A method, device, equipment and storage medium for updating a biometric database

Abstract

The present application relates to a method, apparatus, device and storage medium for updating a biometric database, which includes: obtaining identification information; determining whether the certificate information exists in the information repository, and if not, adding the identification information to the information repository and setting a similarity threshold for each reference image; if it exists, calculating the similarity value between the reference image corresponding to the target reference information and the corresponding feature image; if there is a first feature image and a second feature image, updating the similarity threshold content of all second reference images to the second similarity value; if the similarity value between each reference image and the corresponding feature image is less than the similarity threshold, updating each reference image corresponding to the target reference information to the corresponding feature image, and resetting the similarity threshold of the reference image corresponding to the target reference information to the corresponding initial value. The present application has the effect of reducing the operating pressure of the device caused by frequent updating of the information feature database.

Description

Technical Field

[0001] This application relates to the field of biometric technologies, and in particular, to a method, apparatus, device, and storage medium for updating a biometric database. Background Art

[0002] Biometric-based identity recognition technology (hereinafter referred to as biometric recognition) refers to a technology that uses the inherent physiological or behavioral characteristics of an organism to identify an individual's identity; among them, biometric characteristics can specifically include facial characteristics, fingerprint characteristics, voice characteristics, etc.; a biometric recognition device can be used to identify a user's identity through biometric characteristics.

[0003] Specifically, information of a user is pre-recorded and stored in the biometric recognition device. The above information can include reference information for distinguishing the user's identity (the reference information is permanent and unique, such as an ID number), and also includes various characteristic information for characterizing the user's biometric characteristics. The characteristic information is stored in the form of an image (hereinafter collectively referred to as a reference image). The reference image can be a facial image, a fingerprint image, etc.; the identity feature recognition device is pre-connected to a device for receiving the above information through communication. After the biometric recognition device receives the above information of the user to be recognized, it determines whether the image information of the user to be recognized is pre-stored in the biometric recognition device by comparing the similarity value of the images, so as to achieve identity identification; if the comparison result of the reference information is consistent and the similarity value of the image information is less than a preset similarity threshold, the pre-stored image information needs to be updated to the current image information of the user to be recognized, so as to achieve information update.

[0004] Regarding the above related technologies, the inventors found that there are at least the following problems in this technology: Based on the above technology, if the similarity between the user's facial feature image and the pre-stored reference image is less than the preset similarity threshold, the pre-stored image information needs to be updated. However, if the change in the user's face is temporary (such as the change duration is only one day), such as wearing makeup; but wearing makeup by the user is a temporary behavior. If the above technology is followed, the large change in the user's face before and after makeup will result in a small image similarity, and then the biometric recognition device will frequently update the corresponding feature image, resulting in an increase in the computational amount of the biometric recognition device and an increase in the operating burden. Summary of the Invention

[0005] In order to improve the technical problem of the increased operating burden of the biometric recognition device caused by the frequent update of the biometric recognition device in the related technologies, this application provides a method, apparatus, device, and storage medium for updating a biometric database.

[0006] In a first aspect, a method for updating a biometric database provided by this application adopts the following technical solution:

[0007] A biometric database update method, comprising the following steps:

[0008] Obtain identification information, where the identification information at least includes document information and several feature images for reflecting different biometric features of a user;

[0009] Determine whether the document information exists in a preset information repository. If it does not exist, add the identification information to the information repository and set a similarity threshold for each reference image according to a preset rule; where the information repository stores reference information corresponding to the document information;

[0010] If there is a target reference information that is consistent with the document information, calculate the similarity value between the reference image corresponding to the target reference information and the corresponding feature image; where each reference information in the information repository corresponds to several reference images, and the reference images and feature images are in one-to-one correspondence, each reference image is preset with a similarity threshold, and each similarity threshold corresponds to an initial value;

[0011] If there is a first similarity value between a first feature image and a corresponding first reference image that is greater than the similarity threshold of the first reference image, and there is a second similarity value between a second feature image and a corresponding second reference image that is less than the similarity threshold of the second reference image, update the similarity threshold content of all the second reference images to the corresponding second similarity value; if the similarity value between each reference image corresponding to the target reference information and the corresponding feature image is less than the similarity threshold of the corresponding reference image, update each reference image corresponding to the target reference information to the corresponding feature image, and reset the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value.

[0012] By adopting the above technical solution, the document information can be an identity card number. It is possible to determine whether the information of the identified user pre-exists in the information repository through the document information. If not, the identification information of the user is stored in the information repository to update the information repository. If it exists, the similarity between each of the reference images corresponding to the target reference information consistent with the document information and the corresponding feature images is calculated. The feature images and the reference images are in one-to-one correspondence and are images used to represent different biometric features of the user. If both the first feature image and the second feature image exist in the identification information, then at this time, the similarity threshold corresponding to the second reference image is replaced with the second similarity value, thereby reducing the similarity threshold and reducing the possibility of the reference image being updated, avoiding the situation of frequently updating the reference image due to the similarity value being less than the similarity threshold. Only when the similarity values between all the feature images in the identification information and the corresponding reference images are less than the similarity threshold, is it necessary to perform an update.

[0013] Preferably, if there is target reference information consistent with the document information, calculating the similarity value between the reference image corresponding to the target reference information and the corresponding feature image includes:

[0014] If there is target reference information consistent with the document information, calculate the similarity value between the reference image corresponding to the target reference information and the corresponding feature image, and store the similarity value, the corresponding reference image, the corresponding feature image, and the calculation time of the similarity value in a preset historical record table in the order of the calculation time.

