Rich media intelligent information editing and auditing method and device

Abstract

The embodiment of the invention provides a rich media intelligent information editing and auditing method and device, and the method comprises the steps: receiving and storing a rich media resource file uploaded by a user, and carrying out the file recognition of the rich media resource file, carrying out file segmentation on the rich media resource file through a self-adaptive segmentation algorithm according to the file type, the file content characteristics and a preset size threshold obtained through identification to obtain one or more file fragments; performing multi-modal content vectorization on the file fragment through a preset rich media vectorization model, extracting to obtain a corresponding semantic meaning array, and determining the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, displaying information of similar rich media resource files in a preset user file editing interface, wherein the similarity between the similar rich media resource files and the current rich media resource files exceeds a preset threshold value; according to the invention, rich media information can be processed more flexibly.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and specifically to a method and device for editing and reviewing rich media intelligent information. Background Art

[0002] In today's era of information explosion, the rapid growth of rich media content has brought huge challenges to information editing. Traditional information editing methods have been unable to meet the requirements of efficiency, accuracy and consistency when facing massive and diverse rich media content. Existing technologies have many limitations and problems in processing rich media information editing.

[0003] First, traditional information editing and review methods mainly rely on manual operations. Editors need to review and edit rich media files one by one, which is not only time-consuming and labor-intensive, but also easily affected by personal subjective judgment, making it difficult to ensure the consistency of review standards. When faced with a large amount of complex rich media content, the efficiency of manual methods is even more difficult to meet the needs, which easily leads to work backlogs and delays.

[0004] Secondly, some existing automated editing tools can only process a single type of media file, such as text or images. However, in reality, rich media content usually contains mixed content in multiple forms such as text, images, audio, and video. These tools are unable to handle complex multimodal content and cannot fully grasp the overall semantics and content characteristics of rich media resources. Summary of the invention

[0005] In view of the problems in the prior art, the present application provides a rich media intelligent information editing method and device, which can process rich media information more flexibly.

[0006] In order to solve at least one of the above problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a rich media intelligent information editing method, comprising:

[0008] Receive and store rich media resource files uploaded by users, perform file identification on the rich media resource files, and perform file segmentation on the rich media resource files using an adaptive segmentation algorithm according to the identified file types, file content features, and a preset size threshold to obtain one or more file segments;

[0009] The file fragment is multimodally vectorized by a preset rich media vectorization model, and a corresponding semantic meaning array is extracted; the similarity between other rich media resource files and the current rich media resource file is determined by a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold is displayed in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type;

[0010] Receive the file editing instruction sent by the user, integrate the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handle the conflict of the similar file contents through a conflict resolution mechanism, and display the file content after the content integration and conflict handling in the preset user file editing interface.

[0011] Furthermore, the performing file identification on the rich media resource file includes:

[0012] Read the binary data stream of the rich media resource file, extract the magic number information and the file extension in the file header, match the extracted magic number information with a preset file type magic number database, and determine a preliminary file type;

[0013] Parse the metadata structure of the file, extract key attribute information, and use the pre-trained deep learning model to extract feature vectors of the file content to obtain file content features. The preliminary file type, the key attribute information and the content feature vector are comprehensively input into the decision tree classifier to obtain the final file type recognition result.

[0014] Furthermore, the rich media resource file is segmented by an adaptive segmentation algorithm according to the identified file type, file content features and preset size threshold to obtain one or more file segments, including:

[0015] According to the identified file type, the corresponding content feature extraction algorithm and segmentation strategy are selected to calculate the total size of the file, and compared with the preset size threshold to determine the initial number of segments;

[0016] The file content is analyzed by a selected feature extraction algorithm to generate a content feature sequence, and an adaptive segmentation algorithm based on dynamic programming is applied to calculate the optimal segmentation point. The file segmentation operation is performed at the calculated segmentation point to generate one or more file fragments, and metadata containing the original file information, segmentation position and fragment sequence number is created for each of the file fragments.

[0017] Furthermore, the multimodal content vectorization of the file fragment is performed by using a preset rich media vectorization model to extract a corresponding semantic meaning array, including:

[0018] Selecting a corresponding pre-trained model according to the type of the file segment to perform format conversion, noise reduction and standardization pre-processing, inputting the pre-processed data into the selected pre-trained model, and extracting a high-dimensional feature vector representation;

[0019] The high-dimensional feature vector representation is input into a pre-trained semantic understanding model, the key semantic features in the high-dimensional feature vector are extracted through the multi-layer self-attention mechanism of the semantic understanding model, the extracted semantic features are clustered using a hierarchical clustering algorithm to form semantic concepts at different levels, the clustering results are mapped to a predefined semantic ontology library, normalized semantic labels are obtained and sorted according to importance, and a structured semantic meaning array is constructed, wherein the elements in the semantic meaning array include semantic labels, importance scores and related metrics.

