Picture information intelligent auditing method, device and equipment and storage medium

Through the deep learning model combined with variant recognition, semantic analysis and context association technology, the image-type materials are customized for review, solving the problem of difficult to identify image-type materials and high false alarm rates in the existing technology, and achieving efficient violation recognition and audit.

CN119006004APending Publication Date: 2024-11-22CHINA CITIC BANK CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411035110.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-31
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The existing technology is difficult to identify and filter illegal content in picture materials, and the false alarm rate is high, so it cannot deal with new threats, especially in the case of massive information dissemination on social media platforms.

Method used

Using deep learning model combined with variant recognition, semantic analysis and context association technology, through a customized audit model for business scenarios, we can identify and review icon features, item features, face features and text content to generate image review results to avoid a single model manslaughter and missed killing.

Benefits of technology

It improves the accuracy and review efficiency of violation recognition, can identify illegal content in pictures, reduce false alarms, adapt to new threats, and enhances the review ability of special characters and emojis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119006004A_ABST
    Figure CN119006004A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of data processing, in particular to a picture information intelligent auditing method and device, equipment and a storage medium. The method comprises the following steps: acquiring to-be-published picture content from to-be-published content of a publishing terminal, inputting the to-be-published picture content into an intelligent auditing system, then preprocessing the to-be-published picture content, generating to-be-audited picture content, then identifying icon features, article features and face features in a picture, and identifying auxiliary features of a face and performing positioning; establishing a data analysis model through a business scene related to the to-be-published content and an auditing strategy; and identifying the suspicious text data by using the data analysis model, obtaining violation picture data, analyzing the violation text data to generate a picture auditing result, and finally returning the picture auditing result to the publishing terminal. According to the method, the content of the violation picture is recognized, the auditing model is customized by using different business scenes of the banking industry, misjudgment and missed judgment of a single model are avoided, and the violation recognition accuracy and the auditing efficiency are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, equipment and storage medium for intelligent review of image information. Background Art

[0002] At present, as various social media platforms gradually mature, the threshold for information dissemination is gradually lowered, and users can easily spread information on the Internet. While the scale of user contributions continues to grow, content chaos is becoming increasingly prominent. Some lawless elements use social media to spread malicious promotional content. These harmful contents not only affect the user experience, but also bring great legal risks to the platform and have a very bad impact on the network environment. Therefore, how to identify and filter harmful content from massive information has become an important issue. At present, the bank's publishing terminal mainly uses the data leakage protection system to scan the image data in the marketing materials to be released to prevent the outflow of image data that violates the company's security and compliance policies. It is difficult to audit image materials in the existing technology. It can only audit the text content or the text content in the picture recognized by OCR, and lacks the ability to audit the content of mixed simplified and traditional Chinese, mixed uppercase and lowercase, special characters, emoticons, etc.

[0003] There is an urgent need for an intelligent image information review method that can efficiently complete compliance review of image data, thereby solving the problem that the data leakage protection system in the existing technology is difficult to identify specially modified illegal content and has a high false alarm rate, and is also difficult to deal with new threats. Summary of the invention

[0004] In order to solve the problem that the prior art can only review text content or text content in images recognized by OCR, but cannot review image materials, the embodiments of this specification provide a method, device, equipment and storage medium for intelligent review of image information, which can identify illegal words and their variants based on deep learning models through variant recognition, semantic analysis, context association and other technologies, and use customized review models for different business scenarios in the banking industry to avoid false positives and missed negatives of a single model, thereby improving the accuracy of violation recognition and the efficiency of review.

[0005] In order to solve the above technical problems, the specific technical solutions of this specification are as follows:

[0006] On the one hand, the embodiments of this specification provide a method for intelligent review of picture information, including:

[0007] Acquire the picture content to be published from the content to be published of the publishing terminal, and input the picture content to be published into the intelligent review system;

[0008] The intelligent audit system pre-processes the image content to be published to generate multiple images to be audited and text content to be audited;

[0009] Identify and locate icon features, object features, face features, and auxiliary features for identifying faces in the plurality of images to be reviewed;

[0010] Establish a corresponding data analysis model based on the business scenarios and review strategies involved in the content to be published;

[0011] According to the corresponding data analysis model, the icon features, object features, face features, and auxiliary features for identifying faces in multiple images to be reviewed are determined to obtain illegal image data;

[0012] Determine the content of the text to be reviewed according to the corresponding data analysis model to obtain illegal text data;

[0013] The illegal text data and illegal image data are analyzed to generate image review results, and returned to the publishing terminal.

[0014] Furthermore, the intelligent audit system pre-processes the image content to be published, and generates a plurality of images to be audited and text content to be audited, further comprising:

[0015] The preprocessing includes locating and identifying the pattern images and text images in the image content to be published;

[0016] The image content to be published is segmented according to the pattern image and the text image to generate the plurality of images to be reviewed and the text content to be reviewed.