[0015] The method further includes:

[0016] At fixed intervals, if there is a third reference image in the historical record table, determine the minimum value among the similarity values corresponding to the calculation times of the three closest to the current time for the third reference image, update the reference image in the information repository that is consistent with the third reference image to the feature image corresponding to the minimum value, and reset the similarity threshold content of the reference image that is consistent with the third reference image to the corresponding preset initial value; wherein, the third reference image refers to a reference image for which the difference between any two of the similarity values corresponding to the calculation times of the three closest to the current time is less than a preset difference.

[0017] By adopting the above technical solution, in the historical record table, for the three feature images corresponding to the first reference image in the most recent three times, the differences between the similarity values of these three feature images are all less than the preset difference. That is, these three feature images are similar, which then indicates that the biometric feature corresponding to the above first reference image has reached a stable state and is not likely to change. At this time, use the feature image corresponding to the minimum value among the similarity values of the first reference image in the most recent three times to replace the reference image that is consistent with the third reference avatar in the information repository, so as to update the reference image.

[0018] Preferably, the method further includes:

[0019] Every preset time period, calculate the average value of all similarity values corresponding to each reference image in the historical record table within the preset time period, and use the average value content to replace the initial value of the similarity threshold corresponding to the reference image.

[0020] By adopting the above technical solution, adjust the initial value of the corresponding similarity threshold according to the average value of the similarity values corresponding to each reference image, and reduce the situation that due to inappropriate setting of the initial value, the deviation between the similarity value of the feature image and the reference image and the set initial value is relatively large, which may further lead to the need to adjust the similarity threshold or even the need to update the reference image. Therefore, the initial value of the similarity threshold can be determined according to the average value of the actually calculated similarity values between each reference image and the corresponding feature image, so as to realize the adjustment of the initial value of the artificially set similarity threshold.

[0021] Preferably, the step of "updating the similarity threshold content of all the second reference images to the corresponding second similarity value" includes:

[0022] Update the similarity threshold content of all the second reference images to the corresponding second similarity value, add all the second reference images to a preset adjustment record table, and store the current time as the addition time in the adjustment record table.

[0023] The method further includes:

[0024] Based on the addition time, every specified time period, determine the number of times each reference image in the adjustment record table appears within the specified time period; according to a preset corresponding table, determine the specific time value of the preset time period corresponding to each reference image in the adjustment record table, and add the specific time value of the corresponding preset time period to the corresponding reference image in the historical record table; wherein, the corresponding table is used to store the corresponding relationship between the number of times the reference image appears and the specific time value of the preset time period, and the more times it appears, the shorter the specific time value of the preset time period.

[0025] At each preset time interval, calculate the average value of all similarity values corresponding to each reference image in the historical record table within the preset time interval, including:

[0026] Based on the corresponding table, determine the specific time value of the preset time interval corresponding to each reference image in the historical record table;

[0027] At each corresponding preset time interval value, calculate the average value of all similarities corresponding to each reference image in the historical record table within the corresponding preset time interval value.

[0028] By adopting the above technical solution, according to the adjustment frequency of the similarity threshold, determine the adjustment frequency of the initial value of the similarity threshold of each reference image, that is, determine the specific duration of the preset time interval corresponding to each first reference image.

[0029] Preferably, updating each reference image corresponding to the target reference information to the corresponding feature image, and resetting the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value, includes:

[0030] Generate and display verification information based on the recognition information corresponding to the target reference image, where the verification information includes the recognition information;

[0031] Receive the confirmation information sent by the user through the terminal. If there is modified document information in the confirmation information, store the modified document information in the information storage repository, and store each feature image in the verification information corresponding to the confirmation information as a new reference image in the information storage repository, and set a similarity threshold for each new reference image according to a preset rule;

[0032] If there is no modified document information in the confirmation information, update each reference image corresponding to the target reference information to the corresponding feature image, and reset the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value.

[0033] By adopting the above technical solution, before updating the reference images in the information storage repository, send the verification information with the recognition information to the user again for the user to check again whether the document information is incorrect. If the document information is correct, directly update it. If the document information is incorrect, add it as a new user to the information storage repository.

[0034] Preferably, all feature images in the identification information corresponding to the confirmation information are included in the confirmation information; if there is modified certificate information in the confirmation information, the modified certificate information is stored in the information repository, and each feature image in the verification information corresponding to the confirmation information is stored in the information repository as a newly added reference image, and a similarity threshold is set for each newly added reference image according to a preset rule, including:

[0035] If there is modified certificate information in the confirmation information, the modified certificate information is compared with each reference information in the information repository. If the modified certificate information does not exist in the information repository, the modified certificate information is stored in the information repository, and each feature image in the verification information corresponding to the confirmation information is stored in the information repository as a newly added reference image, and a similarity threshold is set for each newly added reference image according to a preset rule;

[0036] If the modified certificate information exists in the information repository, new identification information is generated based on the modified certificate information and all feature images in the corresponding confirmation information.

[0037] The obtaining of the identification information includes:

[0038] Obtaining the identification information sent by the user through the intelligent terminal or obtaining the automatically generated new identification information.

