Content identification method and related device

Abstract

The invention discloses a content identification method and a related device. The method comprises the following steps: acquiring to-be-processed interaction data generated by interaction of to-be-identified content; and performing first exposure determination on the to-be-identified content corresponding to the to-be-processed interaction data to obtain a first exposure. And performing object type identification on a to-be-identified object corresponding to the to-be-processed interaction data to obtain an object identification result. If it is determined that the to-be-recognized object is the target object based on the object recognition result, the target object concentration of the to-be-recognized content is determined. And if the first exposure reaches an exposure threshold value and the target object concentration reaches a concentration threshold value, determining the to-be-identified content as the target content. Whether the to-be-recognized content is the target content or not is determined by combining the first exposure and the target object concentration, the set exposure threshold is reduced to a certain extent, the target content can be found under the lower exposure under the same human input, the situation that a large number of users touch uncomfortable content due to the fact that the exposure threshold is set to be too high is avoided, and the user experience is improved. And therefore, great influence is avoided.

Description

Technical Field

[0001] This application relates to the field of computers, and in particular, to a content recognition method and related devices. Background Art

[0002] With the wide application of Internet technology in various industries, the number of users producing content every day has reached hundreds of millions. The accumulation of massive big data has greatly enriched people's spiritual and material lives, but the risks and hidden dangers of Internet content are becoming increasingly prominent. To effectively control content risks, content recognition technology can be used to quickly identify risky content from massive data.

[0003] Content recognition technology is a technology applied to content security prevention and control. The purpose is to collect, process, and analyze a large amount of content data produced by Internet users (such as articles, short videos) in real time to discover and warn of potential malicious content exposure risks. Related technologies can discover target content based on exposure. Usually, certain rules are set, and after a certain exposure is reached, relevant personnel are warned to review.

[0004] However, this method is usually limited by manpower and will set a relatively high exposure threshold. Often, after the exposure reaches such a threshold and then manual intervention is carried out, a large number of users have already come into contact with inappropriate content, causing a greater impact. Summary of the Invention

[0005] To solve the above technical problems, this application provides a content recognition method and related devices, which can reduce the set exposure threshold to a certain extent, and can discover target content at a lower exposure with the same human input, avoiding a large number of users coming into contact with inappropriate content due to too high an exposure threshold, and thus avoiding a greater impact.

[0006] The embodiments of this application disclose the following technical solutions:

[0007] On the one hand, the embodiments of this application provide a content recognition method, and the method includes:

[0008] Obtain the to-be-processed interaction data, where the to-be-processed interaction data is generated by interacting with the to-be-recognized content;

[0009] Determine a first exposure degree for the to-be-recognized content corresponding to the to-be-processed interaction data to obtain the first exposure degree of the to-be-recognized content, and identify the object type of the to-be-recognized object corresponding to the to-be-processed interaction data to obtain an object recognition result, where the first exposure degree represents the number of times of interacting with the to-be-recognized content, and the to-be-recognized object is the object that generates the to-be-processed interaction data;

[0010] If it is determined that the object to be recognized is the target object based on the object recognition result, determine the target object concentration of the content to be recognized, where the target object concentration represents the frequency of occurrence of the target object among all objects that have generated the interactive data to be processed for the content to be recognized;

[0011] If the first exposure reaches the exposure threshold and the target object concentration reaches the concentration threshold, determine that the content to be recognized is the target content.

[0012] On the one hand, an embodiment of the present application provides a content recognition device, and the device includes an acquisition unit, a determination unit, and an identification unit:

[0013] The acquisition unit is used to acquire interactive data to be processed, and the interactive data to be processed is generated by interacting with the content to be recognized;

[0014] The determination unit is used to determine the first exposure of the content to be recognized corresponding to the interactive data to be processed, and obtain the first exposure of the content to be recognized, where the first exposure represents the number of times of interacting with the content to be recognized;

[0015] The identification unit is used to identify the object type of the object to be recognized corresponding to the interactive data to be processed, and obtain an object recognition result, where the object to be recognized is the object that generates the interactive data to be processed;

[0016] The determination unit is further used to, if it is determined that the object to be recognized is the target object based on the object recognition result, determine the target object concentration of the content to be recognized, where the target object concentration represents the frequency of occurrence of the target object among all objects that have generated the interactive data to be processed for the content to be recognized;

[0017] The determination unit is further used to, if the first exposure reaches the exposure threshold and the target object concentration reaches the concentration threshold, determine that the content to be recognized is the target content.

[0018] On the one hand, an embodiment of the present application provides a computer device, and the computer device includes a processor and a memory:

[0019] The memory is used to store a computer program and transmit the computer program to the processor;

[0020] The processor is used to execute the method described in any of the foregoing aspects according to the instructions in the computer program.

[0021] On the one hand, an embodiment of the present application provides a computer-readable storage medium, and the computer-readable storage medium is used to store a computer program, and when the computer program is executed by a processor, the processor is caused to execute the method described in any of the foregoing aspects.

[0022] On the one hand, an embodiment of the present application provides a computer program product, including a computer program, which when executed by a processor implements the method described in any of the foregoing aspects.

[0023] As can be seen from the above technical solutions, when interacting with the content to be recognized, the to-be-processed interaction data generated by interacting with the content to be recognized is obtained. Then, a first exposure degree is determined for the content to be recognized corresponding to the to-be-processed interaction data, and the first exposure degree is obtained, which represents the number of times of interacting with the content to be recognized. The object type of the object to be recognized corresponding to the to-be-processed interaction data is recognized to obtain an object recognition result. The object to be recognized is the object that generates the to-be-processed interaction data, and the object recognition result can reflect whether the object to be recognized is a target object. The target object is usually the object that produces, disseminates, and consumes the target content. If the exposure of a content is caused by a sufficient number of target objects interacting with it, then this content is very likely to be target content, that is, malicious content with exposure risk. Therefore, if it is determined based on the object recognition result that the object to be recognized is a target object, then the target object concentration of the content to be recognized is determined, and the target object concentration represents the frequency of occurrence of target objects among all the objects that have generated the to-be-processed interaction data of the content to be recognized, so as to combine the first exposure degree and the target object concentration to determine whether the exposure of the content to be recognized is caused by a sufficient number of target objects interacting with it. If the first exposure degree reaches the exposure degree threshold and the target object concentration reaches the concentration threshold, it indicates that the content to be recognized has a high exposure degree and this exposure degree is caused by a sufficient number of target objects interacting with it, then the content to be recognized is determined to be target content. The present application combines the first exposure degree and the target object concentration to determine whether the content to be recognized is target content. If the target object concentration is high enough, it indicates that the exposure of the content to be recognized is caused by a sufficient number of target objects interacting with it, and the content to be recognized is very likely to be malicious content with exposure risk. Then, even if the set exposure degree threshold is relatively low, when manual intervention is carried out, there is a high probability of detecting more target content. Thus, to a certain extent, the set exposure degree threshold is lowered, and target content can be discovered at a lower exposure degree with the same human input, avoiding a large number of users coming into contact with inappropriate content due to too high an exposure degree threshold, and thus avoiding causing a greater impact. Description of the Drawings

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0025] Figure 1 An application scenario architecture diagram of a content recognition method provided by an embodiment of the present application;

[0026] Figure 2 A flowchart of a content recognition method provided by an embodiment of the present application;

[0027] Figure 3 An overall process schematic diagram of a content recognition method provided by an embodiment of the present application;

[0028] Figure 4 A structural diagram of a content recognition device provided by an embodiment of the present application;

[0029] Figure 5 A structural diagram of a terminal provided by an embodiment of the present application;

[0030] Figure 6 A structural diagram of a server provided by an embodiment of the present application. Detailed implementation manners

[0031] The embodiments of the present application will be described below with reference to the accompanying drawings.