[0017] Furthermore, the preprocessing further includes:

[0018] Performing text recognition on the text images in the image content to be published, restoring the text in special fonts or special layouts, and generating text content;

[0019] The text content is subjected to text denoising, normalization of information of the same category, word segmentation, and removal of stop words; and the text content to be reviewed is generated.

[0020] Furthermore, establishing a corresponding data analysis model according to the business scenarios and review strategies involved in the content to be published further includes:

[0021] Selecting target patterns and target words according to the business scenarios involved in the content to be published;

[0022] Performing multiple deformation processes on the target pattern to generate a pattern recognition library;

[0023] The target word is expanded based on the target word and a preset semantic similarity to generate a scene vocabulary library.

[0024] Furthermore, establishing a data analysis model according to the business scenarios and review strategies involved in the content to be published further includes:

[0025] The audit strategy includes at least one of the following: variant identification, semantic analysis, and context association;

[0026] Acquiring a pattern recognition model of the pattern recognition library through a preset variant pattern algorithm;

[0027] Obtaining a variant recognition model of the scene vocabulary library through a preset variant word algorithm;

[0028] Acquiring a semantic recognition model of the scene vocabulary library through a preset semantic algorithm;

[0029] The association recognition model of the scene vocabulary library is obtained through a preset context association algorithm.

[0030] Further, according to the corresponding data analysis model, the icon features, object features, face features, and auxiliary features for identifying faces in the multiple images to be reviewed are determined, and obtaining the illegal image data further includes:

[0031] Determining the icon features and the object features in the plurality of images to be reviewed according to the pattern recognition model;

[0032] Generate facial information according to the facial features and auxiliary features for identifying the face, and determine the facial information;

[0033] The illegal image data is generated according to the determination results of the icon features and the object features and the determination results of the face information.

[0034] Further, judging the content of the text to be reviewed according to the corresponding data analysis model, obtaining the illegal text data further includes:

[0035] The text content to be reviewed is input into the variant recognition model, the semantic recognition model and the association recognition model to generate the illegal text data.

[0036] Furthermore, analyzing the illegal text data and illegal image data to generate image review results further includes:

[0037] Determining the confidence level of the to-be-published picture content according to the amount of illegal text data and illegal picture data;

[0038] Perform risk classification according to the contents of the illegal text data and illegal image data;

[0039] According to the confidence level of the content of the picture to be released and the risk classification, a violation handling opinion and a distribution of violation content are generated as the picture review result.

[0040] Furthermore, the intelligent audit system further includes synchronous audit and asynchronous audit.

[0041] If the image content to be published is a small image file, the synchronous review service is called;

[0042] If the image content to be published is a large image file, the asynchronous review service is called.

[0043] Further, analyzing the illegal text data to generate image review results, and returning them to the publishing terminal further includes:

[0044] If the synchronous audit service is called, the audit result will be returned directly;

[0045] If the asynchronous audit service interface is called, the audit serial number is returned, and the publishing terminal calls the audit result query interface for polling according to the serial number, or the publishing terminal sets a callback address, and the intelligent audit system sends the audit result to the callback address.

[0046] On the other hand, the embodiment of the present specification provides a picture information intelligent review device, including:

[0047] A content acquisition unit, used to acquire the picture content to be published from the content to be published of the publishing terminal, and input the picture content to be published into the intelligent review system;

[0048] A preprocessing unit, used for the intelligent audit system to preprocess the image content to be published, and generate a plurality of images to be audited and text content to be audited;

[0049] A content positioning unit, used to identify and locate icon features, object features, face features, and auxiliary features for identifying faces in the plurality of images to be reviewed;

[0050] A model building unit, used to determine the icon features, object features, face features, and auxiliary features for identifying faces in multiple images to be reviewed according to a corresponding data analysis model, and obtain illegal image data;

[0051] An illegal picture acquisition unit, used to determine icons, objects, faces and related features in a plurality of images to be reviewed according to the data analysis model, and acquire illegal picture data;

[0052] An illegal text acquisition unit, used to determine the content of the text to be reviewed according to a corresponding data analysis model, and acquire illegal text data;

[0053] The result generating unit is used to analyze the illegal text data and illegal picture data to generate picture review results, and return them to the publishing terminal.

[0054] On the other hand, an embodiment of the present specification further provides a computer device, including a memory, a processor, and a computer program stored in the memory, and the processor implements the above method when executing the computer program.

[0055] On the other hand, an embodiment of the present specification further provides a computer storage medium on which a computer program is stored. When the computer program is executed by a processor of a computer device, the above method is executed.

