Content auditing method and device, storage medium and computer equipment

Abstract

The invention provides a content auditing method, which comprises the following steps of: acquiring a work link and a work identifier of each new work, and opening each work link by adopting a DrissiPage to carry out screenshot to obtain a corresponding work screenshot; after the work screenshot of each newly-added work is converted into a picture link, the work link, the work identifier and the picture link of each newly-added work are stored in an existing work information base; calculating the similarity between each new work and other works in the existing work information base according to the picture link stored in the existing work information base, and updating the existing work information base according to the calculation result; and when a content auditing request is received, obtaining the work identifier of the work to be audited from the content auditing request, and determining a content auditing result of the work to be audited according to the plurality of corresponding similarities of the work identifier of the work to be audited in the existing work information base. In this way, the efficiency and accuracy of content auditing are effectively improved through the automatic process.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a content review method, device, storage medium and computer equipment. Background Art

[0002] In modern marketing activities, many companies increase brand exposure by encouraging users to participate in interactions and attract more users through reward mechanisms. Therefore, marketing activities often rely on user-generated content, especially video or graphic content, to enhance the coverage and effectiveness of activities and motivate users to publish creative content related to the brand.

[0003] However, as user participation increases, some users have resorted to unfair means to obtain more rewards. For example, users may increase their interactions by repeatedly posting the same video or graphic content, or plagiarizing other people's works, which may cause the reward mechanism to fail. In addition, existing content review methods usually rely on manual inspections, which require a lot of manual input to check the similarity of content one by one, and after the work is deleted, it is impossible to effectively trace the historical content, making it difficult to ensure the fairness of reward distribution. Summary of the invention

[0004] The purpose of this application is to solve at least one of the above-mentioned technical defects, especially the technical defect that the existing content review method usually relies on manual inspection, which requires a lot of manual input to check the similarity of the content one by one, and after the work is deleted, it is impossible to effectively trace the historical content, making it difficult to ensure the fairness of the reward distribution.

[0005] In a first aspect, the present application provides a content review method, the method comprising:

[0006] Obtain the work link and work identifier of each newly added work, and use DrissionPage to open each work link and take a screenshot to obtain the corresponding work screenshot;

[0007] After converting the work screenshot of each newly added work into an image link, the work link, work identifier and image link of each newly added work are saved to the existing work information library to update the existing work information library;

[0008] Based on the image links stored in the existing work information database, calculate the similarity between each newly added work and other works in the existing work information database, and update the existing work information database based on the calculation results;

[0009] When a content review request is received, the work identifier of the work to be reviewed is obtained from the content review request, and the content review result of the work to be reviewed is determined based on multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information database.

[0010] In one embodiment, the steps of using DrissionPage to open each work link and take a screenshot to obtain the corresponding work screenshot include:

[0011] For each work link, DrissionPage is used to open the work link. When the information page corresponding to the work link has been loaded, a screenshot of the loaded information page is taken to obtain a screenshot of the work corresponding to the work link.

[0012] In one embodiment, the step of calculating the similarity between each newly added work and other works in the existing work information database according to the image link stored in the existing work information database includes:

[0013] Determine the image features of all works in the existing work information database through the image links stored in the existing work information database;

[0014] Calculate the cosine similarity between the image features of each newly added work and the image features of other works.

[0015] In one embodiment, the step of determining the content review result of the work to be reviewed according to multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information database includes:

[0016] Searching for multiple corresponding similarities in an existing work information database according to the work identifier of the work to be reviewed;

[0017] If at least one of the similarities corresponding to the work identifier of the work to be reviewed satisfies the preset similarity condition, the content review result of the work to be reviewed is failure to pass, otherwise the content review result of the work to be reviewed is passing.

[0018] In one embodiment, the method further comprises:

[0019] Constructing a screenshot task table, and after saving the acquired work link and work identifier of each newly added work to the screenshot task table, marking the task status corresponding to the work link of each newly added work as unfinished;

[0020] When the screenshot task is triggered, DrissionPage is used to open the work link with the task status of unfinished in the screenshot task table, take a screenshot, and update the corresponding task status mark to completed.

