Template feature updating method and device and storage medium

Abstract

The invention provides a template feature updating method and device and a storage medium. In the method, verification features used for describing the concentrated trend of m to-be-identified features matched with template features in a historical time period can be obtained, and the verification features and the template features are subjected to weighted fusion to obtain updated template features. In the template feature updating method, the updated template feature is related to the verification feature and the template feature before updating, and the verification feature describes the centralized trend of the m to-be-identified features matched with the template feature in the historical time period, so that the difference between the updated template feature and the real feature is smaller, and the template feature updating accuracy is improved. FRR and / or FAR of biological recognition can be effectively reduced, so that the accuracy of biological recognition can be improved.

Description

Technical Field

[0001] The present application relates to the field of terminal technology, and in particular to a template feature updating method, device and storage medium. Background Art

[0002] Biometrics refers to a series of related technologies that use human physiological characteristics (e.g., facial images, irises, etc.) or behavioral characteristics (e.g., voice, gait, etc.) to verify identity. For example, taking face recognition as an example, face recognition can be performed on the detected face image to determine the user's identity.

[0003] In the related art, it is usually necessary to collect the user's template features in advance (such as the features of the face image during the face recognition process). When the user needs to be biometrically identified, the user's features to be identified are matched with the template features. When the features to be identified match the template features, identity events are triggered, such as unlocking, payment, opening access rights, etc. As time goes by, the user's real features change, and the template features may differ greatly from the real features, resulting in biometric errors. Therefore, the template features need to be updated. The existing solution is to directly replace the template features with the features to be identified, which is prone to high false rejection rate (FRR) and / or false acceptance rate (FAR) of biometrics. Summary of the invention

[0004] The embodiments of the present application provide a template feature updating method, device and storage medium. The template feature updating method can effectively reduce the FRR and / or FAR of biometrics, thereby facilitating the improvement of the accuracy of biometrics.

[0005] In the first aspect, an embodiment of the present application provides a template feature updating method, the method comprising: obtaining m features to be identified that match the template features of an object within a historical time period; wherein the template features are reference features of the object in a biometric recognition process; m is a positive integer greater than 1; based on an averaging algorithm, obtaining a verification feature for describing a central trend of the m features to be identified within a historical time period; performing weighted fusion of the verification feature and the template feature to obtain the updated template feature of the object.

[0006] It can be seen that in the method described in this aspect, a verification feature for describing the central tendency of the m features to be identified that match the template feature in the historical time period can be obtained, and the verification feature is weightedly fused with the template feature to obtain the updated template feature. Compared with the solution of replacing the template feature with the feature to be identified, in the method described in this aspect, the updated template feature is related to the verification feature and the template feature before the update, and the updated template feature changes with the change of the verification feature and / or the template feature before the update, so that the updated template feature is conducive to the timely change of the real feature, and the difference between the updated template feature and the real feature is smaller, which can effectively reduce the FRR and / or FAR of biometrics and improve the accuracy of biometrics. At the same time, the verification feature describes the central tendency of the m features to be identified that match the template feature in the historical time period, and the difference between the verification feature and the real feature is smaller. When the template feature is updated using the verification feature, the update direction is more accurate, so that the difference between the updated template feature and the real feature is smaller.

[0007] In combination with the first aspect, in an optional implementation, based on an average algorithm, a verification feature is obtained for describing the central trend of m features to be identified in a historical time period, including: determining the exponential decay weight of each feature to be identified based on each feature to be identified and the template feature; based on the exponential decay weight of each feature to be identified, weighting the m features to be identified to obtain the verification feature. Exemplarily, the verification feature can be calculated by formula (2) of the embodiment of the present application, which will not be described in detail.

[0008] It can be seen that in this embodiment, the mean feature obtained by exponential weighting can be used to represent the central trend of m features to be identified in the historical time period, so that the weights of the first feature to be identified to the mth feature to be identified in the verification feature increase successively, that is, the features to be identified that are closer to the current moment have a larger weight in the verification feature, and the features to be identified that are farther away from the current moment have a smaller weight in the verification feature, which is conducive to further narrowing the difference between the verification feature and the real feature.

[0009] In combination with the first aspect, in an optional embodiment, based on each feature to be identified and the template feature, the exponential decay weight of each feature to be identified is determined, including: based on each feature to be identified and the template feature, determining the corresponding similarities between each feature to be identified and the template feature; based on the corresponding similarities, determining the exponential decay weight of each feature to be identified.

[0010] It can be seen that in this embodiment, the exponential decay weight of the feature to be identified is positively correlated with the similarity. For example, the higher the similarity, the larger the exponential decay weight. When the feature to be identified is weighted by the exponential decay weight obtained in this way to obtain the verification feature, the verification feature is more accurate in the rate or trend of change of the real feature, which is conducive to the updated template feature to approach the real feature more quickly.

[0011] In combination with the first aspect, in an optional embodiment, based on each feature to be identified and the template feature, the similarity between each feature to be identified and the template feature is determined, including: based on the similarity between the mth feature to be identified and the mth template feature, determining the exponential decay weight of the mth feature to be identified.

[0012] Optionally, the number of template features that match the m features to be identified in the historical time period is m, and each of the m features to be identified matches the template features in the same order among the m template features. Exemplarily, taking m=3 as an example, for example, the first feature to be identified matches the first template feature, the second feature to be identified matches the second template feature, the third feature to be identified matches the third template feature, the first template feature is extracted based on template biological data, the second template feature is obtained by updating the first template feature, and the third template feature is obtained by updating the second template feature.

[0013] Optionally, taking the mth template feature update process as an example, based on each feature to be identified and the template feature, the similarity between each feature to be identified and the template feature is determined, including: sequentially obtaining the exponential decay weights of the first feature to be identified to the m-1th feature to be identified from the storage area, and based on the similarity between the mth feature to be identified and the mth template feature, determining the exponential decay weight of the mth feature to be identified. It can be seen that the electronic device stores the exponential decay weights obtained by historical calculations, which can avoid repeated calculations, save computing resources, and is simpler and more convenient.

[0014] In combination with the first aspect, in an optional embodiment, based on the similarity between the mth feature to be identified and the template feature, the exponential decay weight of the mth feature to be identified is determined, including: obtaining the quality score of the mth biological data to be identified; the mth biological data to be identified is used to provide the mth feature to be identified; based on the quality score and the similarity between the mth feature to be identified and the template feature, the exponential decay weight of the mth feature to be identified is determined.

