Heterogeneous portrait archive data secondary aggregation method, system, device and storage medium

Abstract

The application discloses a heterogeneous portrait archive data secondary aggregation method, system, device and storage medium, which comprises the following steps: acquiring confidence portrait archive data of a first algorithm manufacturer; performing secondary confidence identity verification on each archive cover photo in the confidence portrait archive data, and judging whether the secondary confidence identity verification result is consistent with the original confidence identity; creating a basic confidence archive library according to a confidence person, and adding field information; performing quality detection on portrait tracks in the confidence portrait archive data, and finding non-track data of the confidence person and cleaning the non-track data to an exception library; acquiring confidence portrait archive data of a next algorithm manufacturer; querying whether an archive has been built through an ID number; cyclically reading the next confidence portrait archive data; calling an archive basic relationship analysis model, acquiring accessory information, and refreshing confidence archive associated basic relationship data; and completing confidence archive data merging; and the scheme has the characteristics of being more complete, more accurate and more detailed.

Description

Technical Field

[0001] This invention relates to the field of facial image archive data aggregation, and particularly to a method, system, device, and storage medium for secondary aggregation of heterogeneous facial image archive data. Background Technology

[0002] Currently, there are relatively few technical solutions for secondary aggregation processing of facial profile data provided by multiple facial recognition algorithm vendors. According to our understanding, the existing technologies on the market mainly include two types:

[0003] One approach involves using third-party facial clustering algorithms for secondary structuring and clustering analysis of the entire dataset. However, this approach has several drawbacks: the cost of purchasing third-party algorithms increases; ensuring higher accuracy in aggregation using these algorithms presents technical risks; secondary structuring and clustering of the entire facial image data is required, significantly increasing GPU computing costs; and there is a lack of quality control checks on archival data. Another approach uses ID card numbers for secondary aggregation of credible archives. This approach also presents several problems: analyzing credible archives using ID card numbers requires merging and associating multiple heterogeneous archives of the same person; it simply associates heterogeneous archive data without performing duplicate or quality checks; and it is only suitable for aggregated analysis of credible archives, as uncredible archives lack ID card numbers, making this approach unsuitable for aggregating uncredible archive data.

[0004] Therefore, it is crucial to achieve secondary aggregation and association of multiple sets of facial image data in the same or different target areas with minimal construction costs; to solve problems such as the coexistence of multiple sets of facial image data, inconsistent data from multiple sources, and poor quality that arise during the construction process in various regions; and to provide a more comprehensive, accurate, and detailed facial image data resource for upper-level practical application systems. Summary of the Invention

[0005] To achieve the above objectives, the inventors provide a method for secondary aggregation of heterogeneous facial image archive data, including acquiring confident facial image archive data and performing secondary aggregation processing, comprising the following steps:

[0006] S101: Obtain the first trusted portrait profile data from an algorithm vendor;

[0007] S102: Perform secondary credible identity verification on each cover photo in the credible facial image archive data, and determine whether the secondary credible identity verification result is consistent with the original credible identity. If they are consistent, continue; if they are inconsistent, put the cover photo into the archive to be confirmed.

[0008] S103: Create a basic confidence profile database by the person making the confidence claim and add new field information;

[0009] S104: Perform quality inspection on the facial trajectory in the confident facial profile data, and find and clean up non-trajectory data into the anomaly database;

[0010] S105: Obtain the confidence profile data of the next algorithm vendor and execute S102;

[0011] S106: Check whether a file has been created by ID card number. If not, proceed with S103 and S104.

[0012] S107: If a file with the same name already exists, execute S104 and perform file data comparison and merging.

[0013] S108: Loop through and read the next confidence portrait profile data, and start execution from S106;

[0014] S109: Call the basic relationship analysis model of archives, obtain the auxiliary information, and refresh the basic relationship data of the archives with confidence;

[0015] S110: Complete the merging of confidence profile data.