[0039] By adopting the above technical solution, after receiving the confirmation information, it is determined whether the modified certificate information exists in the information repository. If it exists, new identification information is automatically generated based on the modified certificate information and all corresponding feature images for subsequent re-verification of the new identification information. If it does not exist, it means that the user corresponding to the modified certificate information is a new user, and the modified certificate information and all corresponding feature images are added to the information repository.

[0040] Preferably, the method further includes:

[0041] Every specified period, it is determined whether there is a fourth reference image or a fifth reference image, where the fourth reference image exists in the information repository and does not exist in the historical record repository; the fifth reference image exists in the historical record repository, and the difference between the latest calculation time corresponding to the fifth reference image in the historical record repository and the current time is greater than a preset time difference;

[0042] If there is a fourth reference image or a fifth reference image, the fourth reference image or the fifth reference image is deleted from the information storage table.

[0043] By adopting the above technical solution, if the customer corresponding to the information entered in the information repository has not performed the operation of identifying or updating the reference image within the specified period, then in order to save the storage space in the information storage table, the corresponding reference image can be deleted.

[0044] In a second aspect, a biometric database updating device provided by the present application includes:

[0045] An identification information acquisition module for acquiring identification information of a user, where the identification information at least includes document information and several feature images for reflecting different biometric features of the user;

[0046] A document information verification module for determining whether the document information exists in a preset information repository. If not, adding the identification information to the information repository and setting a similarity threshold for each reference image according to a preset rule; where the information repository stores reference information corresponding to the document information;

[0047] A feature image verification module for calculating a similarity value between the reference image corresponding to the target reference information and the corresponding feature image when there is target reference information consistent with the document information; where each reference information in the information repository corresponds to several reference images, and the reference images and the feature images are in one-to-one correspondence, each reference image corresponds to a preset similarity threshold, and each similarity threshold corresponds to an initial value;

[0048] A similarity threshold adjustment module for updating the similarity threshold content of all the second reference images to the corresponding second similarity value when there is a first similarity value between the first feature image and the corresponding first reference image greater than the similarity threshold of the first reference image and there is a second similarity value between the second feature image and the corresponding second reference image less than the similarity threshold of the second reference image;

[0049] A feature image update module for updating each reference image corresponding to the target reference information to the corresponding feature image and resetting the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value when the similarity value between each reference image corresponding to the target reference information and the corresponding feature image is less than the similarity threshold of the corresponding reference image.

[0050] In a third aspect, a biometric database updating device provided by the present application includes a memory and a processor, and a computer program capable of being loaded and executed by the processor, such as any method in the first aspect, is stored on the memory.

[0051] Fourthly, a computer-readable storage medium provided by the present application stores a computer program that can be loaded and executed by a processor to perform any of the methods in the first aspect.

[0052] In summary, the present application includes at least one of the following beneficial technical effects:

[0053] 1. The biometric database update method first determines whether the document information exists in the information repository; if not, the user is regarded as a new user and the information of the new user is entered. If it exists, the similarities between the face feature image and the fingerprint feature image and the previously entered reference image are calculated respectively. If the similarities of the above two feature images are both lower than the preset similarity threshold, an update is required. If the similarity of any one of the feature images is higher than the preset similarity threshold, it is only necessary to lower the similarity threshold value of the reference image corresponding to the feature image with a similarity lower than the preset similarity threshold to the calculated similarity value, so that the similarity calculated between the feature image and the reference image can be higher than the similarity threshold, thereby reducing the frequency of updating the reference image and reducing the computational burden caused by frequently updating the reference image. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0055] Figure 1 It is a flowchart of a biometric database update method in an embodiment.

[0056] Figure 2 It is a structural block diagram of a biometric update device in an embodiment.

[0057] Description of the reference numerals: 1. Identification information acquisition module; 2. Document information verification module; 3. Feature image verification module; 4. Similarity threshold adjustment module; 5. Feature image update module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will further describe the embodiments of the present invention in detail with reference to the drawings.

[0059] The embodiments of the present application disclose a biometric database update method. Refer to Figure 1, the biometric database update method is used to verify the latest obtained document information and feature image of a user based on the pre-stored reference information of the user (such as the document number for identifying the user's identity, like the ID card number), and the reference image (such as the face image representing the user's facial features initially entered, the fingerprint image representing the user's fingerprint features, etc.). Specifically, it realizes the verification of the user's identity by comparing the document information with the reference information and calculating the similarity between the reference image and the feature image. When the document information is inconsistent, or the image similarity is less than the preset similarity threshold, the reference information and / or the reference image are updated.

[0060] The execution subject of the biometric database update method disclosed in this application is a biometric database update device, supplemented by a touch display screen, a shooting device, and a document recognition device; the touch display screen, the shooting device, and the document recognition device are all communicatively connected to the biometric update device; among them, the shooting device can be a digital camera for shooting images reflecting the user's biometric features, and the document recognition device can be an ID card reader for reading the user's ID card number; an identification button can be displayed on the touch display screen. The biometric database update device includes a memory and a processor, which stores the reference information and reference image of the user. When the user touches the identification button, the biometric database update device receives the signal sent by the touch display screen, controls the shooting device to shoot the image of the user's biometric features, and controls the document recognition device to recognize the user's document information. Finally, it receives the above-mentioned document information and image and performs subsequent processing operations.

[0061] The following combines the attached Figure 1 , and details the specific optimization process of the above biometric database update method as follows:

[0062] Step 101, obtain identification information, where the identification information includes at least document information and several feature images for reflecting different biometric features of the user.