[0015] It can be seen that in this embodiment, the exponential decay weight of the feature to be identified is positively correlated with the similarity and the quality score. For example, the higher the similarity and the higher the quality score, the greater the exponential decay weight. When the exponential decay weight obtained in this way is used to weight the feature to be identified to obtain the verification feature, the verification feature is more accurate in the rate or trend of change of the real feature, which is conducive to the updated template feature to approach the real feature more quickly.

[0016] In combination with the first aspect, in an optional embodiment, the exponential decay weight of the mth feature to be identified is determined based on the quality score and the similarity between the mth feature to be identified and the mth template feature, including: substituting the quality score and the similarity between the mth feature to be identified and the mth template feature into the two-dimensional Gaussian distribution function to obtain the exponential decay weight of the mth feature to be identified.

[0017] In combination with the first aspect, in an optional implementation, the verification feature and the template feature are weightedly fused to obtain the updated template feature of the object, including: obtaining a first preset weight of the verification feature and a second preset weight of the template feature; based on the first preset weight of the verification feature and the second preset weight of the template feature, the verification feature and the template feature are weightedly fused to obtain the updated template feature of the object. Exemplarily, the updated template feature can be calculated by formula (5) of the embodiment of the present application, which will not be described in detail.

[0018] In combination with the first aspect, in an optional implementation, the template feature updating method is executed by a hardware abstraction layer of the electronic device. Optionally, the hardware abstraction layer may include a weight calculation module and / or a template feature updating module.

[0019] In a second aspect, an embodiment of the present application further provides an electronic device, comprising: one or more processors and one or more memories; the one or more memories are coupled to the one or more processors, the one or more memories are used to store computer program codes, the computer program codes include computer instructions, and when the one or more processors execute the computer instructions, the electronic device executes the method of the first aspect.

[0020] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, including computer instructions, which, when executed on an electronic device, enables the electronic device to execute the method of the first aspect.

[0021] In a fourth aspect, an embodiment of the present application further provides a chip or chip system, including a processing circuit and an interface circuit, wherein the interface circuit is used to receive code instructions and transmit them to the processing circuit, and the processing circuit is used to run the code instructions to execute the method of the first aspect.

[0022] In a fifth aspect, an embodiment of the present application further provides a computer program product, which, when executed on an electronic device, enables the electronic device to execute the method of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments of the present application will be described below.

[0024] Figure 1 This is a schematic diagram of a face recognition scenario provided by an embodiment of the present application;

[0025] Figure 2 is a schematic diagram of another face recognition scenario provided by an embodiment of the present application;

[0026] Figure 3 It is a flow chart of a template feature updating method provided in an embodiment of the present application;

[0027] Figure 4 It is a flowchart of another template feature updating method provided in an embodiment of the present application;

[0028] Figure 5 It is a flowchart of another template feature updating method provided in an embodiment of the present application;

[0029] Figure 6 It is a schematic diagram of an interface for recording a facial image of an electronic device provided in an embodiment of the present application;

[0030] Figure 7 It is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application;

[0031] Figure 8 It is a schematic diagram of the software structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0032] Some terms in this application are explained below to facilitate understanding by those skilled in the art.

[0033] The application (Application, App) involved in the embodiments of the present application, which can also be called an application program, is a software program that can realize one or more specific functions. Usually, multiple applications can be installed in an electronic device, for example, instant messaging applications, audio applications, image capture applications, and the like. Among them, instant messaging applications, for example, may include text messaging applications, WeChat (WeChat), DingTalk, and the like. Audio applications, for example, may include Kugou Music, Xiami Music, and the like. Image capture applications, for example, may include camera applications (factory setting system camera or third-party camera application). The applications mentioned in the following embodiments may be applications that are installed on the electronic device when it leaves the factory, or they may be applications that the user downloads from the Internet or obtains from other electronic devices during the use of the electronic device.

[0034] It should be understood that, in this application, unless otherwise specified, " / " means or, for example, A / B can mean A or B; "and / or" in this application is only a way to describe the association relationship of associated objects, indicating that there can be three relationships, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. "At least one" means one or more, and "more than one" means two or more.

[0035] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0036] It should be noted that the "successful match" mentioned in this application can be simply referred to as "match", and the "failed match" can be simply referred to as "mismatch", without limitation.

[0037] It should be pointed out that the "equal to" involved in the embodiments of the present application can be used in conjunction with "greater than", and is applicable to the technical solution adopted when "greater than" is used, and can also be used in conjunction with "less than", and is applicable to the technical solution adopted when "less than" is used. It should be noted that when "equal to" is used in conjunction with "greater than", it cannot be used in conjunction with "less than"; when "equal to" is used in conjunction with "less than", it cannot be used in conjunction with "greater than".

[0038] Optionally, the difference between the features involved in the embodiments of the present application can be measured by the similarity (or matching score) between the features. For example, when the similarity between the features is greater than or equal to a preset value, it is determined that the difference between the features is small, and when the similarity between the features is less than the preset value, it is determined that the difference between the features is large.

[0039] With the rapid development of terminal technology, electronic devices with cameras and microphones have been widely used. Based on the image capture function of the camera and the voice collection function of the microphone, electronic devices have realized more and more functions, such as paying living bills, logging into bank accounts, checking personal social security information, and checking personal tax information. Among them, many functions involve private information such as user identity and property, so it is necessary to verify the user's identity based on biometric technology. The biometric recognition can include but is not limited to one or more of face recognition, iris recognition, voiceprint recognition, gait recognition, or fingerprint recognition.

[0040] Exemplarily, taking face recognition as an example, face recognition usually involves image acquisition and processing. When a user initiates unlocking, payment, login authority verification, etc. through an electronic device, the electronic device turns on the camera function, obtains the user's face image to be recognized through the camera, and extracts feature information of multiple feature points from the face image to be recognized to obtain the features to be recognized, and matches them with the pre-stored template features. When matching the user's features to be recognized with the template features, if the similarity (or matching score) between the features to be recognized and the template features is greater than or equal to the first threshold, it is determined that the current features to be recognized match the template features, so that the electronic device successfully triggers the unlocking function to open the desktop, triggers the payment function to complete the transaction, or triggers the login function to open the login authority.