[0021] In one embodiment, the process of triggering the screenshot task includes:

[0022] If the time difference between the current time and the end time of the last screenshot task reaches a preset time interval, multiple threads are started to execute the screenshot task in each thread.

[0023] In a second aspect, the present application provides a content review device, the device comprising:

[0024] The work screenshot acquisition module is used to obtain the work link and work identifier of each newly added work, and use DrissionPage to open each work link to take a screenshot, and obtain the corresponding work screenshot;

[0025] An image link generation module is used to convert the work screenshot of each newly added work into an image link, and save the work link, work identifier and image link of each newly added work to the existing work information library to update the existing work information library;

[0026] A similarity calculation module is used to calculate the similarity between each newly added work and other works in the existing work information database based on the image links stored in the existing work information database, and update the existing work information database based on the calculation results;

[0027] The work content review module is used to obtain the work identifier of the work to be reviewed from the content review request when a content review request is received, and determine the content review result of the work to be reviewed based on multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information database.

[0028] In one embodiment, the work screenshot acquisition module includes:

[0029] The work screenshot acquisition unit is used to use DrissionPage to open each work link, and when the information page corresponding to the work link has been loaded, take a screenshot of the loaded information page to obtain the work screenshot corresponding to the work link.

[0030] In a third aspect, the present application provides a storage medium: the storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of any content review method in the above embodiments.

[0031] In a fourth aspect, the present application provides a computer device, comprising: one or more processors, and a memory;

[0032] Computer-readable instructions are stored in the memory. When the computer-readable instructions are executed by one or more processors, the steps of the content review method in any of the above embodiments are performed.

[0033] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0034] In the content review method, device, storage medium and computer equipment provided in this application, the efficiency and accuracy of content review are effectively improved through automated processes. First, each new work link is automatically opened and screenshotted through DrissionPage. After obtaining the screenshot of the work, it is converted into a picture link and saved to the existing work information library, and the work library is updated in time; then, according to the similarity calculation results between the new work and the existing work, the work information library is updated, so that the behavior of repeated publishing or plagiarism of other people's works can be accurately detected; afterwards, when a content review request is received, other similar works are retrieved from the work library according to the identifier of the work to be reviewed, and the review result of the work is determined based on the similarity. In this way, not only the workload of manual review is reduced, manual misjudgment or omission is avoided, but also historical content can be efficiently traced, the fairness of the reward mechanism is guaranteed, and improper means are avoided to affect the issuance of rewards. In addition, automated updates and similarity analysis can handle large-scale user-generated content, further optimizing the content review and reward mechanism in marketing activities. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0036] Figure 1 A flowchart of a content review method provided in an embodiment of the present application;

[0037] Figure 2 A schematic diagram of the structure of a content review device provided in an embodiment of the present application;

[0038] Figure 3 A schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0040] The present application provides a content review method. The following embodiments are described by taking the method applied to a computer device as an example. It can be understood that the computer device can be any device with data processing functions, including but not limited to a single server, a server cluster, a personal laptop computer, a desktop computer, etc. Figure 1 As shown, the method may include the following steps:

[0041] S101: Obtain a work link and a work identifier of each newly added work, and use DrissionPage to open each work link and take a screenshot to obtain a corresponding work screenshot.

[0042] Among them, newly added works refer to new creative content submitted or uploaded by users in marketing activities, such as videos, pictures and texts. The work link points to the URL (uniform resource locator) of a specific work page, which is used to identify and access the network location of the work. The work identifier is a unique identifier for each work, generally a set of numbers or IDs related to the work, to ensure that each work can be distinguished and identified individually. DrissionPage is a tool library or API for automating browser operations, supporting functions such as opening web pages and taking web page screenshots. Screenshots refer to capturing the current screen or web page content from the display device and saving it in image format.