[0020] Furthermore, the determining the similarity between other rich media resource files and the current rich media resource file by using a multi-dimensional weighted similarity algorithm according to the semantic meaning array includes:

[0021] Map each semantic label in the semantic meaning array to a predefined feature space, assign a weight to each dimension according to the importance score of the semantic label, and generate a weighted feature vector as the input for similarity calculation;

[0022] The similarities of the three dimensions of content, structure and semantics are calculated separately, the weight of each dimension is determined according to the file type, and the weighted similarity results of the three dimensions are combined to obtain the final similarity score.

[0023] Furthermore, the receiving of the file editing instruction sent by the user and integrating the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction includes:

[0024] Parse the contents of the file editing instructions sent by the user, identify the instruction type and operation parameters, and convert the parsed instructions into an operation sequence that can be executed by the system;

[0025] Load the contents of the current rich media resource file and similar rich media resource files selected by the user, select the corresponding content integration algorithm according to the file type to perform the content integration operation, generate a new integrated rich media resource file, and retain the version information of the original file, wherein the content integration operation includes content merging, replacing, deleting or inserting.

[0026] Furthermore, the conflict resolution mechanism is used to handle the conflict of similar file contents, including:

[0027] Compare the content structures of the current rich media resource file and similar rich media resource files to identify repeated, contradictory or inconsistent content segments, and mark conflict locations and types;

[0028] Handle some conflicts according to the conflict type and preset resolution strategy. For conflicts that cannot be resolved automatically, generate corresponding user options, receive and execute the user's conflict resolution selection, update the integrated rich media resource file content, and eliminate the conflict.

[0029] In a second aspect, the present application provides a rich media intelligent information editing and reviewing device, comprising:

[0030] A file segmentation module is used to receive and store rich media resource files uploaded by users, perform file identification on the rich media resource files, and perform file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file types, file content characteristics, and a preset size threshold to obtain one or more file segments;

[0031] A similarity display module, used to perform multimodal content vectorization on the file fragment through a preset rich media vectorization model, extract a corresponding semantic meaning array, determine the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and display information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type;

[0032] The similarity merging module is used to receive the file editing instruction sent by the user, integrate the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handle the conflict of the content of similar files through the conflict resolution mechanism, and display the file content after the content integration and conflict handling in the preset user file editing interface.

[0033] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the rich media intelligent information editing method when executing the program.

[0034] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the rich media intelligent information editing method when executed by a processor.

[0035] In a fifth aspect, the present application provides a computer program product, including a computer program / instruction, which implements the steps of the rich media intelligent information editing method when executed by a processor.

[0036] It can be seen from the above technical scheme that the present application provides a rich media intelligent information editing method and device, which receives and stores rich media resource files uploaded by users, performs file identification on the rich media resource files, and performs file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file type, file content characteristics and preset size threshold, so as to obtain one or more file fragments; performs multimodal content vectorization on the file fragments through a preset rich media vectorization model, extracts the corresponding semantic meaning array, determines the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and displays the information of similar rich media resource files whose similarity with the current rich media resource file exceeds the preset threshold in the preset user file editing interface, thereby being able to process rich media information more flexibly. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] 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 some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0038] Figure 1 This is one of the flowcharts of the rich media intelligent information editing method in the embodiment of the present application;

[0039] Figure 2 This is a second flow chart of the rich media intelligent information editing method in the embodiment of the present application;

[0040] Figure 3 This is a third flow chart of the method for editing and reviewing rich media intelligent information in an embodiment of the present application;

[0041] Figure 4 This is a fourth flow chart of the rich media intelligent information editing method in the embodiment of the present application;

[0042] Figure 5 This is a fifth flow chart of the rich media intelligent information editing method in the embodiment of the present application;

[0043] Figure 6 This is a sixth flow chart of the rich media intelligent information editing method in the embodiment of the present application;

[0044] Figure 7 FIG7 is a flowchart of the method for editing and reviewing rich media intelligent information in an embodiment of the present application;

[0045] Figure 8 It is a structural diagram of the rich media intelligent information editing and reviewing device in the embodiment of the present application;

[0046] Fig. 9 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application.

[0047] Reference numerals:

[0048] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are 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.

[0050] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0051] In view of the problems existing in the prior art, the present application provides a rich media intelligent information editing method and device, which receives and stores rich media resource files uploaded by users, performs file identification on the rich media resource files, and performs file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file type, file content characteristics and preset size threshold, so as to obtain one or more file fragments; performs multimodal content vectorization on the file fragments through a preset rich media vectorization model, extracts the corresponding semantic meaning array, determines the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and displays the information whose similarity with the current rich media resource file exceeds the preset threshold in the preset user file editing interface, thereby processing rich media information more flexibly.