[0032] Content recognition technology can discover and warn of potential malicious content exposure risks in real time by collecting, processing, and analyzing a large amount of content data produced by Internet users. Related technologies can discover target content based on the exposure degree, usually setting certain rules, and warning relevant personnel for review after reaching a certain exposure degree. Relevant personnel can discover malicious content with exposure risks on the product and formulate effective countermeasures in a timely manner to reduce the negative impact of malicious content on the product.

[0033] Based on analysis and actual verification, it can be known that most of the content with a relatively high exposure degree is normal content spread by ordinary users, and a small part is malicious content. If a relatively small exposure degree is set, for example, the exposure degree threshold is set to 100 (that is, the number of interactions reaches 100), then when manual intervention is carried out for review, among 1000 interaction data, there may be only one piece of content corresponding to the interaction data that is the target content. Limited by the labor cost and to avoid ineffective manual intervention, usually a relatively high exposure degree threshold is set. For example, the exposure degree threshold is set to 10,000 (that is, the number of interactions reaches 10,000), then when manual intervention is carried out for review, among 1000 interaction data, it may be found that a large number of content corresponding to the interaction data is the target content. In this way, the effect of manual intervention to discover the target content is obvious, and the waste of manpower is avoided.

[0034] However, often after the exposure degree reaches a relatively high exposure degree threshold and then manual intervention is carried out, a large number of users have already been exposed to inappropriate content, causing a greater impact.

[0035] To solve the above technical problems, an embodiment of the present application provides a content recognition method. This method determines whether the content to be recognized is target content by combining the first exposure level and the target object concentration. If the target object concentration is high enough, it indicates that the exposure of the content to be recognized is caused by sufficient interactions of a large number of target objects. The content to be recognized is very likely to be malicious content with exposure risks. Then, even if the exposure level threshold is set relatively low, when manual intervention is involved, there is a high probability of detecting a relatively large number of target contents. Thus, to a certain extent, the set exposure level threshold can be reduced, and target contents can be discovered at a lower exposure level with the same human input, so as to prevent and control earlier, and avoid a large number of users being exposed to inappropriate content due to too high an exposure level threshold, thereby avoiding a greater impact.

[0036] It should be noted that the content recognition method provided by the embodiment of the present application can be applied to various scenarios such as cloud technology, artificial intelligence, social media platforms, e-commerce platforms, etc. There are various types of content in these scenarios, and these contents may be target contents. To prevent target contents from being exposed to a large number of users and causing adverse effects, the method provided by the embodiment of the present application can be used to identify target contents, so as to prevent and control. The embodiment of the present application will be mainly introduced by taking social media platforms and e-commerce platforms as examples.

[0037] In the scenario of a social media platform, the method provided by the embodiment of the present application is used for public exposure risk control. Specifically, the method provided by the embodiment of the present application can help identify and discover target contents. The target contents can be malicious contents with exposure risks, so as to provide real-time risk assessment and monitoring for the publicly exposed contents on the social media platform, and help the social media platform take corresponding measures to reduce the greater exposure risk of malicious contents through early warnings. In this scenario, malicious contents can be contents that violate laws and social ethics, including but not limited to contents involving violations and violence, etc.

[0038] In the scenario of an e-commerce platform, the method provided by the embodiment of the present application can be used for security management. Specifically, the method provided by the embodiment of the present application can identify target contents, provide real-time risk assessment and monitoring for the e-commerce platform, and help the e-commerce platform take appropriate measures to ensure the security and credibility of transactions through early warnings. In this scenario, malicious contents can be illegal commodities, malicious comments and other illegal contents.

[0039] It should be noted that the content recognition method provided in the embodiments of the present application can be executed by a computer device, which can be, for example, a server or a terminal. The server can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The terminal includes, but is not limited to, smart phones, computers, intelligent voice interaction devices, smart home appliances, vehicle-mounted terminals, aircraft, etc.

[0040] As Figure 1 shown, Figure 1 FIG. shows an application scenario architecture diagram of a content recognition method. The application scenario architecture diagram may include a terminal 100 and a server 200. The terminal 100 can provide content for an interaction object, and the interaction object can interact with the content through the terminal 100. Among them, the interaction object can be an object that interacts with the content by performing an interaction operation. The interaction object can be, for example, a user. The content can be various types of content, such as articles, videos, etc. Interaction can refer to a consumption behavior of consuming the content, such as clicking, browsing, forwarding, liking, commenting, collecting, etc.

[0041] The terminal 100 can provide content for the interaction object through various content providing platforms. The content providing platform can be a platform that provides content for the interaction object, such as a social media platform, an e-commerce platform, etc. The content provided by different content providing platforms may also be different. For example, the content on the social media platform may be articles, and the content on the e-commerce platform may be all products, comments, etc. In this case, the server 200 can be a server that provides services for the content providing platform.

[0042] When the interaction object interacts with a certain content through the terminal 100, the terminal 100 can generate interaction data based on the interaction. In order to determine whether the content is a target content, the content can be used as the content to be recognized, and the generated interaction data can be used as the interaction data to be processed. The server 200 can obtain the interaction data to be processed generated by interacting with the content to be recognized. Among them, the content to be recognized is the content targeted by the interaction and needs to be judged whether it is a target content. The interaction data to be processed is the interaction data generated by interacting with the content to be recognized.

[0043] Then, the server 200 determines a first exposure degree for the content to be recognized corresponding to the interaction data to be processed, and obtains a first exposure degree, where the first exposure degree represents the number of times of interacting with the content to be recognized. And the server 200 performs object type recognition on the object to be recognized corresponding to the interaction data to be processed, and obtains an object recognition result. The object to be recognized is the object that generates the interaction data to be processed, and the object recognition result can reflect whether the object to be recognized is a target object.

[0044] The target object is usually the object that produces, disseminates, and consumes the target content. If the exposure of a piece of content is caused by a sufficient number of target objects interacting with it, then it is very likely that the content is the target content, that is, malicious content with exposure risk. Therefore, if the server 200 determines that the object to be recognized is a target object based on the object recognition result, it determines the target object concentration of the content to be recognized. The target object concentration represents the frequency of occurrence of target objects among all the objects that have generated the interaction data to be processed for the content to be recognized. The higher the target object concentration, the more likely it is that the content to be recognized is the target content. Therefore, the first exposure degree and the target object concentration can be combined to determine whether the exposure of the content to be recognized is caused by a sufficient number of target objects interacting with it.