[0043] In this step, the link and identifier of each work are extracted from the newly submitted data. After receiving the command to open the link, DrissionPage initiates an HTTP request, loads the corresponding web page and renders the page content. After loading, DrissionPage triggers the screenshot function through its automation API, takes a screenshot of the web page content and saves it as an image file.

[0044] It can be understood that by obtaining the work link and identifier of each newly added work, it is possible to accurately identify and locate the work submitted by each user, avoiding omissions or erroneous records. By automatically opening the work link and taking screenshots through DrissionPage, the time cost of manual operation is reduced and the efficiency of batch processing is improved. After obtaining the screenshot of the work, the web page content is converted into an image file to ensure that the original content of the work is preserved. The automated screenshot operation not only improves the processing speed, but also ensures the consistency of content review and avoids deviations in manual review. In addition, this method can efficiently track and manage a large amount of user-generated content, support rapid retrieval and comparison, and effectively improve the fairness and execution of marketing activities.

[0045] S102: After converting the work screenshot of each newly added work into an image link, the work link, work identifier and image link of each newly added work are saved to the existing work information library to update the existing work information library.

[0046] The image link refers to a URL pointing to a screenshot image generated in some way, which is used to access or store the image on the Internet. The existing work information library is a database or storage system that contains various information about uploaded works, such as work links, work identifiers, screenshot links, etc.

[0047] In this step, after the screenshot of the work is generated, the image file is saved to a specified storage location, such as a local server or cloud storage, and the URL of the image, i.e., the image link, is generated according to the image storage path. The link is the network address of the image and can be accessed. For example, the image can be uploaded to a cloud storage platform to generate a unique access URL. Next, the relevant information of the newly added work is saved to the existing work information library through database operations, that is, the information is inserted into the work information table as a new record, wherein the consistency and integrity of the data need to be ensured to avoid duplicate data or information loss. As new data is inserted, the data in the existing work information library will be updated to ensure that the library contains the latest work information.

[0048] It can be understood that, firstly, converting the screenshots of the works into image links avoids the loss or inaccessibility of the screenshot files. Secondly, by storing the work links, work identifiers and image links in the existing work information library, the centralized management and unified storage of data are realized, which not only facilitates the subsequent rapid query and content review, but also ensures the unique identification and traceability of the works, thereby improving the efficiency of data management, ensuring the integrity and consistency of the content, and avoiding errors in manual processing. In addition, with the continuous updating of the work information library, new works can be tracked and managed in real time, thereby improving the automation and accuracy of content review and optimizing the overall marketing activity management.

[0049] S103: Calculate the similarity between each newly added work and other works in the existing work information database according to the image links stored in the existing work information database, and update the existing work information database according to the calculation results.

[0050] Among them, similarity is an indicator to measure the degree of similarity between two works, usually calculated by an algorithm to measure the similarity between screenshots of the works.

[0051] In this step, by reading the image link of the newly added work and the image link of other works in the existing work information library, relevant screenshots are downloaded and compared. Similarity calculation can use image processing and computer vision technology, such as image feature extraction, deep learning model, traditional image matching algorithm, etc. The screenshots of the newly added work are compared with the screenshots of each existing work one by one, the key features of the two pictures are extracted and the similarity between them is calculated. The calculation method can be pixel-based similarity calculation, or the similarity evaluation can be performed after extracting image features through advanced models.

[0052] After the similarity is calculated, the similarity results between each newly added work and the existing work are saved in the database, and the relevant records in the work information library are updated. For example, a "similarity" field can be added to the database table, or an association table can be established between the work information table and the similarity result table to record the similarity data between the newly added work and other works. In this way, the work information library can contain the similarity relationship between the works.