[0056] Finally, the embodiments of this specification also provide a computer program product, which includes a computer program. When the computer program is executed by a processor, the above method is performed.

[0057] By using the embodiments of this specification, it is realized that the content to be published in the publishing terminal is reviewed by the intelligent review system. First, the image content to be published is obtained from the content to be published in the publishing terminal and input into the intelligent review system, so as to ensure that the intelligent review system obtains the complete image content to be published. Then, the intelligent review system pre-processes the image content to be published to generate multiple images to be reviewed and text content to be reviewed, so as to conduct image and text review respectively, and at the same time avoid useless interference information in the image content from affecting the review process. Then, a data analysis model is established according to the business scenarios and review strategies involved in the content to be published, so as to improve the pertinence of the data analysis model, avoid the audit efficiency being affected by excessive data volume or the audit efficiency being too low due to insufficient pertinence, and then use the data analysis model to identify suspicious images to be reviewed and text content to be reviewed, obtain illegal image data and illegal text data, and can also identify faces and their facial information to identify them, so as to realize the recognition of multiple variants through learning models, and finally generate image review results and return them to the publishing terminal. Image content review technology based on deep learning algorithms can identify illegal images and texts through variant recognition, semantic analysis and other technologies; it improves the efficiency of image review and avoids missed inspections caused by users maliciously evading inspections. At the same time, it uses customized review models for different business scenarios in the banking industry to improve the targetedness of the review model, thereby improving the efficiency of text data review. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of this specification 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 this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0059] Figure 1 It is a schematic diagram of a system for implementing a method for intelligent review of picture information according to an embodiment of this specification;

[0060] Figure 2 It is a flow chart of a method for intelligent review of picture information according to an embodiment of this specification;

[0061] Figure 3 The process of establishing a data analysis model according to business scenarios and audit strategies in the embodiments of this specification is shown;

[0062] Figure 4 The figure shows the process of identifying and obtaining illegal image data according to the data analysis model in the embodiment of this specification;

[0063] Figure 5 The process of classifying and identifying pictures in the embodiment of this specification is shown;

[0064] Figure 6 The text description embodiment shown in the figure analyzes the illegal picture data and generates the picture review result;

[0065] Figure 7 The figure shows a schematic diagram of the structure of an intelligent review device for picture information according to an embodiment of the present specification;

[0066] Figure 8 The figure is a schematic diagram of the structure of the computer device according to the embodiment of the present specification.

[0067]

Description of the accompanying drawings

[0068] 101. Release terminal;

[0069] 102. Intelligent audit system;

[0070] 701, content acquisition unit;

[0071] 702, pre-processing unit;

[0072] 703, content positioning unit;

[0073] 704, model building unit;

[0074] 705. Illegal picture acquisition unit;

[0075] 706. Illegal text acquisition unit;

[0076] 707, result generation unit;

[0077] 802. Computer equipment;

[0078] 804. Processing equipment;

[0079] 806. Storage resources;

[0080] 808, driving mechanism;

[0081] 810, input / output module;

[0082] 812. Input device;

[0083] 814. Output device;

[0084] 816. Presentation equipment;

[0085] 818. Graphical user interface;

[0086] 820, network interface;

[0087] 822, communication link;

[0088] 824. Communication bus. DETAILED DESCRIPTION

[0089] The following will be combined with the drawings in the embodiments of this specification to clearly and completely describe the technical solutions in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0090] It should be noted that the terms "first", "second", etc. in the description and claims of this specification and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of this specification described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, device, product or equipment that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0091] like Figure 1The figure shows a schematic diagram of a system for implementing an intelligent audit method for picture information according to an embodiment of the present invention, which may include a publishing terminal 101 and an intelligent audit system 102. A communication connection is established between the publishing terminal 101 and the intelligent audit system 102 to enable data interaction. The publishing terminal 101 may send content to be published to the intelligent audit system 102, including picture content to be published, which may also include the business scenarios and audit strategies involved, such as bank marketing scenarios. In addition, due to the protection of information security, the content to be published is desensitized, and the fields involving account information, contact information, etc. are replaced with special symbols. The intelligent audit system 102 then audits the content to be published, and then returns the audit results to the publishing terminal.

[0092] In the embodiments of this specification, the intelligent audit system 102 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks (CDN, Content Delivery Network), and big data and artificial intelligence platforms.

[0093] In an optional embodiment, the publishing terminal 101 may include, but is not limited to, electronic devices such as smart phones, desktop computers, tablet computers, laptop computers, smart speakers, digital assistants, augmented reality (AR) / virtual reality (VR) devices, and smart wearable devices. Optionally, the operating system running on the electronic device may include, but is not limited to, Android, IOS, Linux, and Windows.