[0016] As a preferred embodiment of the present invention, the method includes acquiring unconfidential portrait profile data and performing secondary aggregation processing, comprising the following steps:

[0017] S201: Obtain unconfidential portrait profile data from the first algorithm vendor;

[0018] S202: Verify the identity of each cover photo in the untrusted portrait archive data. If the verification is successful, proceed to S106; otherwise, continue.

[0019] S203: Initialize and create the basic unconfidential archive, adding new field information;

[0020] S204: Perform quality inspection on the facial trajectory in the unconfidential facial profile data, and find and clean up non-trajectory data into the anomaly database;

[0021] S205: Obtain unconfidential portrait profile data from the next algorithm vendor;

[0022] S206: Read the first unconfidential portrait file data in sequence and call step S204 to perform quality inspection processing;

[0023] S207: Determine whether an archive has been created by comparing the archive cover photo with the full archive cover photo in a 1:N ratio.

[0024] S208: If no file has been created, then execute S203 and S204;

[0025] S209: If files have already been created, begin the process of comparing and merging the file data;

[0026] S210: Read the next unconfidential portrait profile data in a loop and start execution from S206;

[0027] S211: Call the basic relationship analysis model of archives, obtain the auxiliary information, and refresh the basic relationship data of unconfidential archives;

[0028] S212, complete the merging process of unconfidential archive data.

[0029] As a preferred embodiment of the present invention, the newly added field information includes: new file type, file manufacturer ID, file merging time and / or a unique video identity ID code; the ancillary information includes peer relationships, frequently visited places, accommodation locations and / or travel patterns.

[0030] As a preferred embodiment of the present invention, the archive data comparison and merging process includes the following steps:

[0031] S301: Obtain the source file records and associated human image trajectory data set;

[0032] S302: Obtain the target file records and associated facial trajectory data set;

[0033] S303: Loop through and read the trajectory data of each target file;

[0034] S304: Perform duplicate detection by comparing the image ID, capture time, device ID, and image URL with the source file trajectory;

[0035] S305: If the trajectory data already exists, only add the manufacturer ID and add it to the duplicate trajectory library;

[0036] S306: If it is not repeated, insert a new trajectory record into the trajectory table associated with the source file;

[0037] S307: Complete the comparison and merging of source and target file data, and return.

[0038] To achieve the above objectives, the inventors also provide a secondary aggregation system for heterogeneous facial image archive data, comprising:

[0039] The data acquisition module is used to acquire reliable portrait data from the first manufacturer.

[0040] The Confidential Identity Verification Module is used to perform secondary Confidential Identity Verification on each cover photo in the Confidential Image Archive Data, and to determine whether the secondary Confidential Identity Verification result is consistent with the original Confidential Identity. If they are consistent, the process continues; if they are inconsistent, the cover photo is placed in the pending confirmation archive.

[0041] The confidence profile processing module is used to create a basic confidence profile database based on the person making the confidence request.

[0042] The video identity ID management module is used to add new field information;

[0043] The document comparison and merging module is used for...

[0044] Perform quality inspection on the facial trajectory in the trusted facial profile data, and find and clean up non-trajectory data into the anomaly database;

[0045] Obtain the confidence profile data of the next algorithm vendor;

[0046] If an ID card number is used to check whether a file has been created, and if not, the process returns to the trusted file processing module for reprocessing. If a file with the same name already exists, the quality of the facial trajectory in the trusted facial image file data is checked, and non-trajectory data is cleaned up and added to the abnormal database before the file data is compared and merged.

[0047] The system continuously reads the next piece of trusted facial image data and checks whether a file has been created by the ID card number, and then makes a judgment and processing accordingly.

[0048] The basic relationship analysis module is used to call the basic relationship analysis model of archives, obtain supplementary information, and refresh the basic relationship data of the archives with confidence.

[0049] The data storage module is used to store the data after the confidence archive data has been merged and processed.