[0053] It can be understood that, first of all, similarity calculation can identify duplicate or similar works, avoiding users from abusing the reward mechanism by repeatedly uploading the same or similar content. By calculating the similarity between works, it is possible to automatically detect and mark those works that may involve plagiarism or plagiarism, reducing the workload of manual review and improving the efficiency and accuracy of the review. Secondly, after updating the calculation results to the existing work information database, the latest status of the work information can be maintained, ensuring that subsequent query, comparison and content review operations have accurate data support. In this way, not only the accuracy of data management is improved, but also the relationship between works can be tracked in real time, providing a data basis for further content analysis, recommendation or content deduplication. In addition, with the continuous updating of the database, the relationship between new works and historical works can be better handled, ensuring that the work information database has strong scalability and flexibility, and supports large-scale content management tasks.

[0054] S104: When a content review request is received, the work identifier of the work to be reviewed is obtained from the content review request, and the content review result of the work to be reviewed is determined according to multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information database.

[0055] Among them, a content review request refers to a request for review of a certain work or content, with the purpose of examining the legality and originality of the work. Works to be reviewed refer to works that currently need to be reviewed. Content review results refer to the conclusions drawn after the review, which may be "passed" or "rejected", or other evaluation results of the work content, such as suspected plagiarism, violation of regulations, etc.

[0056] In this step, when a content review request is received, the identifier of the work to be reviewed is extracted from the request. After that, the identifier of the work to be reviewed is used to search for relevant similarity data in the existing work information library. After obtaining the similarity data between the work to be reviewed and other works, the work to be reviewed is judged according to the set rules. For example, if the similarity exceeds a certain threshold, the work may be judged to be duplicate or plagiarized, and the review result is "rejected". On the contrary, if the similarity is low, the work may be judged to be original, and the review result is "passed".

[0057] It can be understood that, firstly, by extracting the work identifier from the content review request and querying the existing work information database, the work to be reviewed can be quickly located and the similarity data related to it can be obtained, thus avoiding the time waste of manual search and being able to efficiently handle large-scale review requests. Secondly, based on the similarity data in the existing work information database, it is possible to quickly determine whether the work to be reviewed is similar or repeated with other works, thereby automatically deciding whether to pass the review. In this way, not only the accuracy of the review is improved, but also the deviation and omission in manual review are avoided.

[0058] In the above embodiment, the efficiency and accuracy of content review are effectively improved through the automated process. First, each newly added work link is automatically opened and screenshotted through DrissionPage. After obtaining the screenshot of the work, it is converted into a picture link and saved to the existing work information library, and the work library is updated in time; then, according to the similarity calculation results between the new work and the existing work, the work information library is updated, so that the behavior of repeated publishing or plagiarism of other people's works can be accurately detected; afterwards, when a content review request is received, other similar works are retrieved from the work library according to the identifier of the work to be reviewed, and the review result of the work is determined based on the similarity. In this way, not only the workload of manual review is reduced, and manual misjudgment or omission is avoided, but also historical content can be efficiently traced, the fairness of the reward mechanism is guaranteed, and improper means are avoided to affect the issuance of rewards. In addition, automated updates and similarity analysis can handle large-scale user-generated content, further optimizing the content review and reward mechanism in marketing activities.

[0059] In one embodiment, the steps of using DrissionPage to open each work link and take a screenshot to obtain the corresponding work screenshot include:

[0060] For each work link, DrissionPage is used to open the work link. When the information page corresponding to the work link has been loaded, a screenshot of the loaded information page is taken to obtain a screenshot of the work corresponding to the work link.

[0061] Among them, the information page refers to the detailed display page of the work, including the pictures, texts, descriptions and other contents of the work.

[0062] Specifically, first open the work link through DrissionPage, that is, use the API of the DrissionPage library to request the target web page. The browser or web driver will load the web page. At this time, DrissionPage will simulate the user's browser behavior, access the target URL in the background, and obtain the page HTML, CSS, JavaScript and other resources. After opening the link, DrissionPage will wait for all resources of the page to be loaded. In order to ensure that the page is fully loaded, DrissionPage can use explicit waiting or detect specific DOM elements to determine whether the page has been loaded. For example, you can wait for certain key elements of the page, such as the work content, pictures, and comment areas, to be loaded to ensure that the web page displays the final state. When the web page is loaded, DrissionPage provides a screenshot function. By calling an API like `driver.get_screenshot_as_file(filename)`, you can capture the image of the current browser window and save it as a local image file. The screenshot content is the complete work information page displayed on the current web page, including all loaded text, images, and other content.