[0094] In addition, it should be noted that in practical applications, Figure 1 The application environment shown can be applied to scenarios such as financial industry information release and customer service, and this manual does not impose any restrictions.

[0095] In order to solve the problems existing in the prior art, the embodiments of this specification provide a method, device, equipment and storage medium for intelligent review of picture information. Figure 2 The flowchart of a method for intelligent review of image information provided by an embodiment of this specification is shown in this figure, which describes the process of reviewing and guiding personnel image information. The order of steps listed in the embodiment is only one way of executing the steps among many orders, and does not represent the only order of execution. When the actual system or device product is executed, it can be executed in the order of the method shown in the embodiment or the accompanying drawings or in parallel. Specifically, Figure 2As shown, the method may include:

[0096] Step 201: obtaining the picture content to be published from the content to be published of the publishing terminal, and inputting the picture content to be published into the intelligent review system;

[0097] Step 202: the intelligent review system pre-processes the image content to be published to generate a plurality of images to be reviewed and text content to be reviewed;

[0098] Step 203: Identify and locate icon features, object features, face features, and auxiliary features for identifying faces in the plurality of images to be reviewed;

[0099] Step 204: Establishing a corresponding data analysis model according to the business scenarios and review strategies involved in the content to be published;

[0100] Step 205: determining icon features, object features, face features, and auxiliary features for identifying faces in the plurality of images to be reviewed according to corresponding data analysis models, and obtaining illegal image data;

[0101] Step 206: judging the content of the text to be reviewed according to the corresponding data analysis model, and obtaining illegal text data;

[0102] Step 207: Analyze the illegal text data and illegal image data to generate image review results, and return them to the publishing terminal.

[0103] Through the method of the embodiment of this specification, it is realized that the content to be published in the publishing terminal is reviewed by the intelligent review system. First, the image content to be published is obtained from the content to be published in the publishing terminal and input into the intelligent review system, so as to ensure that the intelligent review system obtains the complete image content to be published. Then, the intelligent review system pre-processes the image content to be published to generate multiple images to be reviewed and text content to be reviewed, so as to conduct image and text review respectively, and at the same time avoid useless interference information in the image content from affecting the review process. Then, according to the business scenarios and review strategies involved in the content to be published, a data analysis model is established, so as to improve the pertinence of the data analysis model, avoid the audit efficiency being affected by excessive data volume or the audit efficiency being too low due to insufficient pertinence. At the same time, according to the data analysis model, the icon features and object features in the image to be reviewed are identified, and the face and its facial information can also be identified to identify them, so as to realize the recognition of multiple variants through the learning model, and finally generate the image review result and return it to the publishing terminal. Image content review technology based on deep learning algorithms can identify illegal images and texts through variant recognition, semantic analysis and other technologies; it improves the efficiency of image review and avoids missed inspections caused by users maliciously evading inspections. At the same time, it uses customized review models for different business scenarios in the banking industry to improve the targetedness of the review model, thereby improving the efficiency of text data review.

[0104] In the embodiments of this specification, the publishing terminal can call the intelligent audit system through the API interface, and the intelligent audit system provides synchronous audit services, asynchronous audit services and result query services to the outside world; wherein the intelligent audit system can identify illegal patterns, illegal words and their variants based on the deep learning model through variant recognition, semantic analysis, context association and other technologies. After the intelligent audit system obtains the business demand text audit request, it first authenticates according to the application information and key to determine whether the publishing terminal has the authority to call the intelligent audit service, and then completes the data preprocessing of the text to be audited, establishes variant recognition, semantic analysis, context association and other data analysis models according to the business scenario and audit strategy, and audits the content to be audited. By customizing the audit model and risk label system according to different business scenarios of the banking industry, it meets the needs of business development, avoids the false positives and missed positives of a single model, and has a high accuracy rate in identifying violations.

[0105] In the embodiment of the present specification, in order to improve the efficiency of text review, the intelligent review system pre-processes the image content to be published, and generates multiple images to be reviewed and text content to be reviewed, further comprising:

[0106] The preprocessing includes locating and identifying the pattern images and text images in the image content to be published;

[0107] The image content to be published is segmented according to the pattern image and the text image to generate the plurality of images to be reviewed and the text content to be reviewed.

[0108] In the embodiments of the present specification, it is necessary to detect the text and icons in the image to be reviewed so as to review them separately, and to review the image content on the basis of ensuring the correctness of the text review. Since the posters, pictures and other contents that need to be reviewed in the marketing scenario may include artistic fonts or special typeset fonts, if special fonts are reviewed as images, misjudgment or omission may occur. Therefore, the pattern images and text images are reviewed separately in the preprocessing stage.