[0045] If the first exposure degree reaches the exposure degree threshold and the target object concentration reaches the concentration threshold, it means that the content to be recognized has a high exposure degree and this exposure degree is caused by a sufficient number of target objects interacting with it. Then the server 200 determines that the content to be recognized is the target content. Among them, the target content can be malicious content with exposure risk. Different content providing platforms may have different malicious contents. For example, when the content providing platform is a social media platform, the malicious content can be content involving violations or violence, etc.; when the content providing platform is an e-commerce platform, the malicious content can be illegal contents such as illegal goods and malicious comments.

[0046] This application combines the first exposure degree and the target object concentration to determine whether the content to be recognized is the target content. If the target object concentration is high enough, it means that the exposure of the content to be recognized is caused by a sufficient number of target objects interacting with it. The content to be recognized is very likely to be malicious content with exposure risk. Then, even if the exposure degree threshold is set relatively low, when manual intervention is carried out, it is very likely to check out more target contents. Thus, to a certain extent, the set exposure degree threshold can be reduced, and under the same human input, target contents can be discovered at a lower exposure degree, avoiding a large number of users coming into contact with inappropriate content due to too high an exposure degree threshold, and thus avoiding causing a greater impact.

[0047] It can be understood that Figure 1 Taking the server executing the content recognition method as an example does not limit the embodiments of this application. It can also be the terminal executing the content recognition method, or the terminal and the server cooperating to execute the content recognition method.

[0048] The method provided by the embodiments of the present application may involve artificial intelligence technology, and content recognition is automatically performed through artificial intelligence technology. Artificial Intelligence (AI) is to use a digital computer or a machine controlled by a digital computer to simulate, extend, and expand human intelligence, and is a theory, method, technology, and application system that can perceive the environment, acquire knowledge, and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a way similar to human intelligence. Artificial intelligence also studies the design principles and implementation methods of various intelligent machines to enable the machines to have the functions of perception, reasoning, and decision-making.

[0049] Artificial intelligence technology is a comprehensive discipline that involves a wide range of fields, including both hardware-level and software-level technologies. Artificial intelligence basic technologies generally include, for example, sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, pre-trained model technology, operation / interaction systems, mechatronics, etc. Among them, the pre-trained model, also known as the large model or the base model, can be widely applied to downstream tasks in various directions of artificial intelligence after fine-tuning. Artificial intelligence software technology mainly includes several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0050] It should be noted that in the specific implementation manner of the present application, relevant data such as user information may be involved throughout the process. When the above embodiments of the present application are applied to specific products or technologies, the user's separate consent or separate permission needs to be obtained, and the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of the relevant countries and regions.

[0051] Next, taking the computer device as a server as an example, the content recognition method provided by the embodiments of the present application will be introduced in conjunction with the accompanying drawings. Refer to Figure 2 , Figure 2 which shows a flowchart of a content recognition method, and the method includes:

[0052] S201. Obtain the interaction data to be processed, where the interaction data to be processed is generated by interacting with the content to be recognized.

[0053] The terminal provides content to the interactive object, and the interactive object can interact with the content through the terminal. The terminal can provide content to the interactive object through various content provision platforms. The content provision platform can be a platform that provides content to the interactive object, such as a social media platform or an e-commerce platform. The content provided by different content provision platforms may also vary. For example, the content on a social media platform may be articles, while the content on an e-commerce platform may be products and reviews. In this case, the server can be the server providing services to the content provision platform.

[0054] Content provision platforms are products used on terminals. If the content on them is target content, widespread dissemination of such content on the product could lead to user churn due to user discomfort, thereby harming product development. Therefore, content identification is necessary to identify content and uncover content risks. Content risk can refer to content that violates laws and social order and good morals, including but not limited to content involving violations of laws, violence, etc. Widespread dissemination of such content on the product could lead to user churn due to user discomfort, thereby harming product development. Therefore, discovering content risk can mean identifying whether the content is target content, thereby promptly discovering whether it possesses the aforementioned content risks. If the content is target content, it indicates that the aforementioned content risks have been discovered.

[0055] Therefore, when an interactive object interacts with a piece of content through a terminal, the terminal can generate interaction data based on this interaction. The interactive object can be, for example, a user. The content can be of various types, such as articles, videos, and so on. Interaction can refer to the consumption of content, including, for example, clicks, browsing, forwarding, liking, commenting, and adding to favorites.

[0056] To determine whether the content is the target content, the content can be identified as the content to be identified, and the generated interaction data can be used as the interaction data to be processed. The server can obtain the interaction data to be processed generated by interacting with the content to be identified. The content to be identified is the content that the interaction is targeting and needs to be determined to be the target content. The interaction data to be processed is the interaction data generated by interacting with the content to be identified.

[0057] It should be noted that the interaction objects can continuously interact with the content on the content providing platform. Therefore, within a certain time period, all interaction data of interacting with the content can be collected in a streaming manner. The interaction data of interacting with the same content can form a consumption behavior sequence. The consumption behavior sequence can refer to the sequence formed by all interaction objects that have a certain consumption behavior (such as click, browse, forward, like, comment, collect, etc.) on the same content within a specific time period. For example, the users who browse an open article on the same day form the consumption behavior sequence of that article on that day. Among them, the collection of interaction data can be achieved through a data collection module, and the process of collecting interaction data can be called the data collection process.

[0058] After obtaining each piece of interaction data, the interaction data can be used as the to-be-processed interaction data, and content recognition can be performed in real time through the method provided in the embodiments of the present application.

[0059] In some cases, some interactions may not be out of the will of the interaction object. These interactions are usually due to the misoperations of the interaction object, which makes the corresponding interaction data difficult to reflect the characteristics of the corresponding object. Such interactions are invalid interactions or low-quality interactions, and the interaction data generated by such interactions is a kind of noise data. Therefore, after collecting the interaction data, the collected interaction data can be denoised to remove the noise data, so as to perform the subsequent content recognition steps on the interaction data of non-noise data. That is to say, in this case, the to-be-processed interaction data is non-noise data. If an interaction data is noise data, the interaction data will be removed and the subsequent content recognition steps will not be performed on the interaction data; if an interaction data is not noise data, the interaction data can be used as the to-be-processed interaction data for subsequent content recognition steps.

[0060] It should be noted that the step of denoising can also be called preprocessing, and the step of denoising can be implemented through a preprocessing module. Through denoising, the interaction data generated by invalid interactions and low-quality interactions can be removed, ensuring the accuracy of subsequent content recognition.

[0061] S202. Determine the first exposure degree of the to-be-identified content corresponding to the to-be-processed interaction data to obtain the first exposure degree of the to-be-identified content, and identify the object type of the to-be-identified object corresponding to the to-be-processed interaction data to obtain an object recognition result.

[0062] To determine whether the content is malicious content with exposure risk, it is necessary to determine whether the content is malicious content and whether the content has exposure risk. In the embodiments of the present application, whether the content has exposure risk can be reflected by the exposure degree (such as the first exposure degree), and whether the content is malicious content can be reflected by whether the object generating interaction data for the content is the target object and the target object concentration. Based on this, after obtaining the to-be-processed interaction data, the first exposure degree of the to-be-identified content corresponding to the to-be-processed interaction data can be determined to obtain the first exposure degree of the to-be-identified content, and the object type of the to-be-identified object corresponding to the to-be-processed interaction data can be identified to obtain the object identification result.