[0041] For example, Figure 1 FIG. 1 is a schematic diagram of a face recognition scenario provided by an embodiment of the present application. Figure 1 As shown in a in FIG, when a user selects a target product in a shopping application, the user can confirm the product order information and submit the order on the purchase interface of the target product. After the order is submitted, an order payment page pops up on the order submission page of the shopping application, as shown in FIG. Figure 1 When the order payment page is presented, the camera function of the electronic device is turned on to obtain the face image of the user to be identified. According to the face image to be identified, the current facial features are extracted as the features to be identified, such as Figure 1 As shown in c. After the features to be identified are extracted, the extracted features to be identified are matched with the template features stored on the server side corresponding to the user's shopping application. If the match is successful, the amount of the user's account is deducted according to the amount corresponding to the order, and the amount of the user's account is transferred to the corresponding payment account to complete the transaction. If the features to be identified fail to match the template features, a re-collection operation button may be displayed, and a prompt message of failed matching may be generated. The user can click the operation button to re-acquire the face image to be identified.

[0042] Another example, Figure 2 FIG. 1 is another schematic diagram of a face recognition scenario provided by an embodiment of the present application. Figure 2 As shown in a in FIG, when a user logs in to an instant messaging application, the electronic device can present a login page. When presenting the login page, the camera function of the electronic device is turned on to obtain the user's face image to be identified, and the current facial features are extracted as the features to be identified based on the face image to be identified. After extracting the features to be identified, the extracted features to be identified are matched with the template features stored on the server side corresponding to the instant messaging application of the user. If the match is successful, the login is confirmed to be successful, and the user is redirected from the login page to the application interface of the instant messaging application, such as Figure 2 If the feature to be identified fails to match the template feature, a re-collection operation button may be displayed, and a prompt message of matching failure may be generated, and the user may click the operation button to re-acquire the face image to be identified.

[0043] It should be noted that the above-mentioned server side has a database, or the server side is connected to a specific database, and the required data (such as template features) can be obtained from the database. In other embodiments, the data can also be stored in the local end of the electronic device, without limitation.

[0044] It should be noted that the above-mentioned possible application scenarios are only exemplary. The application may also include other similar application scenarios, such as unlocking, sign-in, iris recognition, fingerprint recognition and other application scenarios. In these similar scenarios, biometrics can also be used, which will not be elaborated.

[0045] Optionally, in different biometric recognition scenarios, the features to be recognized and the template features are used to represent the features of different types of biometric data to be recognized. For example, taking the face recognition scenario as an example, the features to be recognized and the template features are used to represent the features of the face image; for another example, taking the voiceprint recognition scenario as an example, the features to be recognized and the template features are used to represent the features of the voice information.

[0046] It should be noted that the "features to be identified" mentioned in this application may refer to "feature vectors to be identified", "feature matrices to be identified", etc., and "template features" may refer to "template feature vectors", "template feature matrices", etc. Optionally, the dimension of the feature to be identified is the same as the dimension of the template feature. For example, the dimension of the verification feature is i*j, and the dimension of the template feature is also i*j, where i and j are positive integers. It should be noted that when the dimension of the feature to be identified is different from the dimension of the template feature, the dimension of the feature to be identified and / or the dimension of the template feature can be adjusted based on the conversion technology, which will not be elaborated in detail.

[0047] As time goes by, due to the change of the real features of the user (or object), the template features may be greatly different from the real features, which can easily misidentify a successful match as a failed match, and / or misidentify a failed match as a successful match, thus causing biometric errors. Therefore, in order to avoid this phenomenon, the template features need to be updated.

[0048] In order to solve the problem of updating template features, a method is proposed in the prior art. Specifically, when the to-be-recognized feature of the object matches the template feature successfully, the to-be-recognized feature is used to replace the template feature before the update as the updated template feature. The method proposed in the prior art can dynamically update the template feature. However, this updating method easily leads to a large difference between the updated template feature and the real feature. For example, when the edge feature that is greatly different from the real feature is used as the updated template feature, or the real feature changes greatly, it is easy to cause a large difference between the updated template feature and the real feature, so that the successful match is easily misidentified as a failed match, resulting in a higher FRR of biometric recognition, and / or, the failed match is misidentified as a successful match, resulting in a higher FAR of biometric recognition.

[0049] Based on this, the embodiment of the present application provides a template feature update method, which can obtain a verification feature for describing the central trend of m features to be identified that match the template feature in a historical time period, and weightedly fuse the verification feature with the template feature to obtain an updated template feature. In the template feature update method, the difference between the updated template feature and the real feature is smaller, which can effectively reduce the FRR and / or FAR of biometrics, thereby helping to improve the accuracy of biometrics.

[0050] See also Figure 3 , Figure 3 Schematic diagram of a template feature updating method provided in an embodiment of the present application. Figure 3 As shown, the method may include but is not limited to steps S101-S103:

[0051] S101. Obtain m features to be identified that match the template features of the object within a historical time period.

[0052] Wherein, m is a positive integer. Optionally, m is equal to 1, or m is a positive integer greater than 1. The following exemplary description is given with m being a positive integer greater than 1.

[0053] Optionally, the historical time period may refer to the time period between the moment when the user actively enters the template biometric data (such as the template face image) and the current moment. For example, taking the moment when the user actively enters the template biometric data as 2024 / 1 / 10 / 9:00 and the current moment as 2024 / 1 / 11 / 9:30 as an example, the historical time period is 2024 / 1 / 10 / 9:00-2024 / 1 / 11 / 9:30.

[0054] Optionally, the m features to be identified that match the template features may refer to m features to be identified whose similarity with the template features is greater than or equal to a first threshold. The first threshold may be set based on experience or business needs, and the specific value is not limited.

[0055] Among them, the template feature is a reference feature of the object in the biometric recognition process. In an optional embodiment, the number of template features of the object in the historical time period is 1, for example, the template feature is the one template feature extracted based on the template biometric data (such as the template face image actively entered by the user). In another optional embodiment, the number of template features of the object in the historical time period is m, and each feature to be identified among the m features to be identified matches the template features in the same order among the m template features. Then, in this case, the first template feature can be extracted based on the template biometric data, and the second to m-th template features can be respectively updated from the first to m-1th template features. Exemplarily, taking m=3 as an example, for example, the first feature to be identified matches the first template feature, the second feature to be identified matches the second template feature, the third feature to be identified matches the third template feature, the first template feature is extracted based on the template biometric data, the second template feature is updated from the first template feature, and the third template feature is updated from the second template feature.

[0056] S102: Based on an average algorithm, a verification feature is obtained for describing a central trend of the m features to be identified in a historical time period.

[0057] The averaging algorithm may include a moving average method and an exponentially weighted average method.

[0058] For the moving average method, step S102 includes: taking the ratio of the sum of the m features to be identified to m as the verification feature. For example, the verification feature can be calculated by formula (1):

[0059]

[0060] in, Used to represent the qth feature to be identified; V weighting Used to represent validation characteristics.