[0050] As a preferred embodiment of the present invention, the first acquisition module further includes a tool for acquiring unreliable portrait profile data from the first manufacturer;

[0051] The credible identity verification module also includes a function to verify the credible identity of each cover photo in the uncredible portrait archive data. If the verification is successful, it will jump to the archive comparison and merging module to check whether the ID number has been created; otherwise, it will continue.

[0052] It also includes an unconfidential archive processing module, used to initialize and create a basic unconfidential archive database;

[0053] The video identity ID management module is used to add new field information;

[0054] It also includes a face database comparison module, which is used to compare untrusted facial images with the face database and obtain identity information;

[0055] The file comparison and merging module is also used for,

[0056] Perform quality inspection on the facial trajectory in the unconfidential facial profile data, and find and clean up non-trajectory data into the anomaly database;

[0057] Acquire unconfidential portrait profile data from the next algorithm vendor;

[0058] Read the first unreliable portrait file data in sequence and perform quality inspection processing;

[0059] By comparing the cover photo of the archive with the cover photo of the entire database in a 1:N ratio, it is determined whether the archive has been created. If it has not been created, the process returns to the unreliable archive processing module for reprocessing. If it has been created, the archive data comparison and merging process begins.

[0060] The system continuously reads the next untrusted portrait file data and calls the untrusted file processing module, video identity ID management module, and file comparison and merging module for continuous processing.

[0061] The basic relationship analysis module is also used to call the basic relationship analysis model of archives, obtain auxiliary information, and refresh the basic relationship data of unconfidential archives;

[0062] The data storage module is also used to complete the data storage after the non-confidence archive data merging process.

[0063] As a preferred embodiment of the present invention, the newly added field information includes: new file type, file manufacturer ID, file merging time and / or a unique video identity ID code; the ancillary information includes peer relationships, frequently visited places, accommodation locations and / or travel patterns.

[0064] To achieve the above objectives, the inventors also provide a device for secondary aggregation of heterogeneous human portrait archive data, comprising: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a circuit;

[0065] The at least one processor invokes the instructions in the memory to cause the heterogeneous portrait data secondary aggregation device to execute the method described in any one of the above-described inventions.

[0066] To achieve the above objectives, the inventors also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any one of the above inventions.

[0067] Unlike existing technologies, the above technical solution achieves the following beneficial effects:

[0068] (1) This solution does not require the introduction of new third-party face structuring and face clustering algorithms, and the construction cost is much lower than that of existing approximate solutions, which can effectively reduce construction costs;

[0069] (2) This solution is compatible with the access and secondary aggregation of both credible and uncredible archival data;

[0070] (3) This solution is applicable not only to the fusion of multiple sets of heterogeneous archive data under the same capture data in the same area, but also to the fusion of multiple sets of heterogeneous archive data under different capture data in different areas.

[0071] (4) In the secondary aggregation process of archives, the scheme adds quality inspection and calibration management of archive trajectory data to ensure that every trajectory data in the archives is accurate and non-repeating;

[0072] (5) This solution provides rich basic relationship attributes for the merged archive data through big data model analysis, including analysis of the people who accompany the archive target, analysis of frequently visited places and analysis of occurrence patterns, etc., providing richer human image archive data resources for upper-level video investigation operations. Attached Figure Description

[0073] Figure 1 The flowchart of the secondary aggregation method for heterogeneous portrait archive data described in the specific implementation method Figure 1 ;

[0074] Figure 2 The flowchart of the secondary aggregation method for heterogeneous portrait archive data described in the specific implementation method Figure 2 ;

[0075] Figure 3 This is a flowchart illustrating the data comparison and merging process described in the specific implementation method.

[0076] Figure 4 This is a framework diagram of the heterogeneous human image archive data secondary aggregation system described in a specific implementation method;

[0077] Figure 5 This is a schematic diagram of the structure of the heterogeneous human image archive data secondary aggregation device described in a specific implementation.