[0063] In implementation, the biometric database update device receives the feature images sent by the shooting device and the document information sent by the document recognition device, and the biometric database generates identification information based on the above information.

[0064] Step 102, determine whether the document information exists in the preset information storage repository. If not, add the identification information to the information storage repository and set a similarity threshold for each reference image according to the preset rule; where the information storage repository stores the reference information corresponding to the document information.

[0065] In implementation, the information repository is used to store the reference information and reference images initially entered by the user. The reference information is used to identify the user's identity document information. Therefore, the reference information corresponds to the document information. The reference image is an image used to reflect the user's biometric characteristics. Therefore, the reference image corresponds to the feature image.

[0066] In addition, in this embodiment, the content of the reference information and the document information is the ID number. Therefore, the reference information is an invariant and unique piece of information. There are two types of reference images, which are respectively used to reflect the user's facial features and fingerprint features. To distinguish different reference images of the same user, each reference image is set with an image name. The naming rule can specifically be the last four digits of the ID number plus a suffix for distinguishing biometric features. The letter suffix 'a' can be used to represent facial features, and the letter suffix 'b' can be used to represent fingerprint features. Therefore, the image name can be, for example: 3679a.

[0067] Correspondingly, the information repository can be stored in the form of a table, specifically including fields such as "ID number", "reference image a", and "reference image b"; both the "reference image aa" field or the "reference image b" field include the image name and the corresponding reference image, as shown in Table 1

[0068] Table 1 Information Repository

[0069]

[0070] The biometric database updating device compares the document information in the recognition information with the reference information in the information repository one by one to determine whether the document information exists in the preset information repository. If it does not exist, it means that the document information has not been entered into the information repository. At this time, the biometric database updating device adds the above-mentioned document information to the information repository and adds the feature image corresponding to the document information as a reference image to the information repository; since there is more than one type of feature image, in order to be able to add the feature image to the correct reference image field (i.e., reference image a or reference image b in Table 1), the biometric database updating device will add an image name to each feature image when generating the recognition information.

[0071] Specifically, each imaging device can be numbered in advance. Imaging devices with different numbers are used to capture different biometric features. The biometric database update device stores the correspondence between the numbers and the biometric features. When the biometric database update device receives a feature image, it will determine the image name suffix (a or b) of the corresponding feature image according to the above correspondence. When generating the recognition information, the biometric database update device will add the last four digits of the certificate information as the prefix of the image name, thus forming the image name of the feature image. When the biometric database update device enters the feature image as a new reference image into the information storage repository, it can determine the corresponding reference image field according to the image name suffix of the feature image, and at the same time, the image name of the feature image will become the image name of the new reference image.

[0072] Step 103: If there is target reference information that is consistent with the certificate information, calculate the similarity value between the reference image corresponding to the target reference information and the corresponding feature image. Each reference information in the information storage repository corresponds to a number of reference images, and the reference images and the feature images are in one-to-one correspondence. Each reference image is preset with a similarity threshold, and each similarity threshold corresponds to an initial value.

[0073] In implementation, after the biometric database update device compares the certificate information with each reference information, if there is target reference information, that is, there is reference information that is consistent with the content of the certificate information, it means that the corresponding certificate information has been previously entered into the information storage repository. At this time, the biometric database update device calculates the image similarity value between each feature image in the recognition information and the reference image corresponding to the target reference information in the information storage repository respectively. Specifically, the biometric database update device will determine the one-to-one correspondence between the feature image and the reference image according to the image name of the feature image and the image name of the reference image. Then, the biometric database update device calculates the image similarity value between the corresponding feature image and the reference image. Among them, the similarity threshold refers to the similarity lower limit value data (as shown in Table 1) set artificially for each reference image and pre-stored in the information storage repository, and each similarity threshold corresponds to an initial value. The initial values of all similarity thresholds are the same. The similarity threshold is used to enable the biometric database update device to compare the calculated similarity value with the similarity threshold of the corresponding reference image after calculating the similarity value, so as to determine the similarity degree between the feature image and its corresponding reference image. If the calculated similarity value is lower than the similarity threshold, it means that the similarity degree is low.

[0074] Step 104, if there exists a first similarity value between the first feature image and the corresponding first reference image that is greater than the similarity threshold of the first reference image, and there exists a second similarity value between the second feature image and the corresponding second reference image that is less than the similarity threshold of the second reference image, then update the similarity threshold content of all the second reference images to the corresponding second similarity value.

[0075] In implementation, the biometric database updating device compares the calculated similarity value with the similarity threshold of the corresponding reference image. The first feature image refers to a feature image whose similarity value with the corresponding reference image is greater than the similarity threshold, that is, a feature image with a high similarity degree to the corresponding reference image. The first reference image refers to the reference image corresponding to the first feature image, and the first similarity value refers to the similarity value obtained after comparing the similarity between the first reference image and the first feature image. The second feature image refers to a feature image whose similarity value with the corresponding reference image is less than the similarity threshold, that is, a feature image with a low similarity degree to the corresponding reference image. Similarly, the second reference image refers to the reference image corresponding to the second feature image, and the second similarity value refers to the similarity value obtained after comparing the similarity between the second reference image and the second feature image.