[0109] In the embodiment of this specification, in order to reduce the image content that needs to be reviewed by the data analysis model, it is necessary to screen the image content to be reviewed, wherein the preprocessing further includes:

[0110] Performing text recognition on the text images in the image content to be published, restoring the text in special fonts or special layouts, and generating text content;

[0111] The text content is subjected to text denoising, normalization of information of the same category, word segmentation, and removal of stop words; and the text content to be reviewed is generated.

[0112] Specifically, the text appearing in the image is identified separately. When encountering tilted, inverted, affine transformed, vertical, handwritten and other special fonts and layouts, data-driven text restoration is used and the rotation angle is added to the recognition network to generate text content; and the text denoising is to delete special symbols in the image content to be published that are not helpful for Chinese word segmentation, such as special punctuation marks such as "·", "*", and "#", so as to prevent publishers of harmful information from using such symbols to segment the original keywords and thus evade the censorship mechanism; normalization of information of the same category is to use regular expressions to replace the account number, mobile phone number, emoticons and other content in the image content to be published with names, so as to prevent publishers from using numbers and other content to evade the censorship mechanism, and at the same time protect the account login privacy information in the content to be published, and avoid the risk of relevant information being obtained by the auditors in the process of generating the audit results.

[0113] In the embodiment of the present specification, in order to improve the pertinence of the data analysis model, in step 204, establishing a corresponding data analysis model according to the business scenario and the review strategy involved in the content to be published further includes:

[0114] Selecting target patterns and target words according to the business scenarios involved in the content to be published;

[0115] Performing multiple deformation processes on the target pattern to generate a pattern recognition library;

[0116] The target word is expanded based on the target word and a preset semantic similarity to generate a scene vocabulary library.

[0117] For example, different target patterns are involved in specific business scenarios. For example, in a bank's wealth management product marketing posters, it is necessary to avoid patterns of other marketing activities, icons of other banks, and various junk advertisements. These patterns are used as target patterns and are subjected to basic deformation processing such as tilting and inverting to generate a pattern recognition library for targeted identification. At the same time, each time the intelligent audit system audits the illegal pattern data, the illegal pattern data will be stored and a target pattern will be generated. At the same time, after each audit, the target pattern library will be updated according to the new audit results to ensure its accuracy.

[0118] Furthermore, in order to establish the data analysis model, further expansion is required, such as Figure 3 As shown, establishing a data analysis model according to the business scenarios and review strategies involved in the content to be published further includes:

[0119] The audit strategy includes at least one of the following: variant identification, semantic analysis, and context association;

[0120] Step 301: obtaining a pattern recognition model of the pattern recognition library through a preset variant pattern algorithm;

[0121] Step 302: obtaining a variant recognition model of the scene vocabulary library through a preset variant word algorithm;

[0122] Step 303: obtaining a semantic recognition model of the scene vocabulary library through a preset semantic algorithm;

[0123] Step 304: Acquire the association recognition model of the scene vocabulary library through a preset context association algorithm.

[0124] In the embodiments of this specification, in order to determine the content of the image to be reviewed, Figure 4 The method further comprises: determining the icon features, object features, face features, and auxiliary features for identifying faces in the plurality of images to be reviewed according to the corresponding data analysis model, and obtaining the illegal image data further comprises:

[0125] Step 401: determining the icon features and item features in the plurality of images to be reviewed according to the pattern recognition model;

[0126] Step 402: Generate facial information according to the facial features and auxiliary features for identifying the face, and determine the facial information;

[0127] Step 403: Generate the illegal image data according to the determination results of the icon features and the object features and the determination results of the face information.

[0128] Specifically, image content review technology based on deep learning algorithms, such as Figure 5 As shown, first, the image will be classified and identified into icon features, object features, facial features, and auxiliary features for identifying faces. The icon features and object features are judged by the pattern recognition model to determine that the patterns and deformations that need to be avoided in the pattern recognition library are not included; secondly, the facial features in the image to be reviewed will be located and analyzed, and the auxiliary features such as the task body, clothing, and movements used to identify the face will be located and feature identified to complete a secondary verification, and finally the judgment result of the facial information will be output. By way of example, for special cases such as blurred faces, small faces, cartoons, side faces, and rotated faces that appear in the image, the auxiliary features of the face are used to locate the faces to avoid missing the faces, and then the face information is judged to confirm that the identity of the person corresponding to the face information meets the requirements of the business scenario and the review strategy, and does not contain other illegal content.

[0129] In the embodiment of this specification, in order to quickly and accurately identify illegal text data, the content of the text to be reviewed is determined according to the corresponding data analysis model, and obtaining the illegal text data further includes:

[0130] The text content to be reviewed is input into the variant recognition model, the semantic recognition model and the association recognition model to generate the illegal text data.