[0061] It can be seen that in this implementation, the mean feature can be used to represent the central trend of m features to be identified in the historical time period, so that the difference between the verification feature and the real feature is small, while improving the convenience of data processing and providing processing efficiency.

[0062] For the exponential weighted average method, step S102 includes: determining the exponential decay weight of each feature to be identified based on each feature to be identified and the template feature, and weighting the m features to be identified based on the exponential decay weight of each feature to be identified to obtain the verification feature. Exemplarily, the verification feature can be calculated by formula (2):

[0063]

[0064] Among them, V weighting Used to represent verification features; Used to represent the mth feature to be identified, Used to represent the m-1th feature to be identified. Used to represent the m-2th feature to be identified, w m Used to represent the exponential decay weight of the mth feature to be identified, w m-1 Used to represent the exponential decay weight of the m-1th feature to be identified, w m-2 Used to represent the exponential decay weight of the m-2th feature to be identified.

[0065] It can be seen that in this implementation method, the mean feature obtained by exponential weighting can be used to represent the central trend of m features to be identified in the historical time period, so that the weights of the first feature to be identified to the mth feature to be identified in the verification feature increase successively, that is, the features to be identified that are closer to the current moment have a larger weight in the verification feature, and the features to be identified that are farther away from the current moment have a smaller weight in the verification feature, which is conducive to further narrowing the difference between the verification feature and the real feature.

[0066] Optionally, when the value of m is small (for example, when the historical time period is short, or when there are fewer features to be identified that match the template feature in the historical time period, the value of m is small), the difference between the verification feature obtained by the exponential weighted average method and the real feature is large. In order to reduce the difference between the verification feature and the real feature, the verification feature can be corrected for deviation. Specifically, the ratio of the verification feature before deviation correction to the difference between the reference value and the weight cumulative value is used as the verification feature after deviation correction. It should be noted that the embodiment of the present application is exemplified with a reference value of 1, without limitation.

[0067] Exemplarily, the deviation correction can be performed by the following formula (3):

[0068]

[0069] Among them, V weighting Used to represent the verification characteristics before bias correction, w cumulation Used to represent the cumulative value of weight, V corrected Used to represent the bias-corrected validation features.

[0070] Optionally, the cumulative value of the weight mentioned above may refer to the cumulative value of m exponentially decaying weights.

[0071] For example, the cumulative value of the weight can be calculated by formula (4):

[0072] w cumulation =w1×w2×w3...w m-2 × m-1 × m (4)

[0073] S103: Perform weighted fusion on the verification feature and the template feature to obtain the updated template feature of the object.

[0074] In an optional implementation, step S103 includes: obtaining a first preset weight of the verification feature and a second preset weight of the template feature, and based on the first preset weight of the verification feature and the second preset weight of the template feature, performing weighted fusion on the verification feature and the template feature to obtain the updated template feature of the object.

[0075] Exemplarily, the weighted fusion step can be described by formula (5):

[0076] V m+1 =aV weighting +bV m (5)

[0077] Among them, V weighting is used to represent the verification feature; a is used to represent the first preset weight; V m is used to represent the template feature (i.e., the template feature before updating), b is used to represent the second preset weight, V m+1 It is used to represent the updated template features. It should be noted that the first preset weight a and the second preset weight b can be set based on experience or business needs, and the specific values ​​are not limited. For example, the value of a can be 0.1, and the value of b can be 1. For another example, the value of a can be 0.1, and the value of b can be 0.9, and so on.

[0078] Optionally, when the number of template features in the historical time period is m, the template feature before updating is the mth template feature in the historical time period, and the template feature after updating can be used as the m+1th template feature. It should be noted that in other implementations, the verification feature V in formula (5) weighting It can also be replaced by the modified verification feature V corrected , without limitation.

[0079] In some embodiments, the method may further include a data normalization step, wherein the value range of each element of the feature after data normalization is in [0,1]. Exemplarily, the data normalization step may be expressed as formula (6):

[0080]

[0081] Among them, V s Used to indicate features that require data normalization; for example, when a data normalization step is required for a validation feature, this V s Can be replaced by V weighting For example, when it is necessary to perform a data standardization step on the modified verification feature, the V s Can be replaced by V corrected For another example, when it is necessary to perform a data standardization step on the updated template features, the V s Can be replaced by V m+1 , etc. Through the data standardization step, the data quality can be improved, the errors and inconsistencies in the data can be reduced, and the accuracy, reliability and completeness of the data can be improved.

[0082] Furthermore, the implementation method of calculating the verification feature by the exponential weighted average method is used as an example to illustrate the embodiment of the present application. Figure 4 As shown, it is a flow chart of another template feature updating method provided in an embodiment of the present application. Figure 4 As shown, the method may include but is not limited to steps s11-s16:

[0083] s11: Get m features to be identified that match the template features of the object within the historical time period.

[0084] s12: Based on each feature to be identified and the template feature, determine the exponential decay weight of each feature to be identified, and based on the exponential decay weight of each feature to be identified, weight the m features to be identified to obtain the verification feature.

[0085] s13: Based on the accumulated value of the weight, the verification feature is bias-corrected to obtain the bias-corrected verification feature.

[0086] s14: Perform data standardization on the deviation-corrected validation features to obtain data-standardized validation features.

[0087] s15: The verification features after data standardization are merged with the template features to obtain the template features after the object is updated.

[0088] s16: Perform data normalization on the updated template features to obtain data normalized template features.

[0089] It should be noted that the detailed description of steps s11 to s16 can be found in steps S101 to S103, which will not be described in detail. It should be noted that some steps in steps s11 to s16 are optional, for example, step s13 may be optional. Figure 4 In the case where the template feature updating method shown does not include step s13, step s14 may be replaced by: performing data standardization on the verification feature to obtain the verification feature after data standardization. For another example, step s14 may be optional. Figure 4 When the template feature updating method shown does not include step s14, step s15 can be replaced by: fusing the verification feature after the deviation correction with the template feature to obtain the template feature after the object is updated.

[0090] It should be noted that the above embodiment is described by taking one update process as an example, and the above steps can be repeated when the template features are dynamically updated. For example, after step S103, the following is performed: m+1 features to be identified that match the template features of the object in the historical time period are obtained, and updated verification features are obtained based on the average algorithm. The updated verification features are used to describe the concentration trend of the m+1 features to be identified in the historical time period, and the updated verification features are merged with the m+1th template features to obtain the updated template features of the object, that is, the m+2th template features.