[0076] If there exists a first feature image and a second feature image, then the biometric database updating device will lower the similarity threshold of all the second reference images, lower the similarity threshold of the second reference image from the initial value to the corresponding second similarity value, thereby restricting the conditions for the reference image to be updated, reducing the probability of the reference image being updated, and reducing the operation burden of the biometric database updating device caused by frequent updating of the reference image.

[0077] Step 105, if the similarity value between each reference image corresponding to the target reference information and the corresponding feature image is less than the similarity threshold of the corresponding reference image, then update each reference image corresponding to the target reference information to the corresponding feature image, and reset the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value.

[0078] In implementation, the biometric database updating device compares the calculated similarity value with the similarity threshold of the corresponding reference image. If the similarity value between each reference image corresponding to the target reference information and the corresponding feature image is less than the similarity threshold of the corresponding reference image, then it indicates that the similarity degree between all the feature images and the corresponding reference images is low. At this time, the corresponding reference image needs to be updated, that is, the biometric database updating device replaces the corresponding reference image with the feature image, and resets the similarity threshold content corresponding to the replaced reference image to the initial value; thereby realizing the update of the reference image.

[0079] Optionally, the step of "if there is a target reference information that matches the certificate information in step 103, calculate the similarity value between the reference image corresponding to the target reference information and the corresponding feature image" further includes the following processing:

[0080] In implementation, after the biometric database update device determines the existence of the target reference information and calculates the similarity value between the reference image corresponding to the target reference information and the corresponding feature image, the biometric update device stores the calculated similarity value, the corresponding reference image, the corresponding feature image, and the calculation time of the similarity value in a preset historical record table in the order of the calculation time, to achieve archiving; where the above calculation time refers to the time when the similarity value is calculated. In addition, during the process of storing the reference image and the feature image in the historical record table, the image name of the reference image and the image name of the feature image will also be stored in the historical record table. The historical record table is shown in Table 2 below:

[0081] Table 2 Historical Record Table

[0082]

[0083] The historical record table distinguishes each piece of data by the reference image. The data with the same reference image occupies one row of the table, and then different calculation times are used to distinguish different data of the same reference image. That is, in each row, the feature image and similarity value information corresponding to different calculation times occupy one column of the row where they are located, and the calculation time 1 in Table 2 is earlier than the calculation time 2.

[0084] The biometric database update method further includes the following processing:

[0085] The biometric database update device determines whether there is a third reference image in the historical record table every fixed duration (such as 72 hours). The third reference image refers to the reference image corresponding to any two similarity values whose difference is less than the preset difference among the similarity values corresponding to the three most recent calculation times (that is, the three rightmost columns in each row of Table 2) in the historical record table. If it exists, it means that the three similarity values corresponding to the three most recent calculation times of the third reference image are small in difference, that is, it means that the feature image corresponding to the reference image has little change in the most recent three recognition processes and reaches a stable state; at this time, the biometric database update device will use the feature image corresponding to the smallest similarity value among the above three similarity values to update the reference image in the information storage library that is consistent with the third reference image.

[0086] Optionally, the biometric database update method further includes the following processing:

[0087] The biometric database updating device will, every preset time period (such as 7 days), calculate the average value of all similarity values corresponding to each reference image in the historical record table within the preset time period, and use the calculated average value to replace the preset initial value of the similarity threshold corresponding to the corresponding reference image; that is, adjust the initial value of the similarity threshold corresponding to the reference image according to the actually calculated similarity values of each reference image in history.

[0088] Optionally, "then update the similarity threshold content of all second reference images to the corresponding second similarity value" in step 104 includes the following processing:

[0089] When the biometric database updating device updates the similarity threshold content of the second reference image to the corresponding second similarity value, it adds the second reference image to a preset adjustment record table, and stores the current time as the addition time in the adjustment record table.

[0090] The biometric database updating method further includes the following processing:

[0091] The biometric database updating device will, every specified time period, based on the addition time, determine the number of times each reference image in the adjustment record table appears within the specified time period, and then, according to a corresponding table prestored to reflect the corresponding relationship between the number of times the stored reference image appears and the specific time value of the preset time period, determine the specific time value of the preset time period corresponding to each reference image in the adjustment record table, where the more the number of appearances, the shorter the specific time value of the preset time period.

[0092] Then the biometric database updating device adds the determined specific time value of the preset time period to the corresponding reference image in the information storage library in an updated manner (as shown in Table 2). If the corresponding specific time value of the preset time period has already been stored at the corresponding reference image in the information storage library before the current time, then the biometric database updating device will overwrite the original specific time value of the preset time period with the currently determined specific time value of the preset time period.

[0093] Correspondingly, "the biometric database updating device will, every preset time period (such as 7 days), calculate the average value of all similarity values corresponding to each reference image in the historical record table" in the foregoing content specifically includes the following processing: The biometric database updating device will first determine the specific time value of the preset time period for each reference image in the historical record table, and then, every determined specific time value of the preset time period, calculate the average value of all similarity values of the reference image corresponding to the above specific time value of the preset time period in the historical record table within the corresponding specific time value of the preset time period.

[0094] Optionally, before updating the reference image, to ensure the accuracy of the document information and avoid the situation where a user mistakenly uses someone else's ID card for identification, the following processing is also included in "updating each reference image corresponding to the target reference information to the corresponding feature image and resetting the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value" in step 105:

[0095] The biometric database update device will generate verification information based on the identification information corresponding to the target reference image and send the verification information to the touch display screen for display. The verification information includes the identification information and a reminder text for reminding the user to verify. The reminder text is, for example, "Please check the following information". When the biometric database update device sends the verification information to the touch display screen, it will control the touch display screen to display an input box, a button for completing the input, and a button for verifying that there is no error.