[0078] Explanation of reference numerals in the attached figures:

[0079] 401. Processor; 402. Power supply; 403. Wired or wireless network interface; 404. Input / output interface; 405. Operating system; 406. Data; 407. Application program; 408. Storage medium; 409. Memory. Detailed Implementation

[0080] To explain in detail the technical content, structural features, objectives, and effects of the technical solution, the following description is provided in conjunction with specific embodiments and accompanying drawings. Example 1

[0081] like Figure 1 As shown, this embodiment provides a method for secondary aggregation of heterogeneous facial profile data, including obtaining confident facial profile data and performing secondary aggregation processing, including the following steps:

[0082] S101: Obtain the first trusted portrait profile data from an algorithm vendor;

[0083] S102: Perform secondary credible identity verification on each cover photo in the credible facial image archive data, and determine whether the TOP1 result of the secondary credible identity verification is consistent with the original credible identity. If they are consistent, continue; if they are inconsistent, put the cover photo into the archive to be confirmed.

[0084] S103: Create a basic confidence profile database by the person making the confidence claim and add new field information;

[0085] S104: Perform quality inspection on the facial trajectory in the confident facial profile data, and find and clean up non-trajectory data into the anomaly database;

[0086] S105: Obtain the confidence profile data of the next algorithm vendor and execute S102;

[0087] S106: Check whether a file has been created by ID card number. If not, proceed with S103 and S104.

[0088] S107: If a file with the same name already exists, execute S104 and perform file data comparison and merging.

[0089] S108: Loop through and read the next confidence portrait profile data, and start execution from S106;

[0090] S109: Call the basic relationship analysis model of archives, obtain the auxiliary information, and refresh the basic relationship data of the archives with confidence;

[0091] S110: Complete the merging of confidence profile data.

[0092] like Figure 2 As shown, in some embodiments, the method further includes acquiring unconfidential portrait profile data and performing secondary aggregation processing, including the following steps:

[0093] S201: Obtain unconfidential portrait profile data from the first algorithm vendor;

[0094] S202: Verify the identity of each cover photo in the untrusted portrait archive data. If the verification is successful, proceed to S106; otherwise, continue.

[0095] S203: Initialize and create the basic unconfidential archive, adding new field information;

[0096] S204: Perform quality inspection on the facial trajectory in the unconfidential facial profile data, and find and clean up non-trajectory data into the anomaly database;

[0097] S205: Obtain unconfidential portrait profile data from the next algorithm vendor;

[0098] S206: Read the first unconfidential portrait file data in sequence and call step S204 to perform quality inspection processing;

[0099] S207: Determine whether an archive has been created by comparing the archive cover photo with the full archive cover photo in a 1:N ratio.

[0100] S208: If no file has been created, then execute S203 and S204;

[0101] S209: If files have already been created, begin the process of comparing and merging the file data;

[0102] S210: Read the next unconfidential portrait profile data in a loop and start execution from S206;

[0103] S211: Call the basic relationship analysis model of archives, obtain the auxiliary information, and refresh the basic relationship data of unconfidential archives;

[0104] S212, complete the merging process of unconfidential archive data.

[0105] In the above embodiments, the newly added field information includes: new file type, file vendor ID, file merging time and / or a unique video identity ID code; the supplementary information includes peer relationships, frequently visited places, place of stay and / or travel patterns.

[0106] In the above embodiments, such as Figure 3 As shown, the data comparison and merging process includes the following steps:

[0107] S301: Obtain the source file records and associated human image trajectory data set;

[0108] S302: Obtain the target file records and associated facial trajectory data set;

[0109] S303: Loop through and read the trajectory data of each target file;

[0110] S304: Perform duplicate detection by comparing the image ID, capture time, device ID, and image URL with the source file trajectory;

[0111] S305: If the trajectory data already exists, only add the manufacturer ID and add it to the duplicate trajectory library;

[0112] S306: If it is not repeated, insert a new trajectory record into the trajectory table associated with the source file;

[0113] S307: Complete the comparison and merging of source and target file data, and return. Example 2

[0114] like Figure 4 As shown, this embodiment also provides a secondary aggregation system for heterogeneous facial profile data, including:

[0115] The data acquisition module is used to acquire reliable portrait data from the first manufacturer.