[0063] The first exposure degree can represent the number of times of interacting with the to-be-identified content. In the embodiments of the present application, for the same content, each time an interaction is performed on the content, the number of times of interacting with the content can be accumulated as the first exposure degree of the content, and the first exposure degree of the content can be stored. When storing, it can be stored in the form of the corresponding relationship between the content identifier and the first exposure degree, specifically in the form of key-Value (KV). For example, the content identifier can be used as K and the first exposure degree can be used as V. Among them, the content identifier can be an identity number (Identity, ID), a hash value calculated based on the Message-Digest Algorithm (MD) (which can also be called the MD value). The message-digest algorithm can be, for example, the fifth-generation message-digest algorithm (message-digest algorithm 5, MD5), the fourth-generation message-digest algorithm (message-digest algorithm 4, MD4), etc. The embodiments of the present application do not make any limitations on this.

[0064] In a possible implementation manner, the first exposure degree can be stored in the first exposure degree database. At this time, the method for determining the first exposure degree of the to-be-identified content corresponding to the to-be-processed interaction data to obtain the first exposure degree of the to-be-identified content can be to obtain the content identifier of the to-be-identified content, and then query in the first exposure degree database based on the content identifier to obtain the first exposure degree of the to-be-identified content. The first exposure degree database is used to store the first exposure degrees of the content indicated by different content identifiers.

[0065] If the first exposure degree is stored in the KV form, the content identifier of the to-be-identified content can be used as K, and the first exposure degree database can be queried in real time. The content identifier of the to-be-identified content is found in the first exposure degree database, and the first exposure degree corresponding to the content identifier of the to-be-identified content is determined as the first exposure degree of the to-be-identified content.

[0066] In the embodiment of the present application, the first exposure of the content to be recognized is obtained by searching the database in real time, which can ensure the real-time nature of obtaining the first exposure, and then improve the efficiency of content recognition, so as to timely discover and control content risks.

[0067] In the embodiment of the present application, the database used (such as the first exposure database) can be various types of databases. For example, it can be a Remote Dictionary Server (Redis). Redis is an open-source KV storage system with fast search speed, which is convenient for realizing real-time search. When the database used is Redis, the first exposure database can be called the first exposure Redis.

[0068] It can be understood that before the generation of the to-be-processed interaction data this time, the first exposure stored in the first exposure database is the first exposure of the content to be recognized before the generation of the to-be-processed interaction data this time. When the to-be-processed interaction data is obtained this time, the number of interactions for the content to be recognized increases by one, that is, the first exposure of the content to be recognized also increases by 1 accordingly. Therefore, in order to ensure the accuracy of the first exposure stored in the first exposure database, the first exposure of the content to be recognized stored in the first exposure database can be updated based on the content identifier. Thus, the accuracy of the first exposure stored in the first exposure database is ensured.

[0069] It should be noted that the embodiment of the present application does not limit the order of the two steps of update and query. For example, it can be queried first and then updated. At this time, the first exposure of the content to be recognized before the generation of the to-be-processed interaction data is queried from the first exposure database based on the content identifier. In this way, the first exposure of the content to be recognized corresponding to the to-be-processed interaction data can be the queried first exposure plus 1. Then, the obtained first exposure of the content to be recognized corresponding to the to-be-processed interaction data is updated to the first exposure database. Another example is that it can be updated first and then queried. At this time, after the to-be-processed interaction data is obtained, an update instruction can be sent to the first exposure database to control the first exposure database to update the first exposure of the content to be recognized, and then the first exposure of the content to be recognized is directly queried from the first exposure database.

[0070] The object to be recognized is the object that generates the to-be-processed interaction data. In the embodiment of the present application, an object list can be stored in advance. The object list is a list composed of objects belonging to the target object. What can be stored in the object list is the object identifier, and the object identifier can be an ID. The embodiment of the present application does not limit this.

[0071] In a possible implementation, the object list can be stored in an object database. At this time, when identifying the object type of the object to be processed corresponding to the interaction data, the way to obtain the object recognition result can be to obtain the object identifier of the object to be recognized, and then query based on the object identifier in the object list stored in the object database to obtain the object recognition result. The object recognition result is used to reflect whether the object to be recognized is the target object.

[0072] When performing the search, the object identifier can be used as K to query the object list in real time, and it is judged whether the object identifier exists in the object list to obtain the object recognition result. The object recognition result is whether the object exists in the object list, thereby reflecting whether the object to be recognized is the target object.

[0073] In the embodiment of the present application, the object recognition result is obtained by searching the database in real time, without the need to construct an object recognition model, which can ensure the real-time nature of determining the object recognition result, and further improve the efficiency of content recognition, so as to timely discover and control content risks.

[0074] In the embodiment of the present application, the database used (such as the object database) can be various types of databases, for example, it can be Redis. When the database used is Redis, the object database can be called object Redis.

[0075] It should be noted that the objects in the object list can be custom-set by the user (content providing platform). Usually, they come from (but are not limited to) external clues of the content providing platform and user complaints. When the user obtains relevant malicious content from different channels, they can, through diffusion and other means, for example, for the objects that have generated interaction data for the corresponding malicious content, form an object list, and then write the object list into the object database. That is to say, the objects in the object list come from the clue system, the diffusion system, the reporting system, etc.

[0076] It can be understood that the object to be recognized is the object that generates the interactive data to be processed, reflecting the source of the consumption behavior corresponding to the interactive data to be processed. In the embodiments of the present application, the source of the consumption behavior can be reflected in different dimensions. For example, the interactive object (user) that conducts the interaction, the device address to be recognized, or the Internet Protocol (IP) address to be recognized. That is to say, the object to be recognized can be the interactive object (such as a user) that executes the interactive operation, or the interactive object to be recognized, the device address to be recognized, or the Internet Protocol (IP) address to be recognized. When the object to be recognized is the interactive object to be recognized, the target object is the target interactive object, and the objects in the object list are the interactive objects belonging to the target interactive object; when the object to be recognized is the device address to be recognized, the target object is the target device address, and the objects in the object list are the device addresses belonging to the target device address; when the object to be recognized is the Internet Protocol address to be recognized, the target object is the target Internet Protocol address, and the objects in the object list are the Internet Protocol addresses belonging to the target Internet Protocol address.

[0077] In the embodiments of the present application, the case where the object to be recognized is the interactive object to be recognized and the target object is the target interactive object is mainly introduced. If the interactive object is a user, the interactive object to be recognized is the user to be recognized. At this time, the target object can be the target user, and the target user can also be referred to as a specific population. The specific population can refer to users collected through the clue system, the diffusion system, and the reporting system for a certain type of specific malicious content, including but not limited to producers (who produce malicious content), disseminators (who spread malicious content), and consumers (who consume malicious content). Similar malicious content will be generated and spread by these users.

[0078] S203. If it is determined based on the object recognition result that the object to be recognized is the target object, determine the target object concentration of the content to be recognized.

[0079] After obtaining the object recognition result, it can be determined based on the object recognition result whether the object to be recognized is the target object. If it is determined that the object to be recognized is the target object, then determine the target object concentration of the content to be recognized. The target object concentration represents the frequency of occurrence of the target object among all the objects that have generated the interactive data to be processed for the content to be recognized. If it is determined that the object to be recognized is not the target object, then end the processing flow.