[0096] When the user finds that the document information is correct and touches the button for verifying that there is no error, the biometric database update device will receive a confirmation information without the modified document information. At this time, the biometric database update device will update each reference image corresponding to the target reference information to the corresponding feature image and reset the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value.

[0097] When the user finds that the document information is incorrect, enters the correct document information in the input box and touches the button for completing the input, the biometric database update device will receive a confirmation information with the modified document information (i.e., the content entered in the input box). At this time, the biometric database update device will compare the modified document information with each reference information in the information storage library. If the modified document information does not exist in the information storage library, the biometric database update device will store each feature image in the verification information corresponding to the confirmation information as a new reference image in the corresponding reference image field in the information storage library and set a similarity threshold for each new reference image, where the similarity threshold content is the preset initial value;

[0098] If the modified certificate information exists in the information repository, the biometric database update device generates new identification information based on the modified certificate information and all the feature images in the corresponding confirmation information. Correspondingly, "obtaining identification information" in step 101 specifically includes two types of identification information. The first type is the identification information generated by the biometric database update device after receiving the certificate information and image information from the imaging device and the certificate recognition device after the user touches the recognition button on the touch display screen. The second type of identification information is the identification information generated by the biometric database update device when it receives the confirmation information from the user and determines that the modified certificate information exists in the information repository, so as to determine the similarity degree between the feature images in the above two types of identification information and the corresponding reference images in subsequent operations.

[0099] Optionally, the biometric database update method further includes the following processing:

[0100] The biometric database update device determines whether there is a fourth reference image or a fifth reference image every specified period (such as 30 days). Specifically, the biometric database update device can determine whether there is a fourth reference image or a fifth reference image according to the image name. Among them, the fourth reference image is a reference image that exists in the information repository but does not exist in the historical record library, indicating that the user corresponding to the fourth reference image has not been recognized and updated in the last 30 days. The fifth reference image refers to the reference image that exists in the historical record table and the difference between the corresponding latest calculation time in the historical record table and the current time is greater than the preset time difference (such as 30 days), indicating that the user corresponding to the fifth reference image has not been recognized and updated within 30 days. If there is a fourth reference image or a fifth reference image, in order to release the storage space of the biometric database update device, the biometric database update device deletes the fourth reference image and the fifth reference image from the information repository.

[0101] In summary, the biometric database update method is used to verify the identity of a user by three methods: identifying the user's document information, face feature image, and fingerprint feature image. Since the user's document information is constant, the biometric database update method first determines whether the document information exists in the information repository. If it does not exist, the user is regarded as a new user and the information of the new user is entered. If it exists, the similarity between the face feature image and the fingerprint feature image and the previously entered reference images is calculated respectively. If the similarity of both of the above feature images is lower than the preset similarity threshold, an update is required. If the similarity of any one of the feature images is higher than the preset similarity threshold, the similarity threshold value of the reference image corresponding to the feature image with a similarity lower than the preset similarity threshold only needs to be adjusted down to the calculated similarity value, so that the similarity calculated between the feature image and the reference image can be higher than the similarity threshold, thereby reducing the frequency of updating the reference image and reducing the computational burden caused by frequently updating the reference image.

[0102] An embodiment of the present application also discloses a biometric database update device. It includes:

[0103] An identification information acquisition module 1, which acquires identification information of a user. The identification information includes at least document information and several feature images for reflecting different biometric features of the user;

[0104] A document information verification module 2, which is used to determine whether the document information exists in a preset information repository. If it does not exist, the identification information is added to the information repository, and a similarity threshold is set for each reference image according to a preset rule. The information repository stores reference information corresponding to the document information;

[0105] A feature image verification module 3, which is used to calculate the similarity value between the reference image corresponding to the target reference information and the corresponding feature image when there is target reference information consistent with the document information. Each reference information in the information repository corresponds to several reference images, and the reference images and the feature images are in one-to-one correspondence. Each reference image is preset with a similarity threshold, and each similarity threshold corresponds to an initial value;

[0106] A similarity threshold adjustment module 4, which is used to update the similarity threshold content of all second reference images to the corresponding second similarity value when there is a first similarity value between a first feature image and a corresponding first reference image greater than the similarity threshold of the first reference image and there is a second similarity value between a second feature image and a corresponding second reference image less than the similarity threshold of the second reference image;

[0107] The feature image update module 5 is used to update each reference image corresponding to the target reference information to the corresponding feature image when the similarity value between each reference image corresponding to the target reference information and the corresponding feature image is less than the similarity threshold of the corresponding reference image, and reset the similarity threshold content of each reference image corresponding to the target reference image to the corresponding initial value.

[0108] Optionally, the feature image verification module 3 is further used to calculate the similarity value between the reference image corresponding to the target reference information and the corresponding feature image when there is target reference information consistent with the certificate information, and store the similarity value, the corresponding reference image, the corresponding feature image, and the calculation time of the similarity value in the preset historical record table in the order of the calculation time.