[0063] In this embodiment, by using DrissionPage to open the work link and take a screenshot, the display content of the work page can be captured and saved efficiently and automatically. First, DrissionPage simulates user operations to open the work link and load the page, avoiding the inefficiency of manual operation. Waiting for the page to load to complete ensures the accuracy of the screenshot, ensuring that the captured content reflects the final presentation state of the page, rather than the partial content being loaded, thereby avoiding incomplete or erroneous screenshots. In addition, the screenshots of the work generated by the screenshot can be used as a conclusive record of the content of the work, which is convenient for subsequent content review, comparison and archiving. Through the automated screenshot process, not only the efficiency is improved and manual intervention is reduced, but also the consistency and accuracy of the data can be guaranteed. Especially when faced with a large number of works, this automated screenshot process can greatly reduce labor costs. By taking screenshots and saving the work page, it can ensure that each work has a clear and traceable record.

[0064] In one embodiment, the step of calculating the similarity between each newly added work and other works in the existing work information database according to the image links stored in the existing work information database includes:

[0065] Determine the image features of all works in the existing work information database through the image links stored in the existing work information database;

[0066] Calculate the cosine similarity between the image features of each newly added work and the image features of other works.

[0067] Here, image features refer to the unique properties or information of an image. Cosine similarity is a metric used to calculate the similarity between two vectors.

[0068] Specifically, first, by reading the image link in the existing work information library, the image file of each work is accessed and its features are extracted. Image feature extraction is usually completed through deep learning models or classic computer vision algorithms. Through these algorithms, each image can be converted into a feature vector representing the main visual features of the image. The extracted feature vector can contain information about the color, texture, shape, etc. of the image. For each new work, the feature vector of its image is extracted, and then compared with the feature vectors of all other works in the existing work information library, and the cosine similarity is calculated. The calculated similarity value can be used to quantify the similarity between the new work and the existing work.

[0069] In this embodiment, by extracting the image features of existing works and calculating the cosine similarity between the newly added works and the existing works, similar or repeated works can be effectively identified. First, extracting image features can convert complex images into digital feature vectors, which makes the comparison between images more efficient and standardized. The calculation of cosine similarity can accurately quantify the similarity between images and can effectively determine whether the image content is similar or repeated. Secondly, works that may be repeatedly uploaded or suspected of plagiarism can be automatically identified, avoiding possible omissions in manual review. In addition, the process of calculating similarity can identify similar content that may be hidden in different images by analyzing the high-level features of the image, and can maintain efficiency when processing large amounts of image data.

[0070] In one embodiment, the step of determining the content review result of the work to be reviewed according to multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information database includes:

[0071] Searching for multiple corresponding similarities in an existing work information database according to the work identifier of the work to be reviewed;

[0072] If at least one of the similarities corresponding to the work identifier of the work to be reviewed satisfies the preset similarity condition, the content review result of the work to be reviewed is failure to pass, otherwise the content review result of the work to be reviewed is passing.

[0073] Among them, the preset similarity conditions are standards set by the platform or review system, which define the degree to which works are considered "similar".

[0074] Specifically, first use the identifier of the work to be reviewed to search for similarity data related to it in the existing work information database. Each work to be reviewed has a corresponding identifier, and the similarity data in the existing work information database is retrieved based on this identifier. Once the similarity data related to the work to be reviewed is obtained from the existing work information database, these similarities are checked one by one to see if they meet the preset similarity conditions. For example, the preset similarity condition is a threshold. If any similarity exceeds this threshold, the work to be reviewed is considered to be similar to the existing work, and the work will be judged to have "failed the review." If none of the similarities of the work to be reviewed meets the preset similarity conditions, the work will be considered to be independent and original, and its review result will be marked as "passed the review."

