index; Determine the first face recognition verification result of the first face recognition verification mode according to the second risk indicator value and / or the account risk index.

5. The method according to claim 4, characterized in that The performing a second risk analysis on the second video to obtain a second risk index value of the second video includes: Extracting features of multiple dimensions from the second video; Inputting the feature of any of the dimensions into the corresponding prediction branch of the image risk identification model to obtain the sub-risk index value corresponding to the dimension; A second risk index value of the second video is determined according to the sub-risk index values ​​corresponding to each of the dimensions.

6. The method according to claim 4, characterized in that The determining, according to the second risk indicator value and / or the account risk index, the first face recognition verification result of the first face recognition verification mode includes: Based on the second risk indicator value and / or the account risk index, determining whether face recognition verification is required again; If it is determined that face recognition verification needs to be performed again, face recognition verification is performed again using a second face recognition verification mode to obtain a second face recognition verification result of the second face recognition verification mode; A first face recognition verification result of the first face recognition verification mode is generated according to the second face recognition verification result of the second face recognition verification mode.

7. The method according to claim 1, characterized in that The training of the pattern prediction model according to the feedback information includes: Determine a corresponding reward coefficient according to the feedback information; Based on the reward coefficient, the pattern prediction model is fine-tuned.

8. The method according to claim 7, characterized in that Determining a corresponding reward coefficient according to the feedback information includes: In response to the feedback information indicating that the verification is accurate, determining that the reward coefficient is an upwardly adjusted set coefficient; In response to the feedback information indicating that the risky first transaction is identified as a normal transaction, determining the reward coefficient to be a first downward setting coefficient; In response to the feedback information indicating that the normal first transaction is identified as a risky transaction, determining the reward coefficient to be a second downward setting coefficient; In response to the feedback information indicating that the static face recognition verification mode is used once more, determining the reward coefficient to be a third downward setting coefficient; In response to the feedback information indicating an additional use of the dynamic face recognition verification mode, the reward coefficient is determined to be a fourth downward setting coefficient.

9. The method according to any one of claims 1 to 8, characterized in that: The performing a first risk analysis on the facial information contained in the foreground of the image in the first video and the environmental information contained in the background of the image in the first video to determine a first risk index value of the first video includes: Performing emotion recognition on the facial information of the user contained in the foreground of the image in the first video to obtain the emotional state of the user; Performing head recognition on the first video to obtain the number of heads; Analyze the environmental information contained in the image background of the first video to obtain the environmental category of the image background; Based on the user's emotional state, the number of people and the environment category, a first risk index value of the first video is generated.

10. A request processing method, characterized in that: The method comprises: In response to a target transaction request operation initiated by a user, calling a camera to capture a first video including a face of the user; If the first video verification passes, obtaining a first face recognition verification mode used for the first transaction request before the first video is captured; Based on the first face recognition verification mode, using a pattern prediction model, determining a second face recognition verification mode for performing face recognition verification on the user; Performing face recognition verification on the user using the second face recognition verification mode to obtain a target face recognition verification result; According to the target face recognition verification result, the transaction processing corresponding to the target transaction request operation is executed.

11. A model training device, characterized in that: The device comprises: A calling module, configured to call a camera to capture a first video including a user's face in response to a target transaction request operation initiated by the user; An analysis module, configured to perform a first risk analysis on facial information contained in the foreground of the image in the first video and environmental information contained in the background of the image in the first video to determine a first risk index value of the first video; A determination module, configured to determine whether the first video has passed verification according to a first risk indicator value of the first video; an acquisition module, configured to acquire feedback information of a first face recognition verification mode adopted by a first transaction request before the first video is collected, if the first video verification passes; wherein the feedback information is determined based on a first face recognition verification result of the first face recognition verification mode and a first risk identification result of the first transaction, and the first transaction request is used to process the first transaction; The training module is used to train the pattern prediction model according to the feedback information.

12. A request processing device, characterized in that: The device comprises: A calling module, configured to call a camera to capture a first video including a user's face in response to a target transaction request operation initiated by the user; an acquisition module, configured to acquire, when the first video verification passes, a first face recognition verification mode used for the first transaction request before the first video is collected; A determination module, configured to determine, based on the first face recognition verification mode, a second face recognition verification mode for performing face recognition verification on the user using a pattern prediction model; A verification module, configured to perform face recognition verification on the user using the second face recognition verification mode to obtain a target face recognition verification result; A processing module is used to execute transaction processing corresponding to the target transaction request operation according to the target face recognition verification result.

13. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 9 or the method according to claim 10.

14. A computer readable medium, characterized in that The computer-readable medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 9 or the method according to claim 10 when executed by a processor.

15. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 9 or the method according to claim 10 when being executed by a processor.

Images