[0109] It further includes a steady-state update module, which is used to determine the minimum value among the similarity values corresponding to the calculation times of the third reference image in the three closest calculation times to the current time if there is a third reference image in the historical record table every fixed period, update the reference image in the information repository that is consistent with the third reference image to the feature image corresponding to the minimum value, and reset the similarity threshold content of the reference image that is consistent with the third reference image to the corresponding preset initial value; wherein, the third reference image refers to a reference image in which the difference between any two of the similarity values corresponding to the calculation times in the three closest calculation times to the current time is less than the preset difference.

[0110] Optionally, it further includes an initial value adjustment module, which is used to calculate the average value of all similarity values corresponding to each reference image in the historical record table every preset period, and replace the initial value of the similarity threshold of the corresponding reference image with the average value content.

[0111] Optionally, the similarity threshold adjustment module 4 is further used to add all the second reference images to the preset adjustment record table, and store the current time as the addition time in the adjustment record table.

[0112] It further includes a preset period determination module, which is used to determine the number of occurrences of each reference image in the adjustment record table every specified period based on the addition time; it is also used to determine the specific time value of the preset period corresponding to each reference image in the adjustment record table according to the preset correspondence table, and add the specific time value of the corresponding preset period to the corresponding reference image in the historical record table; wherein, the correspondence table is used to store the correspondence relationship between the number of occurrences of the reference image and the specific time value of the preset period.

[0113] The initial value adjustment module is further configured to determine the specific time value of the preset duration corresponding to each reference image in the historical record table based on the corresponding table; and is further configured to, at every preset duration time value, calculate the average value of all similarities corresponding to each reference image in the historical record table within the corresponding preset duration time value.

[0114] Optionally, the feature image update module 5 is further configured to generate and display verification information based on the recognition information corresponding to the target reference image, where the verification information includes the recognition information; and is further configured to receive the confirmation information sent by the user through the terminal. If there is modified document information in the confirmation information, the modified document information is stored in the information storage repository, and each feature image in the verification information corresponding to the confirmation information is stored in the information storage repository as a newly added reference image, and a similarity threshold is set for each newly added reference image according to a preset rule; and is further configured to, if there is no modified document information in the confirmation information, update each reference image corresponding to the target reference information to the corresponding feature image, and reset the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value.

[0115] Optionally, the confirmation information includes all feature images in the recognition information corresponding to the confirmation information; the feature image update module 5 is further configured to, if there is modified document information in the confirmation information, compare the modified document information with each reference information in the information storage repository; and is further configured to, if the modified document information does not exist in the information storage repository, store the modified document information in the information storage repository, and store each feature image in the verification information corresponding to the confirmation information in the information storage repository as a newly added reference image, and set a similarity threshold for each newly added reference image according to a preset rule; and is further configured to, if the modified document information exists in the information storage repository, generate new recognition information based on the modified document information and all feature images in the corresponding confirmation information.

[0116] The recognition information acquisition module 1 is further configured to acquire the recognition information sent by the user through the intelligent terminal or acquire the automatically generated new recognition information.

[0117] Optionally, it further includes a reference image deletion module, configured to determine whether there is a fourth reference image or a fifth reference image at every specified period, where the fourth reference image exists in the information storage repository but does not exist in the historical record repository; the fifth reference image exists in the historical record repository, and the difference between the latest calculation time corresponding to the fifth reference image in the historical record repository and the current time is greater than a preset time difference; and is further configured to, if there is a fourth reference image or a fifth reference image, delete the fourth reference image or the fifth reference image from the information storage table.

[0118] Those of ordinary skill in the art can understand that all or part of the steps to implement the above embodiments can be completed by hardware, or can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a disk, an optical disc, etc.

[0119] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A biometric database update method, characterized in that: It includes the following steps: Obtain identification information, where the identification information at least includes document information and several feature images for reflecting different biometric features of the user; Determine whether the document information exists in a preset information repository. If it does not exist, add the identification information to the information repository and set a similarity threshold for each reference image according to a preset rule; Wherein the information repository stores reference information corresponding to the document information; If there is a target reference information consistent with the document information, calculate the similarity value between the reference image corresponding to the target reference information and the corresponding feature image; Wherein each reference information in the information repository corresponds to several reference images, and the reference images and feature images are in one-to-one correspondence. Each reference image is preset with a similarity threshold, and each similarity threshold corresponds to an initial value; If the first similarity value between the first feature image and the corresponding first reference image is greater than the similarity threshold of the first reference image, and the second similarity value between the second feature image and the corresponding second reference image is less than the similarity threshold of the second reference image, update the similarity threshold content of all the second reference images to the corresponding second similarity value; If the similarity value between each reference image corresponding to the target reference information and the corresponding feature image is less than the similarity threshold of the corresponding reference image, update each reference image corresponding to the target reference information to the corresponding feature image, and reset the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value.

2. The biometric database updating method according to claim 1, wherein: The step of calculating the similarity value between the reference image corresponding to the target reference information and the corresponding feature image when there is a target reference information consistent with the document information includes: If there is a target reference information consistent with the document information, calculate the similarity value between the reference image corresponding to the target reference information and the corresponding feature image, and store the similarity value, the corresponding reference image, the corresponding feature image, and the calculation time of the similarity value in a preset historical record table in the order of the calculation time; The method further includes: At fixed intervals, if there is a third reference image in the historical record table, determine the minimum value among the similarity values corresponding to the three calculation times closest to the current time of the third reference image, update the reference image in the information repository that is consistent with the third reference image to the feature image corresponding to the minimum value, and reset the similarity threshold content of the reference image that is consistent with the third reference image to the preset initial value; wherein, the third reference image refers to a reference image where the difference between any two similarity values corresponding to the calculation times closest to the current time in the last three times is less than a preset difference.