[0091] In some optional embodiments, the electronic device may store one template feature, that is, the electronic device stores the template feature obtained by the last update. For example, taking the electronic device performing m update operations as an example, the electronic device may store the m+1th template feature; for another example, taking the electronic device performing m+1 update operations as an example, the electronic device may store the m+2th template feature; and so on.

[0092] It can be seen that in this embodiment, a verification feature for describing the central tendency of the m features to be identified that match the template feature in the historical time period can be obtained, and the verification feature is weightedly fused with the template feature to obtain the updated template feature. Compared with the solution of replacing the template feature with the feature to be identified, in the embodiment of the present application, the updated template feature is related to the verification feature and the template feature before the update, and the updated template feature changes with the change of the verification feature and / or the template feature before the update, so that the updated template feature is conducive to the timely change of the real feature, and the difference between the updated template feature and the real feature is smaller, which can effectively reduce the FRR and / or FAR of biometrics and improve the accuracy of biometrics. At the same time, the verification feature describes the central tendency of the m features to be identified that match the template feature in the historical time period, and the difference between the verification feature and the real feature is smaller. When the template feature is updated using the verification feature, the update direction is more accurate, so that the difference between the updated template feature and the real feature is smaller.

[0093] In addition, in this embodiment, the number of template features stored in the electronic device may be 1. Compared with the solution in which the electronic device stores multiple template features at the same time, this solution can avoid template feature redundancy and space waste. At the same time, it also avoids uneven distribution of template features, which leads to matching failure.

[0094] The following takes the face recognition scenario as an example to describe in detail the flow chart of the template feature update method in the face recognition system. Optionally, the face recognition system may include but is not limited to a weight calculation module and / or a template feature update module. Figure 5 , Figure 5 FIG. 2 shows a flow chart of another template feature updating method. Figure 5 As shown, including but not limited to steps S201-S210:

[0095] S201, obtaining an input template face image.

[0096] Combine the following Figure 6 A schematic diagram of an interface for a user to input a template face image on an electronic device is described in detail.

[0097] See also Figure 6 As shown in (a) of FIG. 1 , the setting interface 61 of the electronic device may include a “biometric identification and password” control 10. The electronic device may detect a touch operation (e.g., a click operation) of the user on the “biometric identification and password” control 10 in the setting interface 61, and in response to the operation, the electronic device may display a biometric identification and password setting interface 62. It should be noted that the electronic device may also trigger the display of the biometric identification and password setting interface 62 in other ways, such as voice, which is not limited in this application.

[0098] See also Figure 6 As shown in (b) in the figure, the electronic device may display a biometric identification and password setting interface 62, which may include a control for face entry, for example, a "face entry" control 20. The electronic device may detect a touch operation (such as a click operation) of the user on the "face entry" control 20 in the biometric identification and password setting interface 62, and in response to the operation, the electronic device may display a face entry interface 63. It should be noted that the electronic device may also trigger the display of the face entry interface 63 in other ways, such as voice, which is not limited in this application.

[0099] See also Figure 6 As shown in (c) of FIG. 5 , the electronic device may display a face recording interface 63, and the face recording interface 63 may include a control for triggering a face recording function, for example, a “start recording” control 30. The electronic device may detect a touch operation (for example, a click operation) of a user on the “start recording” control 30 in the face recording interface 63. In response to the operation, the electronic device may turn on a camera and record a template face image through the camera, such as Figure 6 It should be noted that the electronic device can also trigger the face recording function in other ways, such as voice, and this application does not limit this.

[0100] S202: Check whether the quality score of the template face image is greater than or equal to a second threshold value. If yes, execute step S203; otherwise, end the step.

[0101] Specifically, step S202 includes: determining the quality score of the template face image, and comparing the quality score of the template face image with a second threshold, if the quality score is greater than or equal to the second threshold, executing step S203, otherwise ending the step.

[0102] In an optional implementation, the quality score of the template face image is related to at least one of a brightness evaluation coefficient, a contrast evaluation coefficient, or a clarity evaluation coefficient. Optionally, the brightness evaluation coefficient may be calculated based on the grayscale value of the template face image; the contrast evaluation coefficient may be calculated based on the histogram variance of the template face image; not limited.

[0103] It should be noted that the second threshold can be set based on experience or business requirements, and the specific value is not limited.

[0104] S203: extracting facial features from the template facial image to obtain template features.

[0105] Optionally, facial feature extraction can be divided into facial feature extraction in the time domain and facial feature extraction in the frequency domain. Facial feature extraction in the time domain generally refers to the global or local structural description of the facial image using methods such as affine, mapping, matrix, and the relationship between pixels in the neighborhood. Facial feature extraction in the frequency domain generally refers to converting the facial image into information and processing it using methods such as Fourier transform, wavelet analysis, and wavelet packet decomposition.

[0106] Optionally, before extracting facial features from the template face image to obtain template features, the template face image may be preprocessed. The preprocessing may include but is not limited to at least one of the following: adjusting the brightness of the template face image, denoising, adjusting the contrast of the template face image, or equalizing the template face image.

[0107] It should be noted that steps S201 to S203 describe a detailed implementation example in which the electronic device obtains the first template feature. For a template feature other than the first template feature, the template feature is updated from the previous template feature. For example, taking the embodiment of the present application as an example to describe the mth template feature update process, the template feature refers to the mth template feature, and the mth template feature is updated from the m-1th template feature.

[0108] S204: Obtain a face image to be recognized.

[0109] Optionally, the specific implementation of step S204 can be found in Figure 1 or Figure 2 The relevant embodiments are not described in detail.

[0110] S205: Check whether the quality score of the face image to be recognized is greater than or equal to a second threshold value. If yes, execute step S206, otherwise, end the step.

[0111] S206: extract facial features from the face image to be recognized to obtain verification features.

[0112] It should be noted that the detailed embodiments of step S205 and step S206 can refer to the relevant description of step S202 and step S203, and will not be elaborated herein.

[0113] It should be noted that, taking the embodiment of the present application used to describe the mth template feature update process as an example, the face image to be identified in steps S204 to S206 refers to the mth face image to be identified, and the feature to be identified refers to the mth feature to be identified.

[0114] S207: Based on each feature to be identified and the template feature, determine the similarity between each feature to be identified and the template feature.

[0115] In an optional implementation, taking the embodiment of the present application used to describe the mth template feature update process as an example, step S207 includes: based on the mth feature to be identified and the mth template feature, determining the similarity between the mth feature to be identified and the mth template feature.