[0052] In order to process rich media information more flexibly, the present application provides an embodiment of a rich media intelligent information editing method, see Figure 1 The rich media intelligent information editing method specifically includes the following contents:

[0053] Step S101: receiving and storing a rich media resource file uploaded by a user, performing file identification on the rich media resource file, and performing file segmentation on the rich media resource file by an adaptive segmentation algorithm according to the identified file type, file content characteristics, and a preset size threshold, to obtain one or more file segments;

[0054] Optionally, in this embodiment, step S101 is the first key step of the rich media intelligent information editing method, which mainly involves the technical process of file reception, storage, identification and segmentation. The implementation of this step not only lays the foundation for subsequent content analysis and processing, but also solves the efficiency problem of large rich media file processing.

[0055] In the specific implementation process, the system first needs to establish a stable file upload and storage mechanism. This may involve the use of a distributed file system or cloud storage solution to ensure that a large number of concurrent upload requests can be efficiently processed and various types of rich media files can be safely and reliably stored. For example, the system may use an object storage service to store uploaded files, which can automatically perform data redundancy and backup to ensure the security and availability of files.

[0056] After receiving the file, the system will perform file identification on the uploaded rich media resource file. This process involves a variety of technologies, including file header analysis, content feature extraction, and machine learning classification. The system will first check the magic number and MIME type of the file, which can usually quickly determine the basic type of the file. For more complex files, such as compound documents, the system may need to further analyze the file structure and content characteristics.

[0057] For example, for a video file, the system not only needs to identify it as a video file, but also needs to determine its encoding format, resolution, frame rate, etc. This information is crucial for subsequent processing, because files of different types and characteristics may require different segmentation and analysis strategies.

[0058] After the file recognition is completed, the system will use the adaptive segmentation algorithm to segment the file according to the recognition results and the preset size threshold. This adaptive segmentation algorithm is one of the core technologies of this step, which can dynamically adjust the segmentation strategy according to the file type and content characteristics.

[0059] For text files, the segmentation algorithm may consider paragraph structure, chapter divisions, or semantic completeness. For example, a long document may be segmented by chapters or topics rather than simply by a fixed word count. For image files, the algorithm may consider the content areas of the image, such as the boundaries of important objects, rather than mechanically chopping the image into equal-sized chunks. For video files, segmentation may be based on scene changes, audio features, or semantic boundaries of the content.

[0060] This adaptive segmentation method solves the content integrity and coherence problems caused by traditional fixed-size segmentation. It can better preserve the semantic structure of the file, which is conducive to subsequent content analysis and processing. At the same time, through reasonable segmentation, the system can more effectively utilize computing resources and improve the efficiency of processing large files.

[0061] The main technical problems solved in this step include: how to efficiently process large rich media files, how to accurately identify different types of media files, and how to effectively segment files while maintaining content integrity. By implementing these technologies, this step achieves the technical effects of improving file processing efficiency, enhancing content analysis accuracy, and improving user experience.

[0062] Specifically, the adaptive segmentation algorithm enables the system to more intelligently process rich media files of various types and sizes, avoiding the content fragmentation problem that may be caused by traditional fixed-size segmentation methods. This not only improves the accuracy of subsequent content analysis, but also provides users with a better editing and viewing experience. For example, when users need to edit a specific part of the video content, they can directly operate on the relevant video clips without having to process the entire large file, which greatly improves editing efficiency.

[0063] In addition, this step also lays the foundation for subsequent parallel processing. Since the file is reasonably divided into multiple segments, the system can process these segments simultaneously, making full use of distributed computing resources and significantly improving the processing speed. This is of great significance for application scenarios that require real-time response, such as online video editing or real-time content analysis.

[0064] In general, step S101 not only improves the efficiency and flexibility of the system in processing rich media files through intelligent file processing and segmentation strategies, but also creates favorable conditions for subsequent content analysis and editing operations, and is an important foundation for the entire rich media intelligent information editing method.

[0065] Step S102: multimodal content vectorization is performed on the file fragment through a preset rich media vectorization model to extract a corresponding semantic meaning array, and similarity between other rich media resource files and the current rich media resource file is determined through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold is displayed in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type;

[0066] Optionally, in this embodiment, step S102 is the core link in the rich media intelligent information editing method, which mainly involves content vectorization, semantic analysis and similarity calculation. The technical process and principle of this step are intended to convert different types of rich media content into comparable mathematical representations and perform intelligent similarity analysis based on these representations.