[0075] In this embodiment, by searching the similarity of the work to be reviewed in the existing work information database and comparing it with the preset similarity conditions, it is possible to automatically and efficiently determine whether the work is duplicated or plagiarized, avoiding the inefficiency and omissions of manual review. First, the process of searching for similarity is based on known similarity data, and all relevant similarity information can be quickly obtained without recalculation. Secondly, comparing the similarity of the work to be reviewed with the preset similarity conditions helps to automate the review process. By setting the similarity threshold, the platform can flexibly adjust the strictness of the review to ensure that works that comply with the platform rules can pass the review smoothly, while duplicate or non-compliant works can be discovered and blocked in time. Such an automated review process can greatly improve the review efficiency, reduce manual intervention, and ensure the originality and quality of the platform content.

[0076] In one embodiment, the method further comprises:

[0077] Constructing a screenshot task table, and after saving the acquired work link and work identifier of each newly added work to the screenshot task table, marking the task status corresponding to the work link of each newly added work as unfinished;

[0078] When the screenshot task is triggered, DrissionPage is used to open the work link with the task status of unfinished in the screenshot task table, take a screenshot, and update the corresponding task status mark to completed.

[0079] The screenshot task table is a database table or data structure used to store relevant information of all pending screenshot tasks. Each record includes fields such as work link, work identifier, and task status. The task status refers to the current execution status of the task, which usually has two states: "unfinished" and "completed". The task status indicates whether the screenshot task has been processed and the screenshot has been completed.

[0080] Specifically, first, you need to create a screenshot task table. Every time a new work is added, the link and identifier of the work are inserted into the screenshot task table, and the task status is initialized to "unfinished", indicating that the screenshot task has not been triggered and executed. When it is detected that a screenshot task is triggered, query all tasks with a status of "unfinished" in the screenshot task table to obtain the corresponding work link. Then, use DrissionPage to automatically open these links, access the work page, and take a screenshot. After the screenshot is completed, update the task status to "completed", indicating that the task has been successfully executed.

[0081] In this embodiment, by constructing a screenshot task table and saving the task information in the table, the screenshot tasks can be effectively managed and scheduled. The link and identifier of each newly added work will be recorded, and the task status is initially set to "unfinished", so that you can clearly track which works have not yet been screenshotted. Updating the task status to "completed" helps ensure the traceability of the task and avoid repeating the same task. In this way, the execution process of the task can be accurately controlled, omissions and errors can be avoided, and the screenshot tasks can be efficiently managed and scheduled, thereby saving resources and reducing labor costs.

[0082] In one embodiment, the process of triggering the screenshot task includes:

[0083] If the time difference between the current time and the end time of the last screenshot task reaches a preset time interval, multiple threads are started to execute the screenshot task in each thread.

[0084] Among them, the current time refers to the current moment obtained by the computer. The last screenshot task end time refers to the time when the last screenshot task was completed. The time difference refers to the difference between the current time and the last screenshot task end time. The preset time interval is a fixed time period set in advance, which means the shortest time interval that must pass between two screenshot tasks. A thread is an independent execution flow in a program, and it is the smallest unit of program execution.

[0085] Specifically, first obtain the current timestamp, compare it with the timestamp of the last screenshot task, and calculate the time difference between the two. If the time difference reaches the preset time interval, such as 3 minutes, continue to execute the subsequent tasks. Once it is determined that the time difference has reached the preset time interval, multiple threads will be started to execute the screenshot task in parallel. Each thread will be responsible for executing an independent screenshot task. In each thread, DrissionPage will be used to open the link of the specified work and take a screenshot. Each thread works independently without interfering with each other to complete its own screenshot task.