[0116] Optionally, before executing the mth template feature update process, the electronic device may store the similarity between the 1st feature to be identified and the 1st template feature to the similarity between the m-1th feature to be identified and the m-1th template feature. That is to say, step S207 specifically includes: obtaining the similarity between the 1st feature to be identified and the 1st template feature to the similarity between the m-1th feature to be identified and the m-1th template feature from the storage area in sequence, and determining the similarity between the mth feature to be identified and the m-1th template feature based on the mth feature to be identified and the mth template feature. It can be seen that the electronic device stores the similarity obtained by historical calculation, which can avoid repeated calculations, save computing resources, and is simpler and more convenient.

[0117] Optionally, the similarity between the feature to be identified and the template feature can be represented by cosine similarity. For example, the template feature V1=(V 1,1 , V 1,2 , V 1,3 ,…V 1,p ), feature V2 to be identified = (V 2,1 , V 2,2 , V 2,3 ,…V 2,p ) as an example, where p is a positive integer, the cosine similarity between the feature to be identified and the template feature can be expressed as formula (7):

[0118]

[0119] In another optional implementation, the similarity between the feature to be identified and the template feature can also be represented by the Euclidean distance. For example, the template feature V1=(V 1,1 , V 1,2 , V 1,3 ,…V 1,p ), feature V2 to be identified = (V 2,1 , V 2,2 , V 2,3 ,…V 2,p ) as an example, the Euclidean distance between the feature to be identified and the template feature can be expressed as formula (8):

[0120]

[0121] It should be noted that the above embodiment is illustrative using cosine similarity and Euclidean distance as examples. In other implementations, the similarity between the feature to be identified and the template feature can also be represented by other contents, such as Hamming distance, which will not be elaborated.

[0122] S208: Detect whether the similarity between the feature to be identified and the template feature is greater than or equal to a first threshold. If yes, execute step S209, otherwise, end the step.

[0123] In an optional implementation, taking the embodiment of the present application used to describe the m-th template feature update process as an example, step S208 includes: detecting whether the similarity between the m-th feature to be identified and the m-th template feature is greater than or equal to the first threshold. If so, execute step S209, otherwise end the step. It should be noted that since the electronic device is executing the m-th template feature update process, the similarity between the 1st feature to be identified to the m-1th feature to be identified and the corresponding template feature is greater than or equal to the first threshold, for example, the similarity between the m-2th feature to be identified and the m-2th template feature is greater than or equal to the first threshold, the similarity between the m-1th feature to be identified and the m-1th template feature is greater than or equal to the first threshold, and so on.

[0124] S209: Determine the exponential decay weight of each feature to be identified based on the corresponding similarities.

[0125] In an optional implementation, taking the embodiment of the present application used to describe the mth template feature update process as an example, step S209 includes: determining the exponential decay weight of the mth feature to be identified based on the similarity between the mth feature to be identified and the mth template feature.

[0126] Optionally, before executing the mth template feature update process, the electronic device may store the exponential decay weights of the 1st feature to be identified to the m-1th feature to be identified. That is, step S209 specifically includes: sequentially obtaining the exponential decay weights of the 1st feature to be identified to the m-1th feature to be identified from the storage area, and determining the exponential decay weight of the mth feature to be identified based on the similarity between the mth feature to be identified and the mth template feature. It can be seen that the electronic device stores the exponential decay weights obtained by historical calculations, which can avoid repeated calculations, save computing resources, and is simpler and more convenient.

[0127] In an optional implementation, based on the similarity between the mth feature to be identified and the mth template feature, the exponential decay weight of the mth feature to be identified is determined, including: substituting the similarity between the mth feature to be identified and the mth template feature into a one-dimensional Gaussian distribution function to obtain the exponential decay weight of the mth feature to be identified.

[0128] Exemplarily, the one-dimensional Gaussian distribution function can be expressed as formula (9):

[0129]

[0130] Among them, score can be used to represent the similarity between the mth feature to be identified and the mth template feature. Optionally, score can be data after data standardization; σ can be used to represent a preset parameter, and the value of σ can be set based on experience or business needs and is not limited. Optionally, the value range of the exponential decay weight calculated by formula (9) can be [0,1].

[0131] It can be seen that in this embodiment, the exponential decay weight of the feature to be identified is positively correlated with the similarity. For example, the higher the similarity, the larger the exponential decay weight. When the feature to be identified is weighted by the exponential decay weight obtained in this way to obtain the verification feature, the verification feature is more accurate in the rate or trend of change of the real feature, which is conducive to the updated template feature to approach the real feature more quickly.

[0132] In another optional embodiment, the exponential decay weight of the mth feature to be identified is determined based on the similarity between the mth feature to be identified and the mth template feature, including: obtaining the quality score of the mth facial image to be identified, and determining the exponential decay weight of the mth feature to be identified based on the quality score and the similarity between the mth feature to be identified and the template feature.

[0133] Optionally, based on the quality score and the similarity between the mth feature to be identified and the template feature, the exponential decay weight of the mth feature to be identified is determined, including: substituting the quality score and the similarity between the mth feature to be identified and the mth template feature into the two-dimensional Gaussian distribution function to obtain the exponential decay weight of the mth feature to be identified.

[0134] Exemplarily, the two-dimensional Gaussian distribution function can be expressed as formula (10):

[0135]

[0136] Among them, score can be used to represent the similarity between the mth feature to be identified and the mth template feature; quality can be used to represent the quality score of the mth face image to be identified; optionally, score can be data after data standardization; quality can also be data after data standardization; σ can be used to represent a preset parameter, and the value of σ can be set based on experience or business needs, and is not limited. Optionally, the value range of the exponential decay weight calculated by formula (10) can be [0, 0.5].

[0137] Optionally, based on the quality score and the similarity between the mth feature to be identified and the template feature, the exponential decay weight of the mth feature to be identified is determined, including: substituting the quality score and the similarity between the mth feature to be identified and the mth template feature into a two-dimensional trigonometric function to obtain the exponential decay weight of the mth feature to be identified.

[0138] Exemplarily, the two-dimensional trigonometric function can be expressed as formula (11):

[0139]

[0140] Among them, score can be used to represent the similarity between the mth feature to be identified and the mth template feature; quality can be used to represent the quality score of the mth face image to be identified.