[0067] First, the system uses a preset rich media vectorization model to vectorize the multimodal content of the file fragments. The technical principle of this process is based on deep learning and representation learning theory, and aims to convert data of different modalities (such as text, images, audio, and video) into a unified vector space representation. For text content, the system may use a pre-trained language model based on the Transformer architecture, such as BERT or GPT; for image content, it may use a convolutional neural network (CNN) or a visual Transformer model; for audio content, it may use waveform-to-spectrum conversion combined with a deep neural network; and for video content, it may combine temporal modeling techniques, such as long short-term memory networks (LSTM) or 3D convolutional networks.

[0068] This multimodal vectorization technology solves the technical problem of how to uniformly process different types of media content. By mapping various types of content into the same vector space, the system can represent and compare information of different modalities in a consistent manner, laying the foundation for subsequent similarity calculations.

[0069] After vectorization, the system extracts the corresponding array of semantic meanings. This process is actually a further semantic analysis and abstraction of the vector representation. The technical principles may involve technologies such as attention mechanism, semantic segmentation or topic modeling. For example, for text content, the system may extract keywords, themes and emotional tendencies; for image content, it may identify the main objects, scenes and visual styles; for audio content, it may analyze timbre, rhythm and emotional characteristics; and for video content, it may comprehensively analyze visual and auditory features to extract temporal semantic information.

[0070] The extraction of this array of semantic meanings solves the technical problem of how to extract high-level semantic information from low-level features, allowing the system to more accurately understand and compare the actual meaning of rich media content rather than just the surface features.

[0071] Next, the system uses a multi-dimensional weighted similarity algorithm to determine the similarity between other rich media resource files and the current file. This algorithm comprehensively considers three dimensions: content similarity, structural similarity, and semantic similarity. Content similarity focuses on the original features of the media, such as the word frequency distribution of text, the color and texture features of images, and the spectral features of audio; structural similarity focuses on the organization of media content, such as the chapter structure of documents, the spatial layout of images, and the temporal structure of videos; semantic similarity focuses on the high-level meaning and topic relevance of the content.

[0072] A key innovation of the algorithm is that it dynamically adjusts the calculation weights of each dimension according to the file type. This adaptive weight mechanism can better adapt to the characteristics of different types of media content. For example, for text files, semantic similarity may be given a higher weight; for image files, content similarity may be more important; for video files, it may be necessary to strike a balance between the three dimensions.

[0073] This multi-dimensional weighted similarity algorithm solves the technical problem of how to comprehensively and accurately evaluate the similarity of different types of rich media content. It not only considers the surface features of the content, but also pays attention to the deep structural and semantic information, so that it can more accurately identify truly relevant content.

[0074] Finally, the system displays the rich media resource file information that exceeds the preset threshold in the preset user file editing interface. This process involves user interface design and information visualization technology. The system may sort the results according to the similarity score and display the key information of similar files in thumbnails, summaries or other intuitive forms.

[0075] This intelligent similar content recommendation mechanism solves the technical problem of how to help users quickly find relevant resources. It can greatly improve user work efficiency and reduce the time of manual search and screening.

[0076] In general, step S102 achieves a number of important technical effects. First, it realizes the unified representation and processing of different types of rich media content, providing a technical basis for cross-modal content analysis and comparison. Secondly, through deep semantic analysis and multi-dimensional similarity calculation, the system can understand and match content more accurately, greatly improving the accuracy of similar content identification. Thirdly, the dynamic weight adjustment mechanism enables the system to adapt to different types of media content more flexibly, improving the versatility and robustness of the algorithm. Finally, through intelligent similar content recommendation, the system significantly improves the user's work efficiency and experience.

[0077] Step S103: receiving the file editing instruction sent by the user, integrating the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handling the conflict of the similar file contents through a conflict resolution mechanism, and displaying the file content after the content integration and conflict handling in the preset user file editing interface.

[0078] Optionally, in this embodiment, step S103 is a key link in the rich media intelligent information editing method, which mainly involves file editing, content integration and conflict handling. The core goal of this step is to provide users with an intelligent file editing environment that can effectively integrate the contents of multiple similar rich media resource files, resolve conflicts that may arise during the integration process, and ultimately present an optimized file content.

[0079] First, the system receives the file editing instructions sent by the user. This process involves human-computer interaction and user intent understanding technology. The system needs to accurately parse the user's editing intentions, which may include operations such as content merging, partial replacement, and structural reorganization. To achieve this goal, the system may use natural language processing technology to understand the user's text instructions, or capture the user's editing intentions through graphical user interface (GUI) operations. This intelligent parsing technology can greatly improve the user's operating efficiency and reduce tedious manual editing steps.

[0080