[0080] If the object recognition result is obtained by querying the object list stored in the object database, the method for determining that the object to be recognized is the target object based on the object recognition result can be that if the object recognition result indicates that the object to be recognized exists in the object list, determine that the object to be recognized is the target object. If the object recognition result indicates that the object to be recognized is not found in the object list, then determine that the object to be recognized is not the target object.

[0081] The frequency of occurrence of the target object among all objects that have generated the interaction data to be processed with the content to be recognized can be represented by the proportion of the number of times the target object has interacted with the content to be recognized in the total number of times. The total number of times can be the number of times of interacting with the content to be recognized. Therefore, a way to determine the target object concentration of the content to be recognized can be to determine the second exposure of the content to be recognized, and obtain the second exposure of the content to be recognized. The second exposure represents the number of times the target object has interacted with the content to be recognized. Then, the ratio between the second exposure and the first exposure is determined as the target object concentration. That is, the target object concentration = second exposure / first exposure, and the value range of this target object concentration is between 0 and 1.

[0082] In the embodiments of the present application, for the same content, every time the target object generates an interaction with the content, the number of times the target object has interacted with the content can be accumulated as the second exposure of the content, and the second exposure of the content is stored. When storing, it can be stored in the form of the corresponding relationship between the content identifier of the content and the second exposure, specifically, it can be stored in the KV form. For example, the content identifier can be used as K, and the second exposure can be used as V.

[0083] In a possible implementation manner, the second exposure can be stored in the second exposure database. At this time, a way to determine the second exposure of the content to be recognized and obtain the second exposure of the content to be recognized can be to obtain the content identifier of the content to be recognized, and then query in the second exposure database based on the content identifier to obtain the second exposure. The second exposure database is used to store the second exposure of the content indicated by different content identifiers.

[0084] If the second exposure is stored in the KV form, the content identifier of the content to be recognized can be used as K, and the second exposure database can be queried in real time. The content identifier of the content to be recognized is found in the second exposure database, and the second exposure corresponding to the content identifier of the content to be recognized is determined as the second exposure of the content to be recognized.

[0085] In the embodiments of the present application, the second exposure of the content to be recognized is obtained by the way of real-time searching in the database, which can ensure the real-time nature of obtaining the second exposure, and further improve the efficiency of content recognition, so as to timely discover and control content risks.

[0086] In the embodiments of the present application, the database used (such as the second exposure database) can be various types of databases, for example, it can be Redis. When the database used is Redis, the second exposure database can be called the second exposure Redis.

[0087] It can be understood that before it is determined that the interaction data to be processed comes from the target object, what is stored in the second exposure database is the second exposure of the content to be recognized before it is determined that the interaction data to be processed comes from the target object. When it is determined that the interaction data to be processed comes from the target object, the number of times the target object interacts with the content to be recognized is increased by one, that is, the second exposure of the content to be recognized is also increased by 1 accordingly. Therefore, in order to ensure the accuracy of the second exposure stored in the second exposure database, the second exposure of the content to be recognized stored in the second exposure database can be updated based on the content identifier. Thus, the accuracy of the second exposure stored in the second exposure database is ensured.

[0088] It should be noted that similar to the update and query of the aforementioned first exposure database, the order of update and query of the second exposure database is not limited either, and will not be elaborated here.

[0089] S204. If the first exposure reaches the exposure threshold and the target object concentration reaches the concentration threshold, determine that the content to be recognized is the target content.

[0090] After obtaining the first exposure and the target object concentration, it is possible to determine whether the content to be recognized is the target content by combining the first exposure and the target object concentration. In the embodiment of the present application, if the first exposure reaches the exposure threshold and the target object concentration reaches the concentration threshold, it means that the content to be recognized has a large exposure, and this exposure is caused by sufficient target objects interacting, then it is determined that the content to be recognized is the target content. Among them, the exposure threshold and the concentration threshold can be set according to actual needs, and the specific values of the exposure threshold and the concentration threshold are not limited in the embodiment of the present application. Thus, in the entire chain from the production to the consumption of a target content, due to the abnormality of the target object concentration and the first exposure, it is timely discovered.

[0091] After determining that the content to be recognized is the target content, the server can also perform content disposal on the content to be recognized, and the content disposal is used to restrict the interaction object from interacting with the content to be recognized. Restricting the interaction object from interacting with the content to be recognized includes, but is not limited to, restricting the interaction object from browsing, forwarding, etc., to avoid continued exposure.

[0092] It should be noted that when performing content disposal, the interaction with the content to be recognized this time can be restricted, making this interaction fail. For example, if the interaction with the content to be recognized this time is forwarding, after the interaction object clicks "forward", it can finally feedback that the forwarding fails, thereby realizing the interaction restriction function. Of course, it is also possible to restrict other interactions after the interaction with the content to be recognized this time, and the embodiment of the present application does not limit this.

[0093] In a possible implementation, the way to restrict the interaction object from interacting with the content to be recognized can be to cancel the interaction function for interacting with the content to be recognized. For example, when the interaction is forwarding, the forwarding control is gray or there is no forwarding control, thus restricting forwarding. The way to restrict the interaction object from interacting with the content to be recognized can also be to directly delete the content to be recognized, so that it is impossible to continue interacting with the content to be recognized. Based on this, the way to perform content disposal on the content to be recognized can be to cancel the interaction function for interacting with the content to be recognized, delete the content to be recognized, or at least one of them.

[0094] In the embodiment of the present application, it can be automatically determined whether the content to be recognized is the target content through S201 - S204. To avoid misrecognition, it can also be re-verified through manual review after automatically determining that the content to be recognized is the target content. Therefore, in a possible implementation, the way to perform content disposal on the content to be recognized can be to send the content to be recognized to the verification platform for re-verification. The verification platform can provide the content to be recognized to the reviewers, and the reviewers will judge the content to be recognized. If the reviewers determine that the content to be recognized is the target content, the verification platform will return a verification passed notification, which indicates that the reviewers have determined that the content to be recognized is the target content. After receiving the verification passed notification returned by the verification platform, the server can perform content disposal on the content to be recognized.

[0095] The above method of re-verifying through manual review can further improve the accuracy of content recognition, avoid unnecessary content disposal caused by misrecognition, and affect the normal interaction with normal content.

[0096] In a possible implementation, if the content to be recognized is the target content, other objects that have interacted with the content to be recognized may also belong to the target objects that produce, spread, and consume the target content. Therefore, after determining that the content to be recognized is the target content, the server can also obtain the historical interaction data generated by interacting with the content to be recognized, and update the object list based on the historical objects corresponding to the historical interaction data. This can ensure that the object list covers the target objects more comprehensively, thus avoiding missed recognition of target objects and further affecting the timely discovery of target content.

[0097] In a possible implementation, after determining that the content to be recognized is the target content, the server can also issue a warning prompt to indicate that the content to be recognized is the target content for timely processing.