[0141] It can be seen that in this embodiment, the exponential decay weight of the feature to be identified is positively correlated with the similarity and the quality score. For example, the higher the similarity and the higher the quality score, the larger the exponential decay weight. When the exponential decay weight obtained in this way is used to weight the feature to be identified to obtain the verification feature, the verification feature is more accurate in the rate or trend of change of the real feature, which is conducive to the updated template feature to approach the real feature more quickly.

[0142] It should be noted that when the quality score of the face image to be identified is less than the second threshold, or the similarity between the feature to be identified and the template feature is less than the first threshold, the exponential decay weight of the feature to be identified can be initialized, for example, the weight of the feature to be identified is 0.

[0143] S210 , based on the exponential decay weight of each feature to be identified, weight the m features to be identified to obtain a verification feature, and weightedly fuse the verification feature with the template feature to obtain an updated template feature of the object.

[0144] It should be noted that the detailed implementation of step S210 can refer to the relevant description of step S102 and step S103, and will not be elaborated herein.

[0145] It should be noted that steps S201 to S210 can be respectively performed by a weight calculation module and / or a template feature update module in the face recognition system. For example, step S209 can be performed by a weight calculation module in the face recognition system; for another example, step S210 can be performed by a template feature update module in the face recognition system; and so on.

[0146] It can be seen that in this embodiment, the exponential decay weight of the feature to be identified is positively correlated with the similarity and / or the quality score. For example, the higher the similarity and / or the higher the quality score, the greater the exponential decay weight. When the exponential decay weight obtained in this way is used to weight the feature to be identified to obtain the verification feature, the verification feature is more accurate in the rate or trend of change of the real feature, which is conducive to the updated template feature to approach the real feature more quickly. At the same time, the template feature update method described in the embodiment of the present application can be applied to the face recognition system of the electronic device, and is implemented by the weight calculation module and / or the template feature update module in the face recognition system.

[0147] Need to explain, Figure 5 The face recognition system is used as an example for illustrative explanation. In other implementations, the method described in the embodiments of the present application can also be applied to other biometric systems, such as voiceprint recognition systems, etc., which will not be elaborated on.

[0148] Exemplarily, the template feature updating method mentioned in the embodiments of the present application can be run in an electronic device, which may be a mobile phone, a tablet computer, a desktop computer, a laptop computer, a handheld computer, a notebook computer, a vehicle-mounted device, an ultra-mobile personal computer (UMPC), a netbook, a cellular phone, a personal digital assistant (PDA), an augmented reality (AR)\virtual reality (VR) device, etc. The embodiments of the present application do not impose any special restrictions on the specific form of the electronic device.

[0149] like Figure 7 As shown, it is a schematic diagram of a hardware structure of an electronic device. Among them, the electronic device may include: a processor, an external memory interface, an internal memory, a Universal Serial Bus (USB) interface, a charging management module, a power management module, a battery, an antenna 1, an antenna 2, a mobile communication module, a wireless communication module, a sensor module, a button, a motor, an indicator, a camera, a display screen, and a SIM card slot, etc. Among them, the audio module may include a speaker, a receiver, a microphone, an earphone interface, etc., and the sensor module may include a pressure sensor, a gyroscope sensor, an air pressure sensor, a magnetic sensor, an acceleration sensor, a distance sensor, a proximity light sensor, a fingerprint sensor, a temperature sensor, a touch sensor, an ambient light sensor, a bone conduction sensor, etc.

[0150] It is to be understood that the structures illustrated in the embodiments of the present application do not constitute specific limitations on the electronic device. In other embodiments, the electronic device may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or arrange the components differently. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0151] Among them, the processor may include one or more processing units, for example: the processor may include an application processor (AP), a modem processor (Modem), a graphics processor (GPU), an image signal processor (ISP), a controller, a memory, a video codec, a digital signal processor (DSP), a neural-network processing unit (NPU), etc.

[0152] Among them, different processing units can be independent devices or integrated into one or more processors. The processor is the nerve center and command center of the electronic device. The controller can generate operation control signals according to the instruction operation code and timing signal to complete the control of fetching and executing instructions.

[0153] The wireless communication function of the electronic device can be implemented through antenna 1, antenna 2, mobile communication module, wireless communication module, and Modem, etc. In some embodiments, the antenna 1 of the electronic device is coupled with the mobile communication module, and the antenna 2 is coupled with the wireless communication module, so that the electronic device can communicate with the network side device and other electronic devices through wireless communication technology.

[0154] The display screen is used to display images, videos, etc. The display screen includes a display panel. The display panel can be a liquid crystal display (LCD), an organic light-emitting diode (OLED), an active-matrix organic light-emitting diode or an active-matrix organic light-emitting diode (AMOLED), a flexible light-emitting diode (FLED), Miniled, MicroLed, a quantum dot light-emitting diode (QLED), etc.

[0155] The audio module can be used to encode / decode speech into electrical signals, and can be used to convert electrical signals into speech. In a specific implementation, the audio module can be connected to at least one of the following: a speaker, a receiver, a microphone, or an earphone interface, and can be used to decode / encode the signal input / output by at least one of the above.

[0156] The camera can capture still images or videos. In a specific implementation, the camera can include one or more image sensors (such as a front sensor module or a rear sensor module, not shown), an image signal processor (ISP, not shown) or a flash LED (not shown).

[0157] Among them, the touch sensor is also called a "touch control device". The touch sensor can be set on the display screen, and the touch sensor and the display screen form a touch screen, also called a "touch control screen". The touch sensor is used to detect touch operations acting on or near it. The touch sensor can transmit the detected touch operation to the AP to determine the type of touch event. Visual output related to the touch operation can be provided through the display screen. In other embodiments, the touch sensor can also be set on the surface of the electronic device, which is different from the position of the display screen. In an embodiment of the present application, the electronic device can detect whether there is a user operation acting on the display screen of the electronic device through the touch sensor. After the touch sensor detects the user operation acting on the application, the electronic device can turn on the camera function to obtain the user's facial image.

[0158] In addition, operating systems are running on the above components, such as the iOS operating system developed by Apple, the Android open source operating system developed by Google, and the Windows operating system developed by Microsoft.

[0159] The operating system of the electronic device can adopt a layered architecture, an event-driven architecture, a micro-core architecture, a micro-service architecture, or a cloud architecture. The embodiment of the present application takes the Android system of the layered architecture as an example to illustrate the software structure of the electronic device. It should be noted that although the embodiment of the present application is illustrated by taking the Android system as an example, its basic principle is also applicable to electronic devices based on operating systems such as iOS or Windows.