[0116] The Confidential Identity Verification Module is used to perform secondary Confidential Identity Verification on each cover photo in the Confidential Image Archive Data, and to determine whether the secondary Confidential Identity Verification result is consistent with the original Confidential Identity. If they are consistent, the process continues; if they are inconsistent, the cover photo is placed in the pending confirmation archive.

[0117] The confidence profile processing module is used to create a basic confidence profile database based on the person making the confidence request.

[0118] The video identity ID management module is used to add new field information;

[0119] The document comparison and merging module is used for...

[0120] Perform quality inspection on the facial trajectory in the trusted facial profile data, and find and clean up non-trajectory data into the anomaly database;

[0121] Obtain the confidence profile data of the next algorithm vendor;

[0122] If an ID card number is used to check whether a file has been created, and if not, the process returns to the trusted file processing module for reprocessing. If a file with the same name already exists, the quality of the facial trajectory in the trusted facial image file data is checked, and non-trajectory data is cleaned up and added to the abnormal database before the file data is compared and merged.

[0123] The system continuously reads the next piece of trusted facial image data and checks whether a file has been created by the ID card number, and then makes a judgment and processing accordingly.

[0124] The basic relationship analysis module is used to call the basic relationship analysis model of archives, obtain supplementary information, and refresh the basic relationship data of the archives with confidence.

[0125] The data storage module is used to store the data after the confidence archive data has been merged and processed.

[0126] In the above embodiments, the first acquisition module further includes a tool for acquiring unconfidential portrait profile data from the first manufacturer;

[0127] The credible identity verification module also includes a function to verify the credible identity of each cover photo in the uncredible portrait archive data. If the verification is successful, it will jump to the archive comparison and merging module to check whether the ID number has been created; otherwise, it will continue.

[0128] It also includes an unconfidential archive processing module, used to initialize and create a basic unconfidential archive database;

[0129] The video identity ID management module is used to add new field information;

[0130] It also includes a face database comparison module, which is used to compare untrusted facial images with the face database and obtain identity information;

[0131] The file comparison and merging module is also used for,

[0132] Perform quality inspection on the facial trajectory in the unconfidential facial profile data, and find and clean up non-trajectory data into the anomaly database;

[0133] Acquire unconfidential portrait profile data from the next algorithm vendor;

[0134] Read the first unreliable portrait file data in sequence and perform quality inspection processing;

[0135] By comparing the cover photo of the archive with the cover photo of the entire database in a 1:N ratio, it is determined whether the archive has been created. If it has not been created, the process returns to the unreliable archive processing module for reprocessing. If it has been created, the archive data comparison and merging process begins.

[0136] The system continuously reads the next untrusted portrait file data and calls the untrusted file processing module, video identity ID management module, and file comparison and merging module for continuous processing.

[0137] The basic relationship analysis module is also used to call the basic relationship analysis model of archives, obtain auxiliary information, and refresh the basic relationship data of unconfidential archives;

[0138] The data storage module is also used to complete the data storage after the non-confidence archive data merging process.

[0139] In the above embodiments, the newly added field information includes: new file type, file vendor ID, file merging time and / or a unique video identity ID code; the ancillary information includes peer relationships, frequently visited places, accommodation locations and / or travel patterns. Example 3

[0140] like Figure 5 As shown, this embodiment also provides a secondary aggregation device for heterogeneous human portrait archive data, including: a memory and at least one processor 401, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a line;

[0141] The at least one processor invokes the instructions in the memory to cause the heterogeneous portrait data secondary aggregation device to execute the method described in any of the above embodiments.