[0098] As can be seen from the above technical solution, when interacting with the content to be recognized, the to-be-processed interaction data generated by the interaction with the content to be recognized is obtained. Then, the first exposure degree of the content to be recognized corresponding to the to-be-processed interaction data is determined to obtain the first exposure degree, where the first exposure degree represents the number of times of interacting with the content to be recognized. The object type of the object to be recognized corresponding to the to-be-processed interaction data is recognized to obtain the object recognition result. The object to be recognized is the object that generates the to-be-processed interaction data, and the object recognition result can reflect whether the object to be recognized is the target object. The target object is usually the object that produces, disseminates, and consumes the target content. If the exposure of a content is caused by a sufficient number of target objects interacting with it, then this content is very likely to be the target content, that is, malicious content with exposure risk. Therefore, if it is determined based on the object recognition result that the object to be recognized is the target object, then the target object concentration of the content to be recognized is determined, where the target object concentration represents the frequency of occurrence of the target object among all the objects that have generated the to-be-processed interaction data of the content to be recognized, so as to combine the first exposure degree and the target object concentration to determine whether the exposure of the content to be recognized is caused by a sufficient number of target objects interacting with it. If the first exposure degree reaches the exposure degree threshold and the target object concentration reaches the concentration threshold, it indicates that the content to be recognized has a high exposure degree and this exposure degree is caused by a sufficient number of target objects interacting with it, then it is determined that the content to be recognized is the target content. This application combines the first exposure degree and the target object concentration to determine whether the content to be recognized is the target content. If the target object concentration is high enough, it indicates that the exposure of the content to be recognized is caused by a sufficient number of target objects interacting with it, and the content to be recognized is very likely to be malicious content with exposure risk. Then, even if the exposure degree threshold is set relatively low, when manual intervention is carried out, there is a high probability of detecting a relatively large number of target contents. Thus, to a certain extent, the set exposure degree threshold can be reduced, and under the same human input, target contents can be discovered at a lower exposure degree, avoiding a large number of users coming into contact with inappropriate content due to too high an exposure degree threshold, and thereby avoiding causing a greater impact.

[0099] Verification shows that for the target content, compared with the content recognition technology based on pure exposure degree, the method provided by the embodiments of this application can newly discover 300% more target contents under the same human input, and can reduce the target content from the original high exposure degree to 2%.

[0100] Compared with determining target content based on a content recognition model or an object recognition model provided by related technologies, constructing a content recognition model or an object recognition model often requires a large amount of high-quality labeled data and complex feature engineering. For new risk patterns, the model (such as a content recognition model or an object recognition model) needs to be retrained. In addition, once adversarial content appears, it can easily bypass the model, resulting in missed recognition. The reason why the target content has continuous high exposure is that the model provided by related technologies covers less than 1% of the target content, resulting in it being unable to be processed in the corresponding production link. However, the embodiments of the present application do not need to construct a model, and the recognition method is simpler. And since there is no need to use a model, it can be quickly launched for prevention and control against new risk patterns, and strong adversarial content can be avoided from bypassing by decoupling from the model.

[0101] The above embodiments introduce a content recognition method, and after discovering the target content, content disposal is performed on the target content. Next, the overall process of the content recognition method provided by the embodiments of the present application will be introduced in combination with an actual application scenario. Refer to Figure 3 As shown, this method can be integrated into a content recognition and disposal system and deployed on a content providing platform. The content recognition and disposal system can be referred to Figure 3 as shown in 301 in. In addition, a storage system 302 and an object list operation system 303 are also provided. The storage system 302 may include a first exposure database 3021, an object database 3022, and a second exposure database 3023. The object list operation system 303 can obtain an object list 3034 based on a lead system 3031, a diffusion system 3032, and a reporting system 3033, and write the object list 3034 into the object database 3022 in real time.

[0102] When performing content recognition, interaction data can be collected (see S301), and the interaction data can be preprocessed (see S302). If the interaction data is not noise data, the interaction data can be used as the interaction data to be processed, so as to obtain the interaction data to be processed. Determine the first exposure degree of the content to be recognized corresponding to the interaction data to be processed (see S303) to obtain the first exposure degree. The first exposure degree can be obtained by querying the first exposure degree database based on the content identifier. After obtaining the first exposure degree, the first exposure degree of the content to be recognized in the first exposure degree database can also be updated. Then, identify the object type of the object to be recognized corresponding to the interaction data to be processed (see S304) to obtain the object recognition result. Determine whether the object to be recognized is the target object based on the object recognition result. The object type recognition can be achieved by querying the object list in the object database. If it is determined that the object to be recognized is the target object based on the object recognition result, the second exposure degree can be determined (see S305). The second exposure degree can be the number of times the target object interacts with the content to be recognized. The second exposure degree can be obtained by querying the second exposure degree database based on the content identifier. After obtaining the second exposure degree, the second exposure degree of the content to be recognized in the second exposure degree database can also be updated. Then, determine the target object concentration (see S306). The target object concentration can be the ratio between the second exposure degree and the first exposure degree. If the first exposure degree reaches the exposure degree threshold and the target object concentration reaches the concentration threshold, determine that the content to be recognized is the target content and issue a warning prompt (see S307). Send the content to be recognized to the verification platform so that the reviewer can judge the content to be recognized, thereby realizing manual review (see S308). If the reviewer determines that the content to be recognized is the target content, the verification platform will return a verification passed notice, and the verification passed notice indicates that the reviewer determines that the content to be recognized is the target content. After the server receives the verification passed notice returned by the verification platform, it can perform content disposal on the content to be recognized (see S309).

[0103] It should be noted that, based on the implementation manners provided in the above aspects of the present application, further combinations can be made to provide more implementation manners.

[0104] Based on the content recognition method provided in the foregoing embodiments, an embodiment of the present application further provides a content recognition device 400. As shown in Figure 4 the content recognition device 400 includes an acquisition unit 401, a determination unit 402, and an identification unit 403:

[0105] The acquisition unit 401 is configured to acquire interaction data to be processed, where the interaction data to be processed is generated by interacting with the content to be recognized;

[0106] The determining unit 402 is configured to determine a first exposure degree of the to-be-recognized content corresponding to the to-be-processed interaction data, so as to obtain the first exposure degree of the to-be-recognized content, where the first exposure degree represents the number of times of interacting with the to-be-recognized content;

[0107] The recognizing unit 403 is configured to perform object type recognition on the to-be-recognized object corresponding to the to-be-processed interaction data, so as to obtain an object recognition result, where the to-be-recognized object is the object that generates the to-be-processed interaction data;

[0108] The determining unit 402 is further configured to determine a target object concentration of the to-be-recognized content if it is determined based on the object recognition result that the to-be-recognized object is a target object, where the target object concentration represents the frequency of occurrence of the target object among all the objects that have generated the to-be-processed interaction data corresponding to the to-be-recognized content;

[0109] The determining unit 402 is further configured to determine that the to-be-recognized content is a target content if the first exposure degree reaches an exposure degree threshold and the target object concentration reaches a concentration threshold.

[0110] In a possible implementation manner, the determining unit 402 is configured to:

[0111] Obtain a content identifier of the to-be-recognized content;

[0112] Query in a first exposure degree database based on the content identifier, so as to obtain the first exposure degree of the to-be-recognized content, where the first exposure degree database is used to store the first exposure degrees of the contents indicated by different content identifiers.

[0113] In a possible implementation manner, the apparatus further includes an updating unit:

[0114] The updating unit is configured to update the first exposure degree of the to-be-recognized content stored in the first exposure degree database based on the content identifier.