[0131] Specifically, the suspicious text data is input into the variant recognition model to obtain the illegal variant recognition result corresponding to the suspicious text data; the suspicious text data is input into the semantic recognition model to obtain the illegal semantic recognition result corresponding to the suspicious text data; the suspicious text data is input into the association recognition model to obtain the illegal association recognition result corresponding to the suspicious text data; the illegal text data is generated according to the illegal variant recognition result, the illegal semantic recognition result and the illegal association recognition result. In the process of identifying suspicious text data, a cutting-edge natural language processing algorithm is used, and the business scenarios and audit strategies involved in the content to be published are combined with a large number of sensitive keyword libraries for modeling to quickly and accurately identify illegal text data. For example, the finite state automaton (FSA) is applied to variant recognition, which can effectively improve the recall rate of online variant black samples while ensuring accuracy; a complex model system built based on NLP technologies such as Fasttext, HMM, CRF, Word2Vec, etc. is used to obtain the illegal semantic recognition results and illegal association recognition results corresponding to the suspicious text data. At the same time, through the joint recognition of multiple models, the misjudgment and omission of a single model can be effectively avoided, and the violation recognition is highly accurate.

[0132] In the embodiment of this specification, in order to determine whether the illegal text data and illegal image data affect the release of the content to be released, Figure 6 As shown, analyzing the illegal text data and illegal image data to generate image review results further includes:

[0133] Step 601: judging the confidence level of the to-be-published picture content according to the quantity of the illegal text data and illegal picture data;

[0134] Step 602: performing risk classification according to the contents of the illegal text data and illegal image data;

[0135] Step 603: Generate violation handling opinions and violation content distribution as the image review result according to the confidence level of the image content to be released and the risk classification.

[0136] Specifically, in one embodiment of the present specification, the confidence level of the to-be-published picture content can be obtained based on the number of illegal text data and illegal picture data. For example, the larger the confidence level score, the greater the risk of the to-be-published content; at the same time, the risk classification of the illegal content is determined by judging the content of the illegal data, such as pornography, terrorism, insults, advertisements, etc.; then, according to the confidence level and the risk classification, a violation handling opinion is generated. For example, if the confidence level score is lower than the threshold and the illegal content is a normal picture such as an advertisement, it is marked as approved; if the confidence level score of the illegal text is higher than the threshold or the illegal content is pornography, terrorism, insults, etc., which are more harmful, it is marked as rejecting the publishing request; if a suspicious picture is encountered and further manual review is required, it is marked as suspicious text. At the same time, the distribution of illegal content is marked in the picture review results, such as the location information of the illegal picture, so as to help the reviewer quickly locate it.

[0137] In the embodiment of this specification, in order to improve the audit efficiency of the intelligent audit system, the intelligent audit system further includes synchronous audit and asynchronous audit.

[0138] If the image content to be published is a small image file, the synchronous review service is called;

[0139] If the image content to be published is a large image file, the asynchronous review service is called.

[0140] Among them, the synchronous review service is for smaller image files, such as avatars, photos, etc., and the synchronous review service can be called, and the review results are returned in real time; the asynchronous review service is for large image files, such as posters, roll-up banners, etc., and the asynchronous review service can be called.

[0141] Specifically, when the intelligent audit system calls different audit services, the way the audit results are returned will also change accordingly. The analysis of the illegal image data generates the image audit results, and returns them to the publishing terminal, further including:

[0142] If the synchronous audit service is called, the audit result will be returned directly;

[0143] If the asynchronous audit service interface is called, the audit serial number is returned, and the publishing terminal calls the audit result query interface for polling according to the serial number, or the publishing terminal sets a callback address, and the intelligent audit system sends the audit result to the callback address.

[0144] Among them, the synchronous audit service is used for the audit needs of small image files, and the audit results are returned in real time; while the asynchronous audit service is used for the audit needs of large image files. The asynchronous audit service is called, and the publishing terminal can obtain the audit results by polling the result query interface or setting the callback callback address. At the same time, depending on the different calling interfaces of the publishing terminal, if the synchronous audit service interface is called, the audit result is returned directly; if the asynchronous audit service interface is called, the transaction serial number is returned first. The business provider can call the audit result query interface to poll according to the serial number, or the publishing terminal can set the callback callback address, and the intelligent audit system will send the audit result to the callback address.