[0142] The secondary aggregation device for heterogeneous facial profile data can vary significantly due to differences in configuration or performance. It may include one or more central processing units (CPUs) and memory 409, and one or more storage media 408 for storing applications 407 or data 406. The memory and storage media can be temporary or persistent storage. The program stored on the storage media may include one or more modules (not shown in the diagram), each module including a series of instruction operations on a single-page application-optimized device. Furthermore, the processor may be configured to communicate with the storage media and execute the series of instruction operations on the storage media on a single-page application-optimized device.

[0143] The device for secondary aggregation of heterogeneous facial image data may also include one or more power supplies 402, one or more wired or wireless network interfaces 403, one or more input / output interfaces 404, and / or one or more operating systems 405, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 5 The structure of the heterogeneous portrait data secondary aggregation device shown in the figure does not constitute a limitation on the heterogeneous portrait data secondary aggregation device. It may include more or fewer components than shown, or combine certain components, or have different component arrangements. Example 4

[0144] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0145] It should be noted that although the above embodiments have been described herein, this does not limit the scope of patent protection of the present invention. Therefore, any changes and modifications made to the embodiments described herein based on the innovative concept of the present invention, or equivalent structural or procedural transformations made using the content of the present invention's specification and drawings, directly or indirectly applying the above technical solutions to other related technical fields, are all included within the scope of patent protection of the present invention.

Claims

1. A method for secondary aggregation of heterogeneous facial image archive data, characterized in that, This includes acquiring reliable facial profile data and performing secondary aggregation processing, including the following steps: S101: Obtain the first trusted portrait profile data from an algorithm vendor; S102: Perform secondary credible identity verification on each cover photo in the credible facial image archive data, and determine whether the secondary credible identity verification result is consistent with the original credible identity. If they are consistent, continue; if they are inconsistent, put the cover photo into the archive to be confirmed. S103: Create a basic confidence profile database by the person making the confidence claim and add new field information; S104: Perform quality inspection on the facial trajectory in the confident facial profile data, and find and clean up non-trajectory data into the anomaly database; S105: Obtain the confidence profile data of the next algorithm vendor and execute S102; S106: Check whether a file has been created by ID card number. If not, proceed with S103 and S104. S107: If a file with the same name already exists, execute S104 and perform file data comparison and merging. S108: Loop through and read the next confidence portrait profile data, and start execution from S106; S109: Call the basic relationship analysis model of archives, obtain the auxiliary information, and refresh the basic relationship data of the archives with confidence; S110: Complete the merging of confidence profile data; The newly added fields include: new file type, file vendor ID, file merging time and / or a unique video identity ID code; the supplementary information includes peer relationships, frequently visited places, accommodations and / or travel patterns.

2. The method for secondary aggregation of heterogeneous facial image archive data according to claim 1, characterized in that, This includes acquiring unconfidential portrait profile data and performing secondary aggregation processing, including the following steps: S201: Obtain unconfidential portrait profile data from the first algorithm vendor; S202: Verify the identity of each cover photo in the untrusted portrait archive data. If the verification is successful, proceed to S106; otherwise, continue. S203: Initialize and create the basic unconfidential archive, adding new field information; S204: Perform quality inspection on the facial trajectory in the unconfidential facial profile data, and find and clean up non-trajectory data into the anomaly database; S205: Obtain unconfidential portrait profile data from the next algorithm vendor; S206: Read the first unconfidential portrait file data in sequence and call step S204 to perform quality inspection processing; S207: Determine whether an archive has been created by comparing the archive cover photo with the full archive cover photo in a 1:N ratio. S208: If no file has been created, then execute S203 and S204; S209: If files have already been created, begin the process of comparing and merging the file data; S210: Read the next unconfidential portrait profile data in a loop and start execution from S206; S211: Call the basic relationship analysis model of archives, obtain the auxiliary information, and refresh the basic relationship data of unconfidential archives; S212, complete the merging process of unconfidential archive data.