3. The biometric database update method according to claim 1, wherein: The method further includes: At preset intervals, calculate the average value of all similarity values corresponding to each reference image in the historical record table within the preset interval, and replace the initial value of the similarity threshold of the corresponding reference image with the average value content.

4. The biometric database updating method according to claim 3, wherein: Updating the similarity threshold content of all the second reference images to the corresponding second similarity value includes: Updating the similarity threshold content of all the second reference images to the corresponding second similarity value, adding all the second reference images to a preset adjustment record table, and storing the current time as the addition time in the adjustment record table; The method further includes: Based on the addition time, at every specified duration, determining the number of times each reference image in the adjustment record table appears within the specified duration; According to a preset correspondence table, determining the specific time value of the preset duration corresponding to each reference image in the adjustment record table, and adding the specific time value of the corresponding preset duration to the corresponding reference image in the historical record table; wherein, the correspondence table is used to store the correspondence relationship between the number of times a reference image appears and the specific time value of the preset duration, and the more times it appears, the shorter the specific time value of the preset duration; At every preset duration, calculating the average value of all the similarity values corresponding to each reference image in the historical record table within the preset duration includes: Based on the correspondence table, determining the specific time value of the preset duration corresponding to each reference image in the historical record table; At every corresponding preset duration time value, calculating the average value of all the similarities corresponding to each reference image in the historical record table within the corresponding preset duration time value.

5. The biometric database update method according to claim 1, wherein: Updating each reference image corresponding to the target reference information to the corresponding feature image, and resetting the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value includes: Generating and displaying verification information based on the recognition information corresponding to the target reference information, where the verification information includes the recognition information; Receiving the confirmation information sent by the user through the terminal. If there is modified document information in the confirmation information, storing the modified document information in the information storage repository, storing each feature image in the verification information corresponding to the confirmation information as a new reference image in the information storage repository, and setting a similarity threshold for each new reference image according to a preset rule; If there is no modified document information in the confirmation information, updating each reference image corresponding to the target reference information to the corresponding feature image, and resetting the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value.

6. The biometric database update method according to claim 5, wherein: The confirmation information includes all the feature images in the recognition information corresponding to the confirmation information; If there is modified document information in the confirmation information, storing the modified document information in the information storage repository, storing each feature image in the verification information corresponding to the confirmation information as a new reference image in the information storage repository, and setting a similarity threshold for each new reference image according to a preset rule includes: If there is modified document information in the confirmation information, comparing the modified document information with each reference information in the information storage repository, If the modified document information does not exist in the information repository, store the modified document information in the information repository, store each feature image in the verification information corresponding to the confirmation information as a newly added reference image in the information repository, and set a similarity threshold for each newly added reference image according to a preset rule; If the modified document information exists in the information repository, generate new identification information based on the modified document information and all feature images in the corresponding confirmation information; The obtaining of the identification information includes: Obtain the identification information sent by the user through the terminal or obtain the automatically generated new identification information.

7. The biometric database update method according to claim 2, wherein: The method further includes: At each specified period, determine whether there is a fourth reference image or a fifth reference image, where the fourth reference image exists in the information repository and does not exist in the historical record table; the fifth reference image exists in the historical record table, and the difference between the latest calculation time corresponding to the fifth reference image in the historical record table and the current time is greater than a preset time difference; If there is a fourth reference image or a fifth reference image, delete the fourth reference image or the fifth reference image from the information storage table.

8. A biometric database update device, characterized in that: Including: An identification information acquisition module (1) for the user to obtain identification information, where the identification information includes at least document information and several feature images for reflecting different biometric features of the user; A document information verification module (2) for determining whether the document information exists in a preset information repository. If not, add the identification information to the information repository and set a similarity threshold for each reference image according to a preset rule; Wherein the information repository stores reference information corresponding to the document information; A feature image verification module (3) for calculating the similarity value between the reference image corresponding to the target reference information and the corresponding feature image when there is target reference information consistent with the document information; Wherein each reference information in the information repository corresponds to several reference images, and the reference images and the feature images are in one-to-one correspondence. Each reference image is preset with a similarity threshold, and each similarity threshold corresponds to an initial value; A similarity threshold adjustment module (4) for updating the similarity threshold content of all the second reference images to the corresponding second similarity value when the first similarity value between the first feature image and the corresponding first reference image is greater than the similarity threshold of the first reference image and the second similarity value between the second feature image and the corresponding second reference image is less than the similarity threshold of the second reference image; A feature image update module (5) for updating each reference image corresponding to the target reference information to the corresponding feature image and resetting the similarity threshold content of each reference image corresponding to the target reference information to the corresponding initial value when the similarity value between each reference image corresponding to the target reference information and the corresponding feature image is less than the similarity threshold of the corresponding reference image.

9. A biometric database update device, characterized in that: It includes a memory and a processor, and a computer program capable of being loaded and executed by the processor, such as any one of the methods recited in claims 1 to 7, is stored on the memory.

10. A computer-readable storage medium, characterized in that, A computer program capable of being loaded and executed by the processor, such as any one of the methods recited in claims 1 to 7, is stored.

Images