Method and system for identifying impersonation of user information

Abstract

The application discloses a kind of impersonation user information identification method and system, it is related to information identification technical field.The method comprises the following steps: obtaining suspect user information, then pre-processing suspect user information, obtaining suspect user data, further matching all the user data to be protected in the preset protection library with suspect user data, descending order sorting is carried out to all the user data to be protected according to the matching result, and the user data to be protected sorting data is obtained, finally, based on the preset recall strategy and the user data to be protected sorting data, the user data to be protected is filtered, the user data to be protected that meets the preset requirement is recorded as filter user data, and the user information corresponding to the filter user data is output as impersonated user information, effectively solve the technical problem that the identification efficiency is low when the existing technology identifies suspect user information (impersonate nickname and / or avatar).

Description

Technical Field

[0001] This invention relates to the field of information recognition technology, and to a method and system for identifying impersonated user information, particularly to a method and system for identifying impersonated nicknames and / or avatars. Background Technology

[0002] When using social media apps, users have unique profile pictures and nicknames, which serve as crucial evidence for identifying users. However, this also leads to instances of impersonation using users' profile pictures and / or nicknames to commit fraud. Perpetrators use the profile pictures and / or nicknames of users they are trying to impersonate to commit fraud. In this scenario, calculating the similarity between the profile picture and / or nickname can effectively identify the impersonating nickname and / or profile picture, thereby detecting fraudulent activities.

[0003] Existing media data, including images, audio, text, and video, can all be similarized after characterization, allowing for intuitive quantification of the similarity between media data. However, different quantification methods will produce different results. In practical anti-impersonation fraud scenarios, matching searches are typically used to determine if a suspected nickname is fraudulent, and the suspected profile picture is manually compared with images in a protected database to determine if it is an imposter. While these methods can solve some cases of complete impersonation of profile pictures and / or nicknames, they suffer from low identification efficiency, hindering the timely detection of impersonation / fraud. Summary of the Invention

[0004] This invention provides a method and system for identifying impersonated user information, which solves the technical problem of low identification efficiency in the prior art when identifying suspected user information (impersonating nicknames and / or avatars).

[0005] This invention provides a method for identifying impersonated user information, comprising:

[0006] S1. Obtain information about suspected users;

[0007] S2. Preprocess the suspected user information to obtain suspected user data;

[0008] S3. Based on a preset matching algorithm, the suspected user data is matched with all the user data to be protected in the preset protection database. The user data to be protected is sorted according to the matching results to obtain the sorted user data to be protected.

[0009] S4. Based on the preset recall strategy and the sorting data of the users to be protected, filter the data of the users to be protected, record the data of the users to be protected that meets the preset requirements as the filtered user data, and output the information of the users to be protected corresponding to the filtered user data as the information of the impersonated users.

[0010] Preferred,

[0011] The suspected user information includes the suspected nickname and the suspected profile picture; the suspected nickname and the suspected profile picture are matched.

[0012] The user data to be protected includes nickname data to be protected and profile picture data to be protected; the nickname data to be protected and the profile picture data to be protected are matched.

[0013] The preset matching algorithm is a preset nickname matching algorithm or a preset avatar matching algorithm;

[0014] The preset recall strategy is either a preset nickname recall strategy, a preset avatar recall strategy, a preset nickname and avatar combined recall strategy, or a preset avatar and nickname combined recall strategy.

[0015] The preset requirements are preset nickname requirements, preset avatar requirements, preset nickname and avatar combined requirements, or preset avatar and nickname combined requirements.

[0016] The filtered user data can be nickname data, avatar data, a combination of nickname and avatar data, or a combination of avatar and nickname data.

[0017] Preferably, step S2 includes:

[0018] S2A: Perform text preprocessing on the suspected nickname to obtain suspected nickname data.

[0019] Preferably, step S2 further includes:

[0020] S2B: Perform image preprocessing on the suspect's headshot to obtain suspect headshot data.

[0021] Preferably, step S3 includes:

[0022] S3A. Based on the preset nickname matching algorithm, the suspected nickname data is matched with all the nickname data to be protected in the preset protection database. The nickname data to be protected is sorted according to the nickname matching results to obtain the nickname data to be protected sorted data.

[0023] Preferably, step S3 includes:

[0024] S3B. Based on the preset avatar matching algorithm, the suspected avatar data is matched with all the avatar data to be protected in the preset protection database. The avatar data to be protected is sorted according to the avatar matching results to obtain the sorted avatar data to be protected.

[0025] Preferably, step S3 includes:

[0026] S3C. Based on the preset nickname matching algorithm, the suspected nickname data is matched with all the nickname data to be protected in the preset protection database. The nickname data to be protected is sorted according to the nickname matching results to obtain the nickname data to be protected sorted data.

[0027] Based on the preset nickname recall strategy and the nickname ranking data to be protected, the nickname data to be protected is filtered, and the nickname data to be protected that meets the preset nickname requirements is recorded as the filtered nickname data, and the avatar data to be protected that matches the filtered nickname data is recorded as the first avatar data to be protected.

[0028] Based on the preset avatar matching algorithm, the suspected avatar data is matched with the first avatar data to be protected, and all the first avatar data to be protected are sorted according to the first avatar matching result to obtain the first avatar data to be protected sorted data.

[0029] Preferably, step S3 includes:

[0030] S3D: Based on the preset avatar matching algorithm, the suspected avatar data is matched with all the avatar data to be protected in the preset protection library. According to the avatar matching results, all the avatar data to be protected are sorted in descending order to obtain the avatar ranking data to be protected.

[0031] Based on the preset avatar recall strategy and the avatar sorting data to be protected, the avatar data to be protected is filtered, and the avatar data to be protected that meets the preset avatar requirements is recorded as the filtered avatar data, and the nickname data to be protected that matches the filtered avatar data is recorded as the first nickname data to be protected.

[0032] Based on the preset nickname matching algorithm, the suspected nickname data is matched with the first nickname data to be protected, and all the first nickname data to be protected is sorted according to the first nickname matching result to obtain the first nickname sorting data to be protected.

[0033] Preferably, after step S3A, step S4 specifically involves:

[0034] S4A. Based on the preset nickname recall strategy and the nickname ranking data to be protected, filter the nickname data to be protected, record the nickname data to be protected that meets the preset nickname requirements as filtered nickname data, and output the nickname to be protected corresponding to the filtered nickname data as the impersonated nickname.

[0035] Preferably, after step S3B, step S4 specifically involves:

[0036] S4B. Based on the preset avatar recall strategy and the avatar sorting data to be protected, filter the avatar data to be protected, record the avatar data to be protected that meets the preset avatar requirements as filtered avatar data, and output the avatar to be protected corresponding to the filtered avatar data as the impersonated avatar.