[0160] See also Figure 8 , is a software structure block diagram of an electronic device provided in an embodiment of the present application. The software structure adopts a layered architecture, which divides the software into several layers, and each layer has a clear role and division of labor. The layers communicate with each other through software interfaces. Taking the Android system as an example, in some embodiments, the Android system is divided from top to bottom into an application layer, an application framework layer (Framework), a hardware abstraction layer (HAL), and a kernel layer (Kernel).

[0161] The application layer may include a series of application packages, including camera, gallery, calendar, call, map, WLAN, Bluetooth, music, video, short message and other APPs.

[0162] Among them, the application framework layer provides an application programming interface (Application Programming Interface, API) and a programming framework for the application of the application layer. The application framework layer includes some predefined functions. For example, the application framework layer may include a window manager, a content provider, a view system, a resource manager, a notification manager, etc. The embodiment of the present application does not impose any restrictions on this. Exemplarily, the above-mentioned window manager is used to manage window programs. The window manager can obtain the size of the display screen, determine whether there is a status bar, lock the screen, capture the screen, etc. The above-mentioned content provider is used to store and obtain data, and make these data accessible to the application. The data may include video, images, audio, dialed and received calls, browsing history and bookmarks, phone books, etc. The above-mentioned view system can be used to build the display interface of the application. Each display interface can be composed of one or more controls. Generally speaking, controls can include interface elements such as icons, buttons, menus, tabs, text boxes, dialog boxes, status bars, navigation bars, widgets, etc. The above-mentioned resource manager provides various resources for the application, such as localized strings, icons, images, layout files, video files, etc. The above notification manager uses an application to display notification information in the status bar, which can be used to convey notification-type messages and can automatically disappear after a short stay without user interaction. For example, the notification manager is used to notify download completion, message reminders, etc. The notification manager can also be a notification that appears in the system top status bar in the form of a chart or scroll bar text, such as a notification of an application running in the background, or a notification that appears on the screen in the form of a dialog window. For example, a text message is prompted in the status bar, a prompt sound is emitted, a vibration is made, an indicator light flashes, etc. In some embodiments, the application framework layer may also include a camera service (Camera Service). The camera service (Camera Service) can interact with the hardware abstraction layer.

[0163] Among them, the hardware abstraction layer is an interface layer located between the kernel layer and the hardware, which can be used to abstract the hardware. In some embodiments, the hardware abstraction layer includes a hardware abstraction layer interface definition language (HIDL) interface. Among them, the hardware abstraction layer may include: camera HAL, display HAL, sensor HAL, etc. In some embodiments, the hardware abstraction layer may include the template feature update module and / or weight calculation module mentioned in the embodiments of the present application.

[0164] The kernel layer is the foundation of the Android operating system, responsible for hardware drivers, network, power, system security, memory management and other functions. The kernel layer is an intermediate layer between hardware and software, and its function is to pass application requests to the hardware. The kernel layer contains at least camera drivers, sensor drivers, display drivers and audio drivers.

[0165] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media integration. The available medium can be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk), etc.

[0166] Those skilled in the art can understand that to implement all or part of the processes in the above-mentioned embodiments, the processes can be completed by computer programs to instruct related hardware, and the programs can be stored in computer-readable storage media. When the programs are executed, they can include the processes of the above-mentioned method embodiments. The aforementioned storage media include: ROM or random access memory RAM, magnetic disk or optical disk and other media that can store program codes.

Claims

1. A template feature updating method, characterized in that: The method comprises: Obtain m features to be identified that match the template features of the object in a historical time period; wherein the template features are reference features of the object in a biometric identification process; and m is a positive integer greater than 1; Based on an average algorithm, obtaining a verification feature for describing a central trend of the m features to be identified in the historical time period; The verification feature is weightedly fused with the template feature to obtain the updated template feature of the object.

2. The method according to claim 1, characterized in that The method of obtaining the verification feature for describing the central tendency of the m features to be identified in the historical time period based on the average algorithm includes: Based on each feature to be identified and the template feature, determining an exponential decay weight of each feature to be identified; Based on the exponential decay weight of each feature to be identified, the m features to be identified are weighted to obtain the verification feature.

3. The method according to claim 2, characterized in that The step of determining the exponential decay weight of each feature to be identified based on each feature to be identified and the template feature comprises: Based on each feature to be identified and the template feature, determining the similarity between each feature to be identified and the template feature respectively; Based on the respectively corresponding similarities, an exponential decay weight of each feature to be identified is determined.

4. The method according to claim 3, characterized in that The determining, based on each feature to be identified and the template feature, the similarity between each feature to be identified and the template feature, respectively, includes: Based on the similarity between the mth feature to be identified and the mth template feature, an exponential decay weight of the mth feature to be identified is determined.

5. The method according to claim 4, characterized in that The step of determining the exponential decay weight of the mth feature to be identified based on the similarity between the mth feature to be identified and the mth template feature includes: Obtaining a quality score of the mth biological data to be identified; the mth biological data to be identified is used to provide the mth feature to be identified; Based on the quality score and the similarity between the mth feature to be identified and the mth template feature, an exponential decay weight of the mth feature to be identified is determined.

6. The method according to claim 5, characterized in that The determining, based on the quality score and the similarity between the mth feature to be identified and the mth template feature, an exponential decay weight of the mth feature to be identified, comprises: Substitute the quality score and the similarity between the m-th feature to be identified and the m-th template feature into a two-dimensional Gaussian distribution function to obtain an exponential decay weight of the m-th feature to be identified.

7. The method according to any one of claims 1 to 6, characterized in that: The step of weightedly fusing the verification feature with the template feature to obtain the updated template feature of the object includes: Obtaining a first preset weight of the verification feature and a second preset weight of the template feature; Based on the first preset weight of the verification feature and the second preset weight of the template feature, the verification feature and the template feature are weightedly fused to obtain the updated template feature of the object.

8. The method according to any one of claims 1 to 7, characterized in that The template feature updating method is executed by a hardware abstraction layer of the electronic device.

9. An electronic device, characterized in that: include: One or more processors and one or more memories; the one or more memories are coupled to the one or more processors, the one or more memories are used to store computer program codes, the computer program codes include computer instructions, and when the one or more processors execute the computer instructions, the electronic device executes the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The method comprises computer instructions, which, when executed on an electronic device, cause the electronic device to execute the method according to any one of claims 1 to 8.

11. A chip or a chip system, characterized in that: The method comprises a processing circuit and an interface circuit, wherein the interface circuit is used to receive code instructions and transmit the code instructions to the processing circuit, and the processing circuit is used to run the code instructions to execute the method according to any one of claims 1 to 8.

Images