[0086] In this embodiment, by calculating the time difference between the current time and the end time of the last screenshot task, it can be ensured that the execution of the screenshot task does not exceed the predetermined interval time. The purpose of setting the preset time interval is to control the frequency of task execution and prevent the screenshot tasks from being too frequent or too sparse, so as to ensure the rational use of system resources. For example, if the preset time interval is 30 minutes, a new screenshot task will only be started after 30 minutes to avoid frequent task scheduling or ineffective resource consumption. When the time difference reaches the preset time interval, multiple threads are started and the screenshot tasks are executed concurrently, which can significantly improve the efficiency of task execution, especially when there are multiple works that need to be screenshotted. Through concurrent execution, the screenshot tasks of multiple works can be processed simultaneously without waiting for the previous task to be completed before starting the next task. In this way, the computing power of the multi-core processor can be fully utilized, the waiting time for task processing can be reduced, and the speed of overall task execution can be accelerated.

[0087] The following describes the content review device provided by the embodiment of the present application. The content review device described below and the content review method described above can be referred to in correspondence with each other. Figure 2 As shown, the present application provides a content review device, which includes:

[0088] The work screenshot acquisition module 201 is used to acquire the work link and work identifier of each newly added work, and use DrissionPage to open each work link to take a screenshot, and obtain the corresponding work screenshot;

[0089] The picture link generation module 202 is used to convert the work screenshot of each newly added work into a picture link, and then save the work link, work identifier and picture link of each newly added work to the existing work information library to update the existing work information library;

[0090] A similarity calculation module 203 is used to calculate the similarity between each newly added work and other works in the existing work information database based on the image links stored in the existing work information database, and update the existing work information database based on the calculation results;

[0091] The work content review module 204 is used to obtain the work identifier of the work to be reviewed from the content review request when a content review request is received, and determine the content review result of the work to be reviewed based on multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information library.

[0092] In one embodiment, the work screenshot acquisition module 201 includes:

[0093] The work screenshot acquisition unit is used to use DrissionPage to open each work link, and when the information page corresponding to the work link has been loaded, take a screenshot of the loaded information page to obtain the work screenshot corresponding to the work link.

[0094] In one embodiment, the similarity calculation module 203 includes:

[0095] A picture feature determination unit, used to determine the picture features of all works in the existing work information database through the picture links stored in the existing work information database;

[0096] The similarity calculation unit is used to calculate the cosine similarity between the image features of each newly added work and the image features of other works.

[0097] In one embodiment, the work content review module 204 includes:

[0098] A similarity search unit, used to search for corresponding multiple similarities in an existing work information database according to the work identifier of the work to be reviewed;

[0099] The work content review unit is used to review the work content as failure if at least one of the similarities corresponding to the work identifier of the work to be reviewed satisfies a preset similarity condition; otherwise, the work content review result is approval.

[0100] In one embodiment, the apparatus further comprises:

[0101] A screenshot task list construction module is used to construct a screenshot task list, and after saving the acquired work link and work identifier of each newly added work to the screenshot task list, mark the task status corresponding to the work link of each newly added work as unfinished;

[0102] The screenshot task execution module is used to use DrissionPage to open the work link with the task status of unfinished in the screenshot task table, take a screenshot, and update the corresponding task status mark as completed when the screenshot task is triggered.

[0103] In one embodiment, the screenshot task execution module includes:

[0104] The screenshot task triggering unit is used to start multiple threads if the time difference between the current time and the end time of the last screenshot task reaches a preset time interval, so as to execute the screenshot task in each thread.

[0105] In one embodiment, the present application also provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the content review method described in any one of the above embodiments.

[0106] In one embodiment, the present application also provides a computer device having computer-readable instructions stored therein. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the content review method as described in any one of the above embodiments.

[0107] Indicatively, if Figure 3 As shown, Figure 3 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. The computer device 300 may be provided as a server. Figure 3 , the computer device 300 includes a processing component 302, which further includes one or more processors, and a memory resource represented by a memory 301 for storing instructions executable by the processing component 302, such as an application. The application stored in the memory 301 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 302 is configured to execute instructions to perform the content audit method of any of the above embodiments.

[0108] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate based on an operating system stored in the memory 301, such as Windows Server TM, Mac OS X TM, Unix TM, Linux TM, Free BSD TM, or the like.