[0145] The embodiment of this specification also provides a picture information intelligent review device, such as Figure 7 As shown, including,

[0146] The content acquisition unit 701 is used to acquire the picture content to be published from the content to be published of the publishing terminal, and input the picture content to be published into the intelligent review system;

[0147] A preprocessing unit 702, used for the intelligent audit system to preprocess the image content to be published, and generate a plurality of images to be audited and text content to be audited;

[0148] A content positioning unit 703, used to identify and locate icon features, object features, face features, and auxiliary features for identifying faces in the plurality of images to be reviewed;

[0149] The model building unit 704 is used to determine the icon features, object features, face features, and auxiliary features for identifying faces in the multiple images to be reviewed according to the corresponding data analysis model to obtain illegal image data;

[0150] The illegal image acquisition unit 705 is used to determine the icons, objects, faces and related features in the multiple images to be reviewed according to the data analysis model to obtain illegal image data;

[0151] The illegal text acquisition unit 706 is used to determine the content of the text to be reviewed according to the corresponding data analysis model and obtain illegal text data;

[0152] The result generating unit 707 is used to analyze the illegal text data and illegal image data to generate image review results, and return them to the publishing terminal.

[0153] Since the principle of solving the problem by the above device is similar to that of the above method, the implementation of the above device can refer to the implementation of the above method, and the repeated parts will not be repeated.

[0154] like Figure 8The structure diagram of the computer device of the embodiment of the present invention is shown. The apparatus in this specification can be the computer device in this embodiment, and executes the method of the above-mentioned specification. The computer device 802 may include one or more processing devices 804, such as one or more central processing units (CPUs), and each processing unit may implement one or more hardware threads. The computer device 802 may also include any storage resource 806, which is used to store any kind of information such as code, settings, data, etc. Non-limitingly, for example, the storage resource 806 may include any one or more combinations of the following: any type of RAM, any type of ROM, flash memory device, hard disk, optical disk, etc. More generally, any storage resource can use any technology to store information. Further, any storage resource can provide volatile or non-volatile retention of information. Further, any storage resource can represent a fixed or removable component of the computer device 802. In one case, when the processing device 804 executes an associated instruction stored in any storage resource or a combination of storage resources, the computer device 802 can perform any operation of the associated instruction. The computer device 802 also includes one or more drive mechanisms 808 for interacting with any storage resources, such as a hard disk drive mechanism, an optical disk drive mechanism, and the like.

[0155] The computer device 802 may also include an input / output module 810 (I / O) for receiving various inputs (via input devices 812) and for providing various outputs (via output devices 814). A specific output mechanism may include a presentation device 816 and an associated graphical user interface (GUI) 818. In other embodiments, the input / output module 810 (I / O), the input device 812, and the output device 814 may not be included, and the computer device 802 may be used as a computer device in a network. The computer device 802 may also include one or more network interfaces 820 for exchanging data with other devices via one or more communication links 822. One or more communication buses 824 couple the components described above together.

[0156] The communication link 822 may be implemented in any manner, for example, through a local area network, a wide area network (e.g., the Internet), a point-to-point connection, etc., or any combination thereof. The communication link 822 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc. governed by any protocol or combination of protocols.

[0157] Corresponds to Figures 2 to 6 The method in the embodiment of the present specification also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described above are executed.

[0158] The embodiment of the present specification also provides a computer-readable instruction, wherein when the processor executes the instruction, the program therein causes the processor to execute the following Figures 2 to 6 The method shown.

[0159] It should be understood that in the various embodiments of this specification, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0160] It should also be understood that in the embodiments of this specification, the term "and / or" is only a description of the association relationship of the associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this specification generally indicates that the associated objects before and after are in an "or" relationship.

[0161] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this specification.

[0162] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0163] In the several embodiments provided in this specification, 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 only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, or it can be an electrical, mechanical or other form of connection.

[0164] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of this specification.

[0165] In addition, each functional unit in each embodiment of this specification may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0166] If 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 specification is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of this specification. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

[0167] Specific embodiments are used in this specification to illustrate the principles and implementation methods of this specification. The description of the above embodiments is only used to help understand the methods and core ideas of this specification. At the same time, for those skilled in the art, according to the ideas of this specification, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on this specification.

Claims

1. A method for intelligent review of image information, characterized in that: The method comprises, Acquire the picture content to be published from the content to be published of the publishing terminal, and input the picture content to be published into the intelligent review system; The intelligent audit system pre-processes the image content to be published to generate multiple images to be audited and text content to be audited; Identify and locate icon features, object features, face features, and auxiliary features for identifying faces in the plurality of images to be reviewed; Establish a corresponding data analysis model based on the business scenarios and review strategies involved in the content to be published; According to the corresponding data analysis model, the icon features, object features, face features, and auxiliary features for identifying faces in multiple images to be reviewed are determined to obtain illegal image data; Determine the content of the text to be reviewed according to the corresponding data analysis model to obtain illegal text data; The illegal text data and illegal image data are analyzed to generate image review results, and returned to the publishing terminal.