[0115] In a possible implementation manner, the recognizing unit 403 is configured to:

[0116] Obtain an object identifier of the to-be-recognized object;

[0117] Query in an object list stored in an object database based on the object identifier, so as to obtain the object recognition result, where the object list is a list composed of objects belonging to the target object;

[0118] The determining unit 402 is configured to determine that the to-be-recognized object is the target object if the object recognition result indicates that the to-be-recognized object exists in the object list.

[0119] In a possible implementation, the device further includes an updating unit:

[0120] The obtaining unit is further configured to obtain historical interaction data generated by interacting with the content to be recognized after determining that the content to be recognized is the target content;

[0121] The updating unit is configured to update the object list based on the historical object corresponding to the historical interaction data.

[0122] In a possible implementation, the object to be recognized is an object to be recognized for interaction, a device address to be recognized, or an Internet Protocol address to be recognized;

[0123] When the object to be recognized is an object to be recognized for interaction, the target object is a target interaction object, and the objects in the object list are interaction objects belonging to the target interaction object;

[0124] When the object to be recognized is a device address to be recognized, the target object is a target device address, and the objects in the object list are device addresses belonging to the target device address;

[0125] When the object to be recognized is an Internet Protocol address to be recognized, the target object is a target Internet Protocol address, and the objects in the object list are Internet Protocol addresses belonging to the target Internet Protocol address.

[0126] In a possible implementation, the determining unit 402 is configured to:

[0127] Determine a second exposure degree of the content to be recognized to obtain the second exposure degree of the content to be recognized, where the second exposure degree represents the number of times of interacting with the content to be recognized through the target object;

[0128] Determine the ratio between the second exposure degree and the first exposure degree as the target object concentration.

[0129] In a possible implementation, the determining unit 402 is configured to:

[0130] Obtain a content identifier of the content to be recognized;

[0131] Query in a second exposure degree database based on the content identifier to obtain the second exposure degree, where the second exposure degree database is used to store the second exposure degrees of the contents indicated by different content identifiers.

[0132] In a possible implementation, the device further includes an updating unit:

[0133] The updating unit is configured to update the second exposure of the content to be recognized stored in the second exposure database based on the content identifier.

[0134] In a possible implementation, the apparatus further includes a handling unit:

[0135] The handling unit is configured to perform content handling on the content to be recognized after determining that the content to be recognized is a target content, and the content handling is used to restrict an interaction object from interacting with the content to be recognized.

[0136] In a possible implementation, the handling unit is configured to implement at least one of the following:

[0137] Cancel the interaction function for interacting with the content to be recognized;

[0138] Or, delete the content to be recognized.

[0139] In a possible implementation, the handling unit is configured to:

[0140] Send the content to be recognized to a verification platform;

[0141] If a verification passed notification returned by the verification platform is received, perform content handling on the content to be recognized.

[0142] As can be seen from the above technical solution, when interacting with the content to be recognized, the to-be-processed interaction data generated by the interaction with the content to be recognized is obtained. Then, a first exposure degree is determined for the content to be recognized corresponding to the to-be-processed interaction data, and the first exposure degree is obtained. The first exposure degree represents the number of times of interacting with the content to be recognized. The object type of the object to be recognized corresponding to the to-be-processed interaction data is recognized, and an object recognition result is obtained. The object to be recognized is the object that generates the to-be-processed interaction data, and the object recognition result can reflect whether the object to be recognized is the target object. The target object is usually the object that produces, disseminates, and consumes the target content. If the exposure of a content is caused by enough target objects interacting with it, then this content is very likely to be the target content, that is, malicious content with exposure risk. Therefore, if it is determined based on the object recognition result that the object to be recognized is the target object, then the target object concentration of the content to be recognized is determined. The target object concentration represents the frequency of the target object appearing among all the objects that have generated the to-be-processed interaction data of the content to be recognized, so as to combine the first exposure degree and the target object concentration to determine whether the exposure of the content to be recognized is caused by enough target objects interacting with it. If the first exposure degree reaches the exposure degree threshold and the target object concentration reaches the concentration threshold, it means that the content to be recognized has a very high exposure degree and this exposure degree is caused by enough target objects interacting with it, then it is determined that the content to be recognized is the target content. This application combines the first exposure degree and the target object concentration to determine whether the content to be recognized is the target content. If the target object concentration is high enough, it means that the exposure of the content to be recognized is caused by enough target objects interacting with it, and the content to be recognized is very likely to be malicious content with exposure risk. Then, even if the exposure degree threshold is set relatively low, when manual intervention is carried out, it is very likely to check out more target content. Thus, to a certain extent, the set exposure degree threshold is reduced, and under the same human input, target content can be found at a lower exposure degree, avoiding a large number of users being exposed to inappropriate content due to too high an exposure degree threshold, and further avoiding causing a greater impact.

[0143] An embodiment of this application also provides a computer device, and this computer device can execute the content recognition method. This computer device can be a terminal. Figure 5 Shown is a structural diagram of a terminal provided by an embodiment of this application. In Figure 5 it, taking the terminal as a smart phone as an example:

[0144] Refer to Figure 5, the smart phone includes components such as a Radio Frequency (RF) circuit 510, a memory 520, an input unit 530, a display unit 540, sensors 550, an audio circuit 560, a Wi-Fi module 570, a processor 580, and a power supply 590. The input unit 530 may include a touch panel 531 and other input devices 532. The display unit 540 may include a display panel 541. The audio circuit 560 may include a speaker 561 and a microphone 562. It can be understood that Figure 5 the structure of the smart phone shown in

[0145] does not limit the smart phone, and it may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements. The memory 520 can be used to store software programs and modules. The processor 580 executes various functional applications and data processing of the smart phone by running the software programs and modules stored in the memory 520. The memory 520 mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the smart phone (such as audio data, a phone book, etc.). In addition, the memory 520 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices.

[0146] The processor 580 is the control center of the smart phone, connects various parts of the entire smart phone using various interfaces and lines, and executes various functions of the smart phone and processes data by running or executing the software programs and / or modules stored in the memory 520, and calling the data stored in the memory 520. Optionally, the processor 580 may include one or more processing units; preferably, the processor 580 may integrate an application processor and a modem processor. Among them, the application processor mainly processes the operating system, user interface, and application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor may not be integrated into the processor 580 either.

[0147] In this embodiment, the processor 580 in the smart phone can execute the content recognition method provided in each embodiment of this application.

[0148] The computer device provided in the embodiments of this application may also be a server. Please refer to Figure 6 as shown Figure 6This is the structural diagram of server 600 provided by the embodiments of the present application. Server 600 may vary significantly due to different configurations or performances, and may include one or more processors, such as Central Processing Units (CPU) 622, and a memory 632, and one or more storage media 630 (such as one or more mass storage devices) for storing application programs 642 or data 644. Among them, the memory 632 and the storage media 630 may be transient storage or persistent storage. The program stored in the storage media 630 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server. Further, the central processor 622 may be configured to communicate with the storage media 630 and execute a series of instruction operations in the storage media 630 on the server 600.

[0149] Server 600 may also include one or more power supplies 626, one or more wired or wireless network interfaces 650, one or more input / output interfaces 658, and / or one or more operating systems 641, such as Windows Server TM , Mac OS X TM , Unix TM , Linux TM , FreeBSD TM and so on.