[0109] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0110] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the term "include", "comprise" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not clearly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements. Herein, "one", "one", "said", "the" and "it" may also include plural forms, unless the context clearly indicates another way. A plurality refers to at least two cases, such as 2, 3, 5 or 8, etc. "And / or" includes any and all combinations of the relevant listed items.

[0111] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can refer to each other.

[0112] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A content review method, characterized in that: The method comprises: Obtain the work link and work identifier of each newly added work, and use DrissionPage to open each of the work links and take a screenshot to obtain a corresponding work screenshot; After converting the work screenshot of each of the newly added works into an image link, the work link, work identifier and image link of each of the newly added works are saved to the existing work information library to update the existing work information library; Calculate the similarity between each of the newly added works and other works in the existing work information database according to the image links stored in the existing work information database, and update the existing work information database according to the calculation results; When a content review request is received, the work identifier of the work to be reviewed is obtained from the content review request, and the content review result of the work to be reviewed is determined based on multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information library.

2. The content review method according to claim 1, characterized in that: The step of using DrissionPage to open each of the work links and take screenshots to obtain corresponding work screenshots includes: For each of the work links, DrissionPage is used to open the work link. When the information page corresponding to the work link has been loaded, a screenshot of the loaded information page is taken to obtain a screenshot of the work corresponding to the work link.

3. The content review method according to claim 1, characterized in that: The step of calculating the similarity between each of the newly added works and other works in the existing work information database according to the picture links stored in the existing work information database comprises: Determine the image features of all works in the existing work information database through the image links stored in the existing work information database; The cosine similarity between the image features of each newly added work and the image features of other works is calculated.

4. The content review method according to claim 1, characterized in that: The step of determining the content review result of the work to be reviewed according to the multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information database includes: Searching for corresponding multiple similarities in an existing work information database according to the work identifier of the work to be reviewed; If at least one of the similarities corresponding to the work identifier of the work to be reviewed satisfies a preset similarity condition, the content review result of the work to be reviewed is failure to review; otherwise, the content review result of the work to be reviewed is passing to review.

5. The content review method according to claim 1, characterized in that: The method further comprises: Constructing a screenshot task table, and after saving the acquired work link and work identifier of each newly added work to the screenshot task table, marking the task status corresponding to the work link of each newly added work as unfinished; When the screenshot task is triggered, DrissionPage is used to open the work link with the task status of unfinished in the screenshot task table, take a screenshot, and update the corresponding task status mark to completed.

6. The content review method according to claim 5, characterized in that: The process of triggering the screenshot task includes: If the time difference between the current time and the end time of the last screenshot task reaches a preset time interval, multiple threads are started to execute the screenshot task in each thread.

7. A content review device, characterized in that: The device comprises: A work screenshot acquisition module is used to obtain the work link and work identifier of each newly added work, and use DrissionPage to open each of the work links to take a screenshot, and obtain the corresponding work screenshot; A picture link generation module, configured to convert the work screenshot of each newly added work into a picture link, and save the work link, work identifier and picture link of each newly added work to an existing work information library to update the existing work information library; A similarity calculation module, used to calculate the similarity between each of the newly added works and other works in the existing work information database according to the image links stored in the existing work information database, and update the existing work information database according to the calculation results; The work content review module is used to obtain the work identifier of the work to be reviewed from the content review request when a content review request is received, and determine the content review result of the work to be reviewed based on multiple similarities corresponding to the work identifier of the work to be reviewed in the existing work information library.

8. The content review method according to claim 7, characterized in that: The work screenshot acquisition module includes: The work screenshot acquisition unit is used to use DrissionPage to open each work link, and when the information page corresponding to the work link has been loaded, take a screenshot of the loaded information page to obtain the work screenshot corresponding to the work link.

9. A storage medium, characterized in that: The storage medium stores computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the content review method as described in any one of claims 1 to 6.

10. A computer device, characterized in that: include: one or more processors, and memory; The memory stores computer-readable instructions, which, when executed by the one or more processors, perform the steps of the content review method as described in any one of claims 1 to 6.

Images