2. The method for intelligent review of picture information according to claim 1, characterized in that: The intelligent audit system pre-processes the content of the picture to be published, and generates a plurality of pictures to be audited and text content to be audited, further comprising: The preprocessing includes locating and identifying the pattern images and text images in the image content to be published; The image content to be published is segmented according to the pattern image and the text image to generate the plurality of images to be reviewed and the text content to be reviewed.

3. The method for intelligent review of picture information according to claim 2, characterized in that: The pre-processing further comprises, Performing text recognition on the text images in the image content to be published, restoring the text in special fonts or special layouts, and generating text content; Performing text denoising, normalizing information of the same category, segmenting words and sentences, and removing stop words on the text content; Generate the text content to be reviewed.

4. The method for intelligent review of picture information according to claim 1, characterized in that: Establishing a corresponding data analysis model according to the business scenarios and review strategies involved in the content to be published further includes: Selecting target patterns and target words according to the business scenarios involved in the content to be published; Performing multiple deformation processes on the target pattern to generate a pattern recognition library; The target word is expanded based on the target word and a preset semantic similarity to generate a scene vocabulary library.

5. The method for intelligent review of picture information according to claim 4, characterized in that: Establishing a data analysis model according to the business scenarios and review strategies involved in the content to be published further includes: The audit strategy includes at least one of the following: variant identification, semantic analysis, and context association; Acquiring a pattern recognition model of the pattern recognition library through a preset variant pattern algorithm; Obtaining a variant recognition model of the scene vocabulary library through a preset variant word algorithm; Acquire a semantic recognition model of the scene vocabulary library through a preset semantic algorithm; The association recognition model of the scene vocabulary library is obtained through a preset context association algorithm.

6. The method for intelligent review of picture information according to claim 5, characterized in that: According to the corresponding data analysis model, the icon features, object features, face features, and auxiliary features for identifying faces in the multiple images to be reviewed are determined, and obtaining the illegal image data further includes: Determining the icon features and the object features in the plurality of images to be reviewed according to the pattern recognition model; Generate facial information according to the facial features and auxiliary features for identifying the face, and determine the facial information; The illegal image data is generated according to the determination results of the icon features and the object features and the determination results of the face information.

7. The method for intelligent review of picture information according to claim 5, characterized in that: Determining the content of the text to be reviewed according to the corresponding data analysis model, obtaining the illegal text data further includes: The text content to be reviewed is input into the variant recognition model, the semantic recognition model and the association recognition model to generate the illegal text data.

8. The method for intelligent review of picture information according to claim 1, characterized in that: Analyzing the illegal text data and illegal image data to generate image review results further includes: Determining the confidence level of the to-be-published picture content according to the amount of illegal text data and illegal picture data; Perform risk classification according to the contents of the illegal text data and illegal image data; According to the confidence level of the content of the picture to be released and the risk classification, a violation handling opinion and a distribution of violation content are generated as the picture review result.

9. The method for intelligent review of picture information according to claim 1, characterized in that: The intelligent audit system further includes synchronous audit and asynchronous audit. If the image content to be published is a small image file, the synchronous review service is called; If the image content to be published is a large image file, the asynchronous review service is called.

10. The method for intelligent review of picture information according to claim 9, characterized in that: Analyzing the illegal text data to generate image review results and returning them to the publishing terminal further includes: If the synchronous audit service is called, the audit result will be returned directly; If the asynchronous audit service interface is called, the audit serial number is returned, and the publishing terminal calls the audit result query interface for polling according to the serial number, or the publishing terminal sets a callback address, and the intelligent audit system sends the audit result to the callback address.

11. An intelligent image information review device, characterized in that: include, A content acquisition unit, used to acquire the picture content to be published from the content to be published of the publishing terminal, and input the picture content to be published into the intelligent review system; A preprocessing unit, used for the intelligent audit system to preprocess the image content to be published, and generate a plurality of images to be audited and text content to be audited; A content positioning unit, used to identify and locate icon features, object features, face features, and auxiliary features for identifying faces in the plurality of images to be reviewed; A model building unit, used to determine the icon features, object features, face features, and auxiliary features for identifying faces in multiple images to be reviewed according to a corresponding data analysis model, and obtain illegal image data; An illegal picture acquisition unit, used to determine icons, objects, faces and related features in a plurality of images to be reviewed according to the data analysis model, and acquire illegal picture data; An illegal text acquisition unit, used to determine the content of the text to be reviewed according to a corresponding data analysis model, and acquire illegal text data; The result generating unit is used to analyze the illegal text data and illegal picture data to generate picture review results, and return them to the publishing terminal.

12. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 10 is implemented.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 10 is implemented.

14. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 10 is implemented.

Citation Information

Cited By

  • Digital media content processing method, device and system

    CN119420970A

  • Image content auditing method and system based on multilayer scene graph structure

    CN121861291A