[0150] In this embodiment, the central processor 622 in the server 600 may execute the content recognition method provided by the embodiments of the present application.

[0151] According to one aspect of the present application, there is provided a computer-readable storage medium for storing a computer program for executing the content recognition method described in the foregoing various embodiments.

[0152] According to one aspect of the present application, there is provided a computer program product, which includes a computer program stored in a computer-readable storage medium. The processor of the computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the methods provided in the various optional implementation manners of the above embodiments.

[0153] The descriptions of the processes or structures corresponding to the above various drawings have their own emphases. For parts not detailed in a certain process or structure, reference may be made to the relevant descriptions of other processes or structures.

[0154] In the description of the present application and the above-mentioned drawings, the terms "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device comprising a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0155] In several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces, and the indirect coupling or communication connection of the device or unit can be in electrical, mechanical, or other forms.

[0156] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0157] In addition, in each embodiment of the present application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0158] When the integrated unit is implemented in the form of 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 application, in essence, or the part that contributes to the prior art, or all or part of this 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 for causing a computer device (which can be a terminal, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes: various media that can store computer programs, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical discs.

[0159] In the embodiments of this application, the term "module" or "unit" refers to a computer program with a predetermined function or a part of a computer program, which works together with other related parts to achieve a predetermined goal, and can be fully or partially implemented by using software, hardware (such as a processing circuit or a memory), or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be a part of the overall module or unit that includes the function of that module or unit.

[0160] As described above, the above embodiments are only used to illustrate the technical solution of this application, rather than to limit it; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of this application.

Claims

1. A content recognition method, characterized in that, The method includes: Obtaining interaction data to be processed, where the interaction data to be processed is generated by interacting with the content to be recognized; Determining a first exposure degree of the content to be recognized corresponding to the interaction data to be processed to obtain the first exposure degree of the content to be recognized, and identifying an object type of the object to be recognized corresponding to the interaction data to be processed to obtain an object recognition result. The first exposure degree represents the number of times of interacting with the content to be recognized, and the object to be recognized is the object that generates the interaction data to be processed; If it is determined based on the object recognition result that the object to be recognized is a target object, determining a target object concentration of the content to be recognized, where the target object concentration represents the frequency of occurrence of the target object among all objects that have generated the interaction data to be processed for the content to be recognized; If the first exposure degree reaches an exposure threshold and the target object concentration reaches a concentration threshold, determining that the content to be recognized is a target content.

2. The method according to claim 1, wherein The determining a first exposure degree of the content to be recognized corresponding to the interaction data to be processed to obtain the first exposure degree of the content to be recognized includes: Obtaining a content identifier of the content to be recognized; Querying in a first exposure degree database based on the content identifier to obtain the first exposure degree of the content to be recognized, where the first exposure degree database is used to store the first exposure degrees of the contents indicated by different content identifiers.

3. The method according to claim 2, wherein The method further includes: Updating the first exposure degree of the content to be recognized stored in the first exposure degree database based on the content identifier.

4. The method according to claim 1, characterized in that The identifying an object type of the object to be recognized corresponding to the interaction data to be processed to obtain an object recognition result includes: Obtaining an object identifier of the object to be recognized; Querying in an object list stored in an object database based on the object identifier to obtain the object recognition result, where the object list is a list composed of objects belonging to the target object; The determining that the object to be recognized is a target object based on the object recognition result includes: If the object recognition result indicates that the object to be recognized exists in the object list, determining that the object to be recognized is the target object.

5. The method according to claim 4, wherein After determining that the content to be recognized is a target content, the method further includes: Obtaining historical interaction data generated by interacting with the content to be recognized; Updating the object list based on the historical object corresponding to the historical interaction data.

6. The method according to claim 4, characterized in that, The object to be recognized is a to-be-recognized interaction object, a to-be-recognized device address, or a to-be-recognized Internet protocol address; When the object to be recognized is a to-be-recognized interaction object, the target object is a target interaction object, and the objects in the object list are interaction objects belonging to the target interaction object; When the object to be recognized is a to-be-recognized device address, the target object is a target device address, and the objects in the object list are device addresses belonging to the target device address; When the object to be recognized is a to-be-recognized Internet protocol address, the target object is a target Internet protocol address, and the objects in the object list are Internet protocol addresses belonging to the target Internet protocol address.

7. The method according to claim 1, characterized in that, Determining the target object concentration of the content to be recognized includes: Determining a second exposure degree for the content to be recognized to obtain the second exposure degree of the content to be recognized, where the second exposure degree represents the number of times of interaction generated by the target object with the content to be recognized; Determining the ratio between the second exposure degree and the first exposure degree as the target object concentration.

8. The method according to claim 7, characterized in that Determining the second exposure degree for the content to be recognized to obtain the second exposure degree of the content to be recognized includes: Obtaining the content identifier of the content to be recognized; Querying in a second exposure degree database based on the content identifier to obtain the second exposure degree, where the second exposure degree database is used to store the second exposure degrees of the contents indicated by different content identifiers.

9. The method according to claim 8, wherein The method further includes: Updating the second exposure degree of the content to be recognized stored in the second exposure degree database based on the content identifier.

10. The method according to any one of claims 1-9, characterized in that, After determining that the content to be recognized is the target content, the method further includes: Performing content disposal on the content to be recognized, where the content disposal is used to restrict the interaction of the interaction object with the content to be recognized.

11. The method according to claim 10, wherein Performing content disposal on the content to be recognized includes at least one of the following: Canceling the interaction function for interacting with the content to be recognized; Or, deleting the content to be recognized.

12. The method according to claim 10, characterized in that Performing content disposal on the content to be recognized includes: Sending the content to be recognized to a verification platform; If a verification passed notification returned by the verification platform is received, performing content disposal on the content to be recognized.

13. A content recognition device, characterized in that, The apparatus includes an acquisition unit, a determination unit, and an identification unit: The acquisition unit is used to acquire interaction data to be processed, where the interaction data to be processed is generated by interacting with the content to be recognized; The determination unit is used to determine a first exposure degree for the content to be recognized corresponding to the interaction data to be processed to obtain the first exposure degree of the content to be recognized, where the first exposure degree represents the number of times of interaction with the content to be recognized; The identification unit is used to perform object type identification on the object to be recognized corresponding to the interaction data to be processed to obtain an object identification result, where the object to be recognized is the object that generates the interaction data to be processed; The determination unit is further used to, if it is determined based on the object identification result that the object to be recognized is a target object, determine the target object concentration of the content to be recognized, where the target object concentration represents the frequency of occurrence of the target object among all the objects that have generated the interaction data to be processed for the content to be recognized; The determination unit is further used to, if the first exposure degree reaches an exposure degree threshold and the target object concentration reaches a concentration threshold, determine that the content to be recognized is the target content.

14. A computer device, characterized in that, The computer device includes a processor and a memory: The memory is used to store a computer program and transmit the computer program to the processor; The processor is used to execute the method according to any one of claims 1-12 based on the instructions in the computer program.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program, and when the computer program is executed by a processor, the processor is caused to execute the method according to any one of claims 1-12.

16. A computer program product, comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1-12.

Images