3. The method for secondary aggregation of heterogeneous facial image archive data according to claim 1, characterized in that, The data comparison and merging process includes the following steps: S301: Obtain the source file records and associated human image trajectory data set; S302: Obtain the target file records and associated facial trajectory data set; S303: Loop through and read the trajectory data of each target file; S304: Perform duplicate detection by comparing the image ID, capture time, device ID, and image URL with the source file trajectory; S305: If the trajectory data already exists, only add the manufacturer ID and add it to the duplicate trajectory library; S306: If it is not repeated, insert a new trajectory record into the trajectory table associated with the source file; S307: Complete the comparison and merging of source and target file data, and return.

4. A secondary aggregation system for heterogeneous facial image archive data, characterized in that, include: The data acquisition module is used to acquire reliable portrait data from the first manufacturer. The Confidential Identity Verification Module is used to perform secondary Confidential Identity Verification on each cover photo in the Confidential Image Archive Data, and to determine whether the secondary Confidential Identity Verification result is consistent with the original Confidential Identity. If they are consistent, the process continues; if they are inconsistent, the cover photo is placed in the pending confirmation archive. The confidence profile processing module is used to create a basic confidence profile database based on the person making the confidence request. The video identity ID management module is used to add new field information; The document comparison and merging module is used for... Perform quality inspection on the facial trajectory in the trusted facial profile data, and find and clean up non-trajectory data into the anomaly database; Obtain the confidence profile data of the next algorithm vendor; If an ID card number is used to check whether a file has been created, and if not, the process returns to the trusted file processing module for reprocessing. If a file with the same name already exists, the quality of the facial trajectory in the trusted facial image file data is checked, and non-trajectory data is cleaned up and added to the abnormal database before the file data is compared and merged. The system continuously reads the next piece of trusted facial image data and checks whether a file has been created by the ID card number, and then makes a judgment and processing accordingly. The basic relationship analysis module is used to call the basic relationship analysis model of archives, obtain supplementary information, and refresh the basic relationship data of the archives with confidence. The data storage module is used to store the data after the confidence archive data has been merged and processed. The newly added fields include: new file type, file vendor ID, file merging time and / or a unique video identity ID code; The supplementary information includes peer relationships, frequently visited locations, accommodations, and / or travel patterns.

5. The heterogeneous facial image data secondary aggregation system according to claim 4, characterized in that: The archive data acquisition module also includes a tool for acquiring unreliable portrait archive data from the first manufacturer; The credible identity verification module also includes a function to verify the credible identity of each cover photo in the uncredible portrait archive data. If the verification is successful, it will jump to the archive comparison and merging module to check whether the ID number has been created; otherwise, it will continue. It also includes an unconfidential archive processing module, used to initialize and create a basic unconfidential archive database; The video identity ID management module is used to add new field information; It also includes a face database comparison module, which is used to compare untrusted facial images with the face database and obtain identity information; The file comparison and merging module is also used for, Perform quality inspection on the facial trajectory in the unconfidential facial profile data, and find and clean up non-trajectory data into the anomaly database; Acquire unconfidential portrait profile data from the next algorithm vendor; Read the first unreliable portrait file data in sequence and perform quality inspection processing; By comparing the cover photo of the archive with the cover photos of the entire database in a 1:N ratio, it can be determined whether an archive has been created. If no file has been created, the process will revert to the unreliable file processing module for reprocessing. If a file has been created, the process of comparing and merging file data will begin. The system continuously reads the next untrusted portrait file data and calls the untrusted file processing module, video identity ID management module, and file comparison and merging module for continuous processing. The basic relationship analysis module is also used to call the basic relationship analysis model of archives, obtain auxiliary information, and refresh the basic relationship data of unconfidential archives; The data storage module is also used to complete the data storage after the non-confidence archive data merging process.

6. A device for secondary aggregation of heterogeneous facial image archive data, characterized in that, include: A memory and at least one processor, wherein the memory stores instructions and the memory and the at least one processor are interconnected via a circuit; The at least one processor invokes the instructions in the memory to cause the heterogeneous human portrait data secondary aggregation device to perform the method described in any one of claims 1-3.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the method described in any one of claims 1-3.

Images