[0037] Preferably, after step S3C, step S4 specifically involves:

[0038] S4C. Based on the preset nickname and avatar joint recall strategy and the first avatar ranking data to be protected, filter the first avatar data to be protected, record the first avatar data to be protected that meets the preset nickname and avatar joint requirements as the first filtered avatar data, output the avatar to be protected corresponding to the first filtered avatar data as the impersonated avatar, and output the nickname to be protected corresponding to the nickname data to be protected that matches the first filtered avatar data as the impersonated nickname.

[0039] Preferably, after step S3D, step S4 specifically includes:

[0040] S4D: Based on the preset avatar-nickname joint recall strategy and the first nickname to be protected sorting data, filter the first nickname to be protected data, record the first nickname to be protected that meets the preset avatar-nickname joint requirements as the first filtered nickname data, output the nickname to be protected corresponding to the first filtered nickname data as the impersonated nickname, and output the avatar to be protected corresponding to the avatar data to be protected that matches the first filtered nickname data as the impersonated avatar.

[0041] Preferably, step S2A specifically includes:

[0042] The suspected nicknames were normalized, converted to uppercase and lowercase letters, converted to simplified and traditional Chinese characters, converted to Martian language, and filtered for common words to obtain the suspected nickname data.

[0043] Alternatively, the suspected nickname can be vectorized based on a preset semantic representation model to obtain the suspected nickname data.

[0044] Preferably, step S2B specifically includes:

[0045] The suspect's headshot is scaled, flipped, binarized, and compressed to obtain suspect headshot data;

[0046] Alternatively, the suspect's image can be vectorized based on a preset image representation model to obtain suspect image data.

[0047] Preferably, after step S4, the method further includes:

[0048] The impersonated user information is stored as updated protected user information, and the preset matching algorithm is updated based on the updated protected user information.

[0049] Preferably, after step S4, the method further includes:

[0050] The impersonated user information and the corresponding suspected user information are packaged, stored, and displayed.

[0051] Preferably, after step S4, the method further includes:

[0052] Send a risk alert about being impersonated to the user whose information was being impersonated.

[0053] A system for identifying impersonated user information includes:

[0054] The suspect information acquisition module acquires information about suspected users.

[0055] The suspect information preprocessing module is used to preprocess the suspect user information to obtain suspect user data;

[0056] The information matching and sorting module is used to match the suspected user data with all the user data to be protected in the preset protection database based on a preset matching algorithm, and sort all the user data to be protected in descending order according to the matching results to obtain the user data to be protected sorting data.

[0057] The identification output module is used to filter the user data to be protected based on a preset recall strategy and the sorting data of the users to be protected, record the user data to be protected that meets the preset requirements as filtered user data, and output the user information to be protected corresponding to the filtered user data as the impersonated user information.

[0058] As can be seen from the above technical solutions, the present invention has the following advantages:

[0059] This application provides a method and system for identifying impersonated user information. The method includes: acquiring suspected user information, preprocessing the suspected user information to obtain suspected user data, further matching the suspected user data with all user data to be protected in a preset protection database, sorting all user data to be protected in descending order according to the matching results to obtain user data to be protected sorting, and finally filtering the user data to be protected based on a preset recall strategy and the user data to be protected sorting, recording the user data to be protected that meets preset requirements as filtered user data, and outputting the user information to be protected corresponding to the filtered user data as impersonated user information.

[0060] This application provides a method for identifying impersonated user information. First, the suspected user information is matched with all / specified user information to be protected in a pre-set protection database. Based on the matching results (similarity), the user information to be protected is sorted. Further, a pre-set recall strategy is used to filter the sorted user information to select user information to be protected that meets the preset requirements. By using a sorting-then-filtering approach to identify suspected user information, the identification efficiency can be effectively improved, and fraudulent activities can be detected at a faster speed. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 A flowchart illustrating a method for identifying impersonated user information provided in this application;

[0063] Figure 2 A schematic diagram of a system for identifying impersonated user information provided in this application. Detailed Implementation

[0064] This invention provides a method and system for identifying impersonated user information, which solves the technical problem of low identification efficiency in the prior art when identifying suspected user information (impersonating nicknames and / or avatars).

[0065] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0066] Embodiment 1 of this application provides a method for identifying impersonated user information. Please refer to [link to documentation]. Figure 1 In Example 1, the method includes:

[0067] S1. Obtain information about the suspected user.

[0068] When users use social media apps, they acquire corresponding user information, which serves as a crucial basis for verifying user identity. In this environment, criminals also impersonate users to commit fraud. To quickly identify fraudulent activities, by obtaining information about suspected users and comparing it with the information of users to be protected, it is possible to effectively identify both the impersonated users and the suspected users, thereby detecting fraudulent activities.

[0069] It should be noted that user information includes, but is not limited to, a unique avatar and nickname. In a preferred embodiment, the suspected user information includes a suspected nickname and / or a suspected avatar.

[0070] Taking a live streaming room as an example, the room may contain the streamer, room administrators, donors with a high degree of connection to the streamer, friends of donors with a high degree of connection to the donors, users with a high degree of connection to the streamer, and users with a low degree of connection to the streamer. Among these, donors with a high degree of connection to the streamer can be understood as those who send high-value gifts to the streamer within a preset time period; users with a high degree of connection to the streamer can be understood as those who interact with the streamer a preset number of times within a preset time period.

[0071] In this environment, perpetrators may impersonate streamers to defraud viewers by changing their profile pictures and / or nicknames, or impersonate live stream administrators to defraud users in the live stream, or impersonate viewers or users to defraud streamers, or impersonate friends of viewers to defraud viewers, etc.

[0072] Taking the above background as an example:

[0073] In one specific scenario, when a new user enters the live stream, the new user is immediately marked as a suspect user. Step S1 needs to be executed to obtain the new user's user information and record the new user's user information as suspect user information.

[0074] In another specific scenario, when a user modifies their user information in the live stream, that user is marked as a suspect. Step S1 needs to be executed to obtain the modified user information and record it as a suspect user.

[0075] In some cases, perpetrators may impersonate a user's nickname to commit fraud, or impersonate a user's avatar to commit fraud, or impersonate a user's nickname or avatar to commit fraud. Therefore, in step S1, obtaining the information of the suspected user may specifically include obtaining the suspected user's suspected nickname and / or suspected avatar.

[0076] S2. Preprocess the suspected user information to obtain suspected user data.

[0077] To evade tracking, perpetrators typically do not use user information that is exactly the same as the user to be protected. Instead, they modify the user information or use similar information as the suspect user information. Based on this, step S2 preprocesses the suspect user information to obtain suspect user data, which specifically includes suspect nickname data and / or suspect avatar data.

[0078] Specifically, when the suspected user information includes a suspected nickname, step S2 obtains the suspected user data including:

[0079] The suspected nicknames are subjected to number normalization, letter case conversion, simplified / traditional Chinese character conversion, Martian language conversion, and general word filtering to obtain suspected nickname data; or the suspected nicknames are subjected to vector conversion based on a preset semantic representation model to obtain suspected nickname data.

[0080] Specifically, when the suspect user information includes a suspect's profile picture, step S2 obtains the suspect user data including:

[0081] The suspect's headshot is scaled, flipped, binarized, and compressed to obtain suspect headshot data;

[0082] Alternatively, the suspect's image can be vectorized based on a preset image representation model to obtain suspect image data.

[0083] The suspected user data obtained after preprocessing the suspected user information has the same format as the type of all user data to be protected in the preset protection database in step S, which facilitates subsequent matching and similarity detection.

[0084] S3. Based on a preset matching algorithm, the suspected user data is matched with all the user data to be protected in the preset protection database. The user data to be protected is sorted according to the matching results to obtain the sorted user data to be protected.

[0085] It should be noted that the pre-set protection database contains a certain number of user data. Among these user data, some user data may be high-risk impersonation data that requires system protection, while the remaining user data may be ordinary user data with low impersonation risk. In some specific embodiments, all user data may be used as user data to be protected as needed. In the specific execution of this embodiment, only all user data to be protected in the pre-set protection database is matched with the suspected user data, which can save computing power.

[0086] Based on the aforementioned step S1:

[0087] In a preferred embodiment, the live streamer, live stream administrator, donors with a high degree of association with the live streamer, donors' friends with a high degree of association with the donors, and users with a high degree of association with the live streamer are considered as users with a high risk of being impersonated and who need to be protected by the system. The user data of the users to be protected is used as user data to be protected, and the user data of the remaining users is used as ordinary user data.

[0088] The users to be protected in the live stream are not fixed. In some specific situations, ordinary users can become users to be protected. For example, if a user with low affinity to the streamer interacts with the streamer a preset number of times within a preset time period, that user with low affinity will be upgraded to a user with high affinity to the streamer, and will change from an ordinary user with low risk of being impersonated to a user to be protected with high risk of being impersonated and requiring system protection. Conversely, in some specific situations, users to be protected can also become ordinary users. For example, if a user with high affinity to the streamer does not interact with the streamer a preset number of times within a preset time period, that user with high affinity will be downgraded to a user with low affinity to the streamer, and will change from a user to an ordinary user with low risk of being impersonated. Those skilled in the art should understand that the above examples are only for reference regarding changes in users to be protected and ordinary users, and are not intended to be limiting.

[0089] In another preferred embodiment, all users in the live broadcast room are considered as users with a high risk of being impersonated and who need to be protected by the system, and the user data of these users is used as the user data to be protected.

[0090] In this step, sorting all protected user data according to the matching results specifically involves sorting based on the similarity of the matches. That is, obtaining the similarity between the data to be protected and the suspected user data, sorting the user data to be protected according to the similarity, and obtaining the sorted user data to be protected, which is the sorted user data to be protected.

[0091] The user data to be protected includes nickname data to be protected and profile picture data to be protected; among them, the nickname data to be protected and the profile picture data to be protected are matched one-to-one.

[0092] Based on the processing steps S1 and S2, it can be determined that the suspected user data can be only suspected nickname data, or only suspected avatar data, or both suspected nickname and suspected avatar data. Therefore, the user sorting data obtained based on step S3 can be nickname sorting data to be protected or avatar sorting data to be protected.

[0093] It should be noted that step S3 only sorts the user data to be protected, without filtering. By sorting the user data to be protected that is highly similar to the suspected user data according to the matching results (matching similarity) in step S3, it is helpful to speed up the filtering process of the users to be protected in step S4.

[0094] S4. Based on the preset recall strategy and the sorting data of the users to be protected, filter the data of the users to be protected, record the data of the users to be protected that meets the preset requirements as the filtered user data, and output the information of the users to be protected corresponding to the filtered user data as the information of the impersonated users.

[0095] Specifically, recording the user data to be protected that meets the preset requirements as filtered user data includes: recording all user data to be protected that meets the preset requirements as filtered user data, or recording the user data to be protected that has the highest similarity to the suspected user data among all user data to be protected that meets the preset requirements as filtered user data.

[0096] Among these features, the recognition strength of suspected user data can be changed by modifying preset requirements to adapt to different recognition application scenarios.

[0097] Furthermore, in a specific embodiment, after step S4, the impersonated user can be identified based on the impersonated user information obtained in step S4, or the suspected user can be identified as having impersonated user data in the preset protection database. Then, a risk alert can be sent to the impersonated user or some of the users to be protected who are related to the impersonated user.

[0098] Based on the aforementioned steps S1-S4:

[0099] In a preferred embodiment, the donors who have a high degree of association with the streamer are designated as users to be protected, i.e., users 1.

[0100] When it is determined that User 1 (a donor with a high degree of association with the streamer) is being impersonated, or when Suspect 1 is impersonating User 1 (a donor with a high degree of association with the streamer), a risk warning will be sent to User 1. Specifically, the warning will include: "The nickname and / or avatar of Suspect 1 is highly similar to your (User 1's) nickname and / or avatar. Please note that Suspect 1 may have impersonated your user information." The specific behavior here is that Suspect 1 initiates communication with User 1, or User 1 initiates communication with Suspect 1. In the latter case, the historical communication behavior between User 1 and Suspect 1 can be used as one of the judgment conditions.

[0101] In this more specific implementation, impersonation can be further judged by combining it with specific behaviors. When it is determined that the user to be protected 1 (a donor with a high degree of association with the streamer) is the impersonated user, or when the suspected user 1 is impersonating the user to be protected 1 (a donor with a high degree of association with the streamer), a risk warning is sent to the user to be protected 1. The specific process is as follows: when the suspected user 1 triggers a preset behavior, the historical nickname and / or image of the suspected user 1 is obtained, and it is determined whether the change of the historical nickname and / or image of the suspected user 1 exceeds a preset threshold. If it does, the obtained historical nickname and / or image of the suspected user 1, as well as the current nickname and image, are sent to the current chat window and / or SMS window of the impersonated user as a warning. The warning content includes: "The changed nickname and / or avatar of the suspected user 1 is different from..." "Your (user data to be protected 1) nickname and / or avatar are highly similar. Please be aware that suspect user 1 may be impersonating you." The preset behavior specifically includes suspect user 1 sending messages to the impersonated user's friends. Message content includes, but is not limited to, money, invitations to make offers, and promises. The preset threshold here can be further processed based on steps S1-S4. That is, although the suspect user's avatar and / or nickname obtained in steps S1-S4 are relatively close to the impersonated user, in this step, it can be further judged manually or by a preset model whether the changes in suspect user 1's historical nickname and / or image exceed the preset threshold to avoid confusion. Of course, in this embodiment, to ensure communication security, the preset threshold judgment can be omitted, and the preset behavior can be judged directly.

[0102] Building upon the previous embodiment, donors with a high degree of association with the streamer are designated as users to be protected 1, and friends of donors with a high degree of association with the donors are designated as users to be protected 2. That is, when identifying users to be protected 1 among the friends of users to be protected 2, the pre-defined behavior is judged based on the contact between suspected user 1 and users to be protected 1. The pre-defined behavior includes: changes in the frequency and content of contact between suspected user 1 and users to be protected 2, for example, suspected user 1 previously had little communication with users to be protected 2, but communication increases in a short period; or, for example, suspected user 1 previously only had simple communication with users to be protected 2, such as suspected user 1 initiating a conversation with users to be protected 2 and receiving no response, very little response, or a very simple formatted response, and now the response is abnormal, as well as the pre-defined behavior in the previous embodiment.

[0103] When it is determined that User 2 (a friend of the donor with a high degree of association with the donor) is being impersonated, or when it is determined that Suspect 1 is impersonating User 2 (a friend of the donor with a high degree of association with the donor), a risk alert is sent to User 1 (a donor with a high degree of association with the streamer), a friend of User 2. The specific process includes: when Suspect 1 is impersonating User 2, determining whether Suspect 1's communication partner is User 1. If so, further determining whether Suspect 1 has engaged in pre-defined behavior. If so, the historical nickname and / or image of Suspect 1, as well as the current nickname and image, are sent to User 1's current chat window and / or SMS window as a reminder. Specifically, it includes: "Suspect 1's modified nickname and / or avatar is highly similar to the nickname and / or avatar of your friend (User 2). Please note that Suspect 1 may have impersonated your friend's user information."

[0104] Alternatively, if it is determined that User 2 (a friend of the donor with a high degree of association with the donor) is the user being impersonated, or if it is determined that Suspect User 1 is impersonating User 2 (a friend of the donor with a high degree of association with the donor), and Suspect User 1 sends a friend request to User 2's friend—User 1 (the donor with a high degree of association with the streamer), a risk warning will be sent to User 1 (the friend of User 2)—the donor with a high degree of association with the streamer. Specifically, the warning will include: "The nickname and / or avatar of Suspect User 1 is highly similar to the nickname and / or avatar of your friend (User 2). Please note that Suspect User 1 may have impersonated your friend's user information."

[0105] In another preferred embodiment, the broadcaster is designated as user 1 to be protected.

[0106] When it is determined that User 2 (the streamer) to be protected is being impersonated, or when it is determined that Suspect User 1 is impersonating User 1 (the streamer), an authorization reminder is sent to User 1. Specifically, this reminder includes the message: "Suspect User 1's modified nickname and / or avatar is highly similar to your (the streamer's) nickname and / or avatar. Please select whether to agree to Suspect User 1's modification request." In this case, the modified user information will not be displayed immediately. Only after User 1 (the streamer) authorizes the change will the suspect's information be updated to the modified user information. By sending an authorization reminder to User 1, potential impersonation can be prevented at the source.

[0107] In one specific embodiment, after step S4, the impersonated user information and the corresponding suspected user information can be packaged, stored, and displayed to facilitate further manual verification by staff and improve the accuracy of the above identification method.

[0108] Example 1 provides a method for identifying impersonated user information. First, the suspected user information is matched against all / specified protected user information in a pre-set protection database. Based on the matching results (similarity), the protected user information is sorted. Further, a pre-set recall strategy is used to filter the sorted user information, selecting those that meet preset requirements. This sorting-then-filtering approach improves identification efficiency, enabling rapid and accurate detection of potential fraud. Furthermore, when an impersonated user is identified, or a suspected user is found to be impersonating protected user data in the pre-set protection database, a risk alert can be sent to the impersonated user or related protected users, preventing potential fraud at its source. Simultaneously, the impersonated user information and the suspected user information can be packaged, stored, and displayed for further manual verification by staff, improving the accuracy of the identification method.

[0109] In the above embodiment 1, the preset matching algorithm is a preset nickname matching algorithm or a preset avatar matching algorithm; the preset recall strategy is a preset nickname recall strategy or a preset avatar recall strategy or a preset nickname and avatar combined recall strategy or a preset avatar and nickname combined recall strategy; the preset requirements are preset nickname requirements or preset avatar requirements or preset nickname and avatar combined requirements or preset avatar and nickname combined requirements; the filtered user data is filtered nickname data or filtered avatar data or filtered nickname and avatar combined data or filtered avatar and nickname combined data.

[0110] For the corresponding situations described above, please refer to Examples 2 to 5 below.

[0111] Based on the aforementioned implementation 1, in embodiment 2, when the suspected user information in step S1 only contains the suspected nickname, or only the suspected nickname in the suspected user information needs to be identified, step S3 specifically involves:

[0112] S3A. Based on the preset nickname matching algorithm, the suspected nickname data is matched with all the nickname data to be protected in the preset protection database. The nickname data to be protected is sorted according to the nickname matching results to obtain the nickname data to be protected sorted data.

[0113] Among them, the pre-set nickname matching algorithm includes, but is not limited to, character-level literal similarity algorithms, such as similarity algorithms based on text edit distance, Euclidean distance, etc., or semantic similarity algorithms based on static / dynamic vector representation.

[0114] A pre-set nickname matching algorithm is used to obtain the matching results (similarity) between the suspected nickname data and all nickname data to be protected. The nickname data to be protected is sorted from largest to smallest similarity. Of course, it can also be sorted from smallest to largest similarity. The sorting method is not specifically limited. The purpose of sorting is to provide cleaner input information for step S4 and improve the filtering speed of nickname data to be protected in step S4.

[0115] Furthermore, following step S3A, step S4 specifically involves:

[0116] S4A. Based on the preset nickname recall strategy and the nickname ranking data to be protected, filter the nickname data to be protected, record the nickname data to be protected that meets the preset nickname requirements as filtered nickname data, and output the nickname to be protected corresponding to the filtered nickname data as the impersonated nickname.

[0117] It should be noted that the preset nickname recall strategy is to filter the nickname data to be protected based on the nickname sorting data in step S3, retain the nickname data to be protected that meets the preset nickname requirements, and then find the impersonated nickname through the nickname data to be protected that meets the preset nickname requirements, or determine that the suspect user has impersonated the nickname to be protected in the preset protection database.

[0118] The process of recording the protected nickname data that meets the preset nickname requirements as filtered nickname data can be either: recording all protected nickname data that meets the preset nickname requirements as filtered nickname data, or recording the protected nickname data that has the highest similarity to the suspected nickname among all protected nickname data that meets the preset nickname requirements as filtered nickname data.

[0119] Furthermore, the recognition strength for suspicious nicknames can be altered by modifying preset nickname requirements to adapt to different recognition application scenarios. The following embodiments are also applicable.

[0120] Furthermore, in a specific embodiment, after step S4A, the impersonated user can be identified based on the impersonated nickname obtained in step S4A, or it can be determined that the suspected user has engaged in impersonating nickname data to be protected in the preset protection database. Then, a risk alert can be sent to the impersonated user or some protected users related to the impersonated user. See the example in Embodiment 1 for details.

[0121] Example 2 provides a method for identifying impersonation of a protected user's nickname. This method involves matching suspected nickname data with all / specified protected nicknames in a pre-set protection database. Based on the matching results (similarity), the protected nickname data is sorted. Furthermore, a pre-set nickname recall strategy is used to filter the sorted protected nickname data, selecting those that meet preset requirements. This identifies the impersonated nickname and / or confirms that the suspected user is impersonating a protected nickname in the pre-set protection database. Similarly, Example 2 uses a sorting-then-filtering approach to identify suspected nicknames, effectively improving accuracy and quickly detecting fraudulent activities.

[0122] Based on the aforementioned implementation 1, in embodiment 3, when the suspected user information in step S1 only contains a suspected avatar, or only the suspected avatar in the suspected user information needs to be identified, step S3 specifically involves:

[0123] S3B. Based on the preset avatar matching algorithm, the suspected avatar data is matched with all the avatar data to be protected in the preset protection database. The avatar data to be protected is sorted according to the avatar matching results to obtain the sorted avatar data to be protected.

[0124] It should be noted that the preset avatar matching algorithms include, but are not limited to, pixel-level similarity measurement algorithms such as MD5, histogram, PSNR, and SSIM, or similarity measurement algorithms based on vector representation.

[0125] A pre-set avatar matching algorithm is used to obtain the matching results (similarity) between the suspected avatar data and all the avatar data to be protected. The avatar data to be protected is sorted from largest to smallest similarity. Of course, it can also be sorted from smallest to largest similarity. The sorting method is not specifically limited. The purpose of sorting is to provide cleaner input information for step S4 and improve the filtering speed of the avatar data to be protected in step S4.

[0126] Furthermore, following step S3B, step S4 specifically involves:

[0127] S4B. Based on the preset avatar recall strategy and the avatar sorting data to be protected, filter the avatar data to be protected, record the avatar data to be protected that meets the preset avatar requirements as filtered avatar data, and output the avatar to be protected corresponding to the filtered avatar data as the impersonated avatar.

[0128] It should be noted that the preset avatar recall strategy is to filter the avatar data to be protected based on the avatar sorting data in step S3, retain the avatar data to be protected that meets the preset avatar requirements, and then find the avatar that has been impersonated or determine that the suspect user has impersonated the avatar in the preset protection database.

[0129] The process of recording the profile picture data that meets the preset profile picture requirements as filtered profile picture data can be either: recording all profile picture data that meets the preset profile picture requirements as filtered profile picture data, or recording the profile picture data that has the highest similarity to the suspect profile picture among all profile picture data that meets the preset profile picture requirements as filtered profile picture data.

[0130] Furthermore, in a specific embodiment, after step S4B, the impersonated user can be identified based on the impersonated avatar obtained in step S4B, or it can be determined that the suspected user has engaged in impersonating avatar data to be protected in the preset protection database. Then, a risk alert is sent to the impersonated user or some protected users related to the impersonated user. See the example in Embodiment 1 for details.

[0131] Example 3 provides a method for identifying impersonation of a user's profile picture. This method involves matching the suspected profile picture data with all / specified profile pictures in a pre-set protection database. Based on the matching results (similarity), the profile picture data is sorted. Furthermore, a pre-set profile picture retrieval strategy is used to filter the sorted profile picture data, selecting those that meet preset requirements. This identifies the impersonated profile picture and / or confirms that the suspected user is impersonating a profile picture in the pre-set protection database. Similarly, Example 3 uses a sorting-then-filtering approach to identify suspected profile pictures, effectively improving accuracy and quickly detecting fraudulent activities.

[0132] Based on the aforementioned implementation 1, in embodiment 4, when the suspected user information in step S1 simultaneously includes both a suspected nickname and a suspected avatar, or when it is necessary to simultaneously identify both the suspected nickname and the suspected avatar in the suspected user information, step S3 can specifically be:

[0133] S3C1. Based on the preset nickname matching algorithm, the suspected nickname data is matched with all the nickname data to be protected in the preset protection database, and all the nickname data to be protected is sorted according to the nickname matching results to obtain the nickname ranking data to be protected.

[0134] S3C2. Based on the preset nickname recall strategy and the nickname ranking data to be protected, filter the nickname data to be protected, record the nickname data to be protected that meets the preset nickname requirements as filtered nickname data, and record the avatar data to be protected that matches the filtered nickname data as the first avatar data to be protected.

[0135] S3C3. Based on the preset avatar matching algorithm, the suspected avatar data is matched with the first avatar data to be protected, and all the first avatar data to be protected are sorted according to the first avatar matching result to obtain the first avatar sorting data to be protected.

[0136] Unlike the aforementioned Embodiment 2, in Embodiment 4, although step S3C also matches the suspected nickname data with all the nickname data to be protected in the preset protection library, the goal is not to obtain the nickname ranking data to be protected, but to use the obtained nickname ranking data to obtain the first avatar ranking data to be protected.

[0137] By acquiring profile picture data that matches the profile picture data that meets the preset nickname requirements, and then further matching the suspected profile picture data with the first profile picture data to be protected, the first profile picture ranking data to be protected is obtained.

[0138] It should be noted that there may be multiple first-to-protect headshot data obtained in step S3C2. Therefore, in step S3C3, the suspect headshot data is matched with all the first-to-protect headshot data obtained in step S3C2 to obtain the first-to-protect headshot sorting data.

[0139] Steps S3C1 to S3C3 above involve first acquiring the first set of profile picture data that meets the preset nickname requirements, then filtering the profile picture data to remove profile picture data that is not very similar to the suspected nickname data, and then further sorting the filtered profile picture data to obtain smaller and more accurate profile picture sorting data. This provides more concise and accurate input information for step S4, improving the filtering accuracy and speed of the profile picture data in step S4.

[0140] Furthermore, after step S3C3, step S4 specifically involves:

[0141] S4C. Based on the preset nickname and avatar joint recall strategy and the first avatar ranking data to be protected, filter the first avatar data to be protected, record the first avatar data to be protected that meets the preset nickname and avatar joint requirements as the first filtered avatar data, output the avatar to be protected corresponding to the first filtered avatar data as the impersonated avatar, and output the nickname to be protected corresponding to the nickname data to be protected that matches the first filtered avatar data as the impersonated nickname.

[0142] It should be noted that the preset nickname and avatar joint recall strategy is to filter the first avatar data to be protected based on the sorting data of the first avatar to be protected in step S3C3, retain the first avatar data to be protected that meets the preset nickname and avatar joint requirements, and then find the avatar and nickname being impersonated through the first avatar data to be protected that meets the preset nickname and avatar joint requirements, or determine that the suspect user has impersonated the avatar and nickname to be protected in the preset protection database.

[0143] It is understandable that, based on steps S3C3 and S4C, the first filtered avatar data corresponds to the first avatar data to be protected, the first avatar data to be protected corresponds to the avatar data to be protected, and the avatar data to be protected corresponds to the nickname data to be protected. Each first filtered avatar data has a corresponding nickname data to be protected.

[0144] The first protected avatar data that meets the preset nickname and avatar combination requirements can be recorded as the first filtered avatar data. This can be either: all the first protected avatar data that meets the preset nickname and avatar combination requirements are recorded as the first filtered avatar data, or: among all the first protected avatar data that meets the preset nickname and avatar combination requirements, the first protected avatar data with the highest similarity to the suspect avatar is recorded as the first filtered avatar data.

[0145] Furthermore, in a specific embodiment, after step S4C, the impersonated user can be identified based on the impersonated avatar and / or impersonated nickname obtained in step S4C, or it can be determined that the suspected user has engaged in impersonating avatar data and / or nickname data to be protected in the preset protection database. Then, a risk alert is sent to the impersonated user or some protected users related to the impersonated user. See the example in Embodiment 1 for details.

[0146] Example 4 provides a method for identifying impersonation of a user's nickname and avatar. It involves matching suspected nickname data with all protected nickname data in a pre-set protection database. The matched avatar data is then filtered to remove those with low similarity to the suspected nickname data. The filtered avatar data is further sorted to obtain a smaller, more accurate sorted dataset. A pre-set nickname and avatar joint recall strategy is then used to filter the sorted first-order avatar data, selecting those that meet the pre-set nickname and avatar joint requirements. This identifies the impersonated avatar and nickname, confirming that the suspected user is impersonating a user in the pre-set protection database. Example 4 uses a sorting-filtering-sorting-filtering approach to identify suspected avatars and nicknames, effectively improving identification efficiency and accuracy, and quickly detecting fraudulent activities.

[0147] Based on the aforementioned implementation 1, in embodiment 5, when the suspected user information in step S1 simultaneously includes both a suspected nickname and a suspected avatar, or when it is necessary to simultaneously identify both the suspected nickname and the suspected avatar in the suspected user information, step S3 may further be as follows:

[0148] S3D1. Based on the preset avatar matching algorithm, the suspected avatar data is matched with all the avatar data to be protected in the preset protection database. The avatar data to be protected is sorted according to the avatar matching results to obtain the sorted avatar data to be protected.

[0149] S3D2. Based on the preset avatar recall strategy and the avatar sorting data to be protected, filter the avatar data to be protected, record the avatar data to be protected that meets the preset avatar requirements as filtered avatar data, and record the nickname data to be protected that matches the filtered avatar data as the first nickname data to be protected.

[0150] S3D3. Based on the preset nickname matching algorithm, the suspected nickname data is matched with the first nickname data to be protected. According to the first nickname matching result, all the first nickname data to be protected are sorted to obtain the first nickname data to be protected sorted data.

[0151] Following step S3D3, step S4 specifically involves:

[0152] Based on the preset avatar-nickname joint recall strategy and the first nickname to be protected sorting data, the first nickname to be protected data is filtered, and the first nickname to be protected data that meets the preset avatar-nickname joint requirements is recorded as the first filtered nickname data. The nickname to be protected corresponding to the first filtered nickname data is output as the impersonated nickname, and the avatar to be protected corresponding to the avatar data to be protected that matches the first filtered nickname data is output as the impersonated avatar.

[0153] Furthermore, in a specific embodiment, after step S4D, the impersonated user can be identified based on the impersonated nickname and / or impersonated avatar obtained in step S4D, or it can be determined that the suspected user has engaged in impersonating protected nickname data and / or protected avatar data in the preset protection database. Then, a risk alert is sent to the impersonated user or some protected users related to the impersonated user. See the example in Embodiment 1 for details.

[0154] The specific implementation methods and effects of the above steps S3D1 to S4D can be found in the above embodiment 4, and will not be repeated here.

[0155] It is understood that in both Embodiment 4 and Embodiment 5 above, the user data to be protected is screened in one round to remove user data with low similarity to suspected user data, and then the screened user data to be protected is sorted to obtain user data to be protected with smaller data volume and more accurate sorting data, which provides more concise and accurate input information for step S4, and improves the filtering accuracy and speed of the avatar data to be protected in step S4.

[0156] In another preferred embodiment, the impersonated nicknames and impersonated avatars obtained in Embodiments 4 and 5 can be further compared, and the impersonated nicknames and impersonated avatars that appear in both identification methods can be selected as the final impersonated nicknames and impersonated avatars, thereby further improving the identification accuracy.

[0157] In a preferred real-time example, based on Embodiment 2, or Embodiment 3, or Embodiment 4, or Embodiment 5, after step S4, the following may also be added:

[0158] The impersonated user information is stored as updated protected user information, and the preset matching algorithm is updated based on the updated protected user information.

[0159] Among them, the impersonated user information includes the impersonated nickname and / or the impersonated avatar.

[0160] Optionally, the impersonated nickname information is stored as updated protected nickname information, and the preset nickname matching algorithm is updated based on the updated protected nickname information;

[0161] Alternatively, the impersonated avatar information can be stored as updated protective avatar information, and the preset avatar matching algorithm can be updated based on the updated protective avatar information;

[0162] Alternatively, the impersonated nickname information and the impersonated avatar information can be stored as updated protected user information, and the preset nickname-avatar joint matching algorithm / preset avatar-nickname joint matching algorithm can be updated based on the updated protected user information.

[0163] In a preferred embodiment, storing the impersonated nicknames and / or avatars facilitates data retrieval by staff to verify the effectiveness of the identification method. On the other hand, the relevant sorting and representation algorithms can be updated based on the impersonated nicknames and / or avatar data to improve the accuracy and speed of the identification method.

[0164] This application also provides an embodiment of a system for identifying impersonated user information; see [link to example]. Figure 2 .

[0165] A system for identifying impersonated user information includes:

[0166] Suspect Information Acquisition Module 1: Acquires suspect user information;

[0167] Suspect information preprocessing module 2 is used to preprocess the suspect user information to obtain suspect user data;

[0168] The information matching and sorting module 3 is used to match the suspected user data with all the user data to be protected in the preset protection database based on a preset matching algorithm, and sort all the user data to be protected according to the matching results to obtain the user data to be protected sorting data.

[0169] The identification output module 4 is used to filter the user data to be protected based on a preset recall strategy and the sorting data of the users to be protected, record the user data to be protected that meets the preset requirements as filtered user data, and output the user information to be protected corresponding to the filtered user data as the impersonated user information.

[0170] In a preferred embodiment, the suspected user information includes a suspected nickname and a suspected avatar; the suspected nickname and the suspected avatar are matched; the user data to be protected includes nickname data to be protected and avatar data to be protected; the nickname data to be protected and the avatar data to be protected are matched.

[0171] The preset matching algorithm is either a preset nickname matching algorithm or a preset avatar matching algorithm; the preset recall strategy is either a preset nickname recall strategy, a preset avatar recall strategy, a preset nickname and avatar combined recall strategy, or a preset avatar and nickname combined recall strategy; the preset requirements are either preset nickname requirements, preset avatar requirements, preset nickname and avatar combined requirements, or preset avatar and nickname combined requirements; the filtered user data is either filtered nickname data, filtered avatar data, filtered nickname and avatar combined data, or filtered avatar and nickname combined data.

[0172] In a preferred embodiment, the suspect information acquisition module 1 includes a suspect nickname acquisition module 101 and a suspect avatar acquisition module 102;

[0173] The suspect nickname acquisition module 101 is used to acquire the suspect's nickname;

[0174] The suspect image acquisition module 102 is used to acquire the suspect's image.

[0175] In a preferred embodiment, the suspect information preprocessing module 2 includes a suspect nickname preprocessing module 201 and a suspect avatar preprocessing module 202;

[0176] The suspected nickname preprocessing module 201 is used to perform text preprocessing on the suspected nickname to obtain suspected nickname data;

[0177] The suspect image preprocessing module 202 is used to perform text preprocessing on the suspect image to obtain suspect image data.

[0178] In a preferred embodiment, the information matching and sorting module 3 includes a nickname matching algorithm module 301, an avatar matching algorithm module 302, a nickname and avatar joint matching algorithm module 303, and an avatar and nickname joint matching algorithm module 304.

[0179] The nickname matching algorithm module 301 is used to match the suspected nickname data with all the nickname data to be protected in the preset protection database based on the preset nickname matching algorithm, and sort all the nickname data to be protected according to the nickname matching results to obtain the nickname ranking data to be protected.

[0180] The avatar matching algorithm module 302 is used to match the suspected avatar data with all the avatar data to be protected in the preset protection database based on the preset avatar matching algorithm, and sort all the avatar data to be protected according to the avatar matching results to obtain the avatar ranking data to be protected.

[0181] The nickname and avatar joint matching algorithm module 303 is used to match the suspected nickname data with all the nickname data to be protected in the preset protection database based on the preset nickname matching algorithm, and sort all the nickname data to be protected according to the nickname matching results to obtain the nickname ranking data to be protected.

[0182] Based on the preset nickname recall strategy and the nickname ranking data to be protected, the nickname data to be protected is filtered, and the nickname data to be protected that meets the preset nickname requirements is recorded as the filtered nickname data, and the avatar data to be protected that matches the filtered nickname data is recorded as the first avatar data to be protected.

[0183] Based on the preset avatar matching algorithm, the suspected avatar data is matched with the first avatar data to be protected, and all the first avatar data to be protected are sorted according to the first avatar matching result to obtain the first avatar data to be protected sorted data.

[0184] The avatar and nickname joint matching algorithm module 304 is used to match the suspected avatar data with all the avatar data to be protected in the preset protection database based on the preset avatar matching algorithm, and sort all the avatar data to be protected according to the avatar matching results to obtain the avatar ranking data to be protected.

[0185] Based on the preset avatar recall strategy and the avatar sorting data to be protected, the avatar data to be protected is filtered, and the avatar data to be protected that meets the preset avatar requirements is recorded as the filtered avatar data, and the nickname data to be protected that matches the filtered avatar data is recorded as the first nickname data to be protected.

[0186] Based on the preset nickname matching algorithm, the suspected nickname data is matched with the first nickname data to be protected, and all the first nickname data to be protected is sorted according to the first nickname matching result to obtain the first nickname sorting data to be protected.

[0187] In a preferred embodiment, the recognition output module 4 includes a nickname recognition output module 401, an avatar recognition output module 402, a nickname and avatar joint recognition output module 403, and an avatar and nickname joint recognition output module 404.

[0188] The nickname recognition output module 401 is used, after the nickname matching algorithm module 301, to filter the nickname data to be protected based on the preset nickname recall strategy and the nickname ranking data to be protected, to record the nickname data to be protected that meets the preset nickname requirements as the filtered nickname data, and to output the nickname to be protected corresponding to the filtered nickname data as the impersonated nickname.

[0189] The avatar recognition output module 402 is used, after the avatar matching algorithm module 302, to filter the first avatar data to be protected based on the preset nickname avatar joint recall strategy and the first avatar ranking data to be protected, to record the first avatar data to be protected that meets the preset nickname avatar joint requirements as the first filtered avatar data, to output the avatar to be protected corresponding to the first filtered avatar data as the impersonated avatar, and to output the nickname to be protected corresponding to the nickname data to be protected that matches the first filtered avatar data as the impersonated nickname.

[0190] The nickname and avatar joint recognition output module 403 is used, after the nickname and avatar joint matching algorithm module 303, to filter the nickname and avatar joint data to be protected based on the preset nickname and avatar joint recall strategy and the nickname and avatar joint sorting data to be protected, to record the nickname and avatar joint data to be protected that meets the preset nickname and avatar joint requirements as filtered nickname and avatar joint data, to output the nickname information to be protected corresponding to the filtered nickname and avatar joint data as the impersonated nickname information, and to output the avatar information to be protected corresponding to the filtered nickname and avatar joint data as the impersonated avatar information.

[0191] The avatar-nickname joint recognition output module 404 is used, after the avatar-nickname joint matching algorithm module 304, to filter the first nickname data to be protected based on the preset avatar-nickname joint recall strategy and the first nickname ranking data to be protected, to record the first nickname data to be protected that meets the preset avatar-nickname joint requirements as the first filtered nickname data, to output the nickname to be protected corresponding to the first filtered nickname data as the impersonated nickname, and to output the avatar to be protected corresponding to the avatar data to be protected that matches the first filtered nickname data as the impersonated avatar.

[0192] In a preferred embodiment, the identification system further includes a storage and display module;

[0193] The storage and display module is used to package, store, and display the impersonated user information and the corresponding suspected user information.

[0194] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process and implementation effect of the systems, devices and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0195] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0196] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0197] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0198] 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 the present 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 the present 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.

[0199] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying impersonated user information, characterized in that, include: S1. Obtain suspect user information; wherein, the suspect user information includes suspect nickname and suspect avatar; the suspect nickname and the suspect avatar are matched; S2. Preprocess the suspected user information to obtain suspected user data, including: preprocessing the suspected nickname with text to obtain suspected nickname data; preprocessing the suspected avatar with image to obtain suspected avatar data; S3. Based on a preset matching algorithm, the suspected user data is matched with all user data to be protected in a preset protection database. The user data to be protected is then sorted according to the matching results to obtain sorted user data to be protected. The user data to be protected includes user nickname data and user avatar data to be protected. The user nickname data and user avatar data to be protected are matched. The preset matching algorithm is either a preset nickname matching algorithm or a preset avatar matching algorithm. Step S3 includes: S3C, based on the preset nickname matching algorithm, matching the suspected nickname data with all nickname data to be protected in the preset protection database, and sorting all the nickname data to be protected according to the nickname matching results to obtain nickname ranking data to be protected; based on the preset nickname recall strategy and the nickname ranking data to be protected, filtering the nickname data to be protected, recording the nickname data to be protected that meets the preset nickname requirements as filtered nickname data, and recording the avatar data to be protected that matches the filtered nickname data as first avatar data to be protected; based on the preset avatar matching algorithm, matching the suspected avatar data with the first avatar data to be protected, and sorting all the first avatar data to be protected according to the first avatar matching results to obtain first avatar ranking data to be protected; S4. Based on the preset recall strategy and the sorting data of the users to be protected, filter the data of the users to be protected, record the data of the users to be protected that meets the preset requirements as the filtered user data, and output the information of the users to be protected corresponding to the filtered user data as the impersonated user information. After step S3C, step S4 specifically involves: S4C, based on the preset nickname and avatar joint recall strategy and the first avatar ranking data to be protected, filtering the first avatar data to be protected, recording the first avatar data to be protected that meets the preset nickname and avatar joint requirements as the first filtered avatar data, outputting the avatar to be protected corresponding to the first filtered avatar data as the impersonated avatar, and outputting the nickname to be protected corresponding to the nickname data to be protected that matches the first filtered avatar data as the impersonated nickname.

2. The method for identifying impersonated user information according to claim 1, characterized in that, Step S3 can also be: S3D, based on the preset avatar matching algorithm, match the suspected avatar data with all the avatar data to be protected in the preset protection library, and sort all the avatar data to be protected in descending order according to the avatar matching results to obtain the avatar sorting data to be protected. Based on the preset avatar recall strategy and the avatar sorting data to be protected, the avatar data to be protected is filtered, and the avatar data to be protected that meets the preset avatar requirements is recorded as the filtered avatar data, and the nickname data to be protected that matches the filtered avatar data is recorded as the first nickname data to be protected. Based on the preset nickname matching algorithm, the suspected nickname data is matched with the first nickname data to be protected, and all the first nickname data to be protected is sorted according to the first nickname matching result to obtain the first nickname sorting data to be protected. Following step S3D, step S4 specifically involves: S4D: Based on the preset avatar-nickname joint recall strategy and the first nickname to be protected sorting data, filter the first nickname to be protected data, record the first nickname to be protected that meets the preset avatar-nickname joint requirements as the first filtered nickname data, output the nickname to be protected corresponding to the first filtered nickname data as the impersonated nickname, and output the avatar to be protected corresponding to the avatar data to be protected that matches the first filtered nickname data as the impersonated avatar.

3. The method for identifying impersonated user information according to claim 1, characterized in that, The suspected nickname is preprocessed to obtain suspected nickname data, including: The suspected nicknames were normalized, converted to uppercase and lowercase letters, converted to simplified and traditional Chinese characters, converted to Martian language, and filtered for common words to obtain the suspected nickname data. Alternatively, the suspected nickname can be vectorized based on a preset semantic representation model to obtain the suspected nickname data.

4. The method for identifying impersonated user information according to claim 1, characterized in that, The suspected headshot is preprocessed to obtain suspected headshot data, specifically including: The suspect's headshot is scaled, flipped, binarized, and compressed to obtain suspect headshot data; Alternatively, the suspect's image can be vectorized based on a preset image representation model to obtain suspect image data.

5. The method for identifying impersonated user information according to claim 1, characterized in that, The process after step S4 also includes: The impersonated user information is stored as updated protected user information, and the preset matching algorithm is updated based on the updated protected user information.

6. The method for identifying impersonated user information according to claim 1, characterized in that, The process after step S4 also includes: The impersonated user information and the corresponding suspected user information are packaged, stored, and displayed.

7. The method for identifying impersonated user information according to claim 1, characterized in that, The process after step S4 also includes: Send a risk alert about being impersonated to the user whose information was being impersonated.

8. A system for identifying impersonated user information, characterized in that, include: The suspect information acquisition module acquires suspect user information, including a suspect nickname and a suspect avatar; the suspect nickname and the suspect avatar are matched. The suspect information preprocessing module preprocesses the suspect user information to obtain suspect user data, including: performing text preprocessing on the suspect nickname to obtain suspect nickname data; or performing image preprocessing on the suspect avatar to obtain suspect avatar data. The information matching and sorting module is used to match the suspected user data with all user data to be protected in a preset protection database based on a preset matching algorithm, and sort all user data to be protected in descending order according to the matching results to obtain user data to be protected sorting; wherein, the user data to be protected includes nickname data to be protected and avatar data to be protected; the nickname data to be protected and the avatar data to be protected are matched; the preset matching algorithm is a preset nickname matching algorithm or a preset avatar matching algorithm; The information matching and sorting module is used to perform the following: S3C, based on the preset nickname matching algorithm, matching the suspected nickname data with all the nickname data to be protected in the preset protection database, and sorting all the nickname data to be protected according to the nickname matching results to obtain the nickname data to be protected sorted; based on the preset nickname recall strategy and the nickname data to be protected sorted, filtering the nickname data to be protected, recording the nickname data to be protected that meets the preset nickname requirements as filtered nickname data, and recording the avatar data to be protected that matches the filtered nickname data as the first avatar data to be protected; based on the preset avatar matching algorithm, matching the suspected avatar data with the first avatar data to be protected, and sorting all the first avatar data to be protected according to the first avatar matching results to obtain the first avatar data to be protected sorted. The identification output module is used to filter the user data to be protected based on a preset recall strategy and the sorting data of the user to be protected, record the user data to be protected that meets the preset requirements as the filtered user data, and output the user information to be protected corresponding to the filtered user data as the impersonated user information. When the information matching and sorting module is used to execute S3C, the identification output module is used for S4C, based on the preset nickname and avatar joint recall strategy and the first avatar to be protected sorting data, to filter the first avatar to be protected data, record the first avatar to be protected data that meets the preset nickname and avatar joint requirements as the first filtered avatar data, output the avatar to be protected corresponding to the first filtered avatar data as the impersonated avatar, and output the nickname to be protected corresponding to the nickname to be protected that matches the first filtered avatar data as the impersonated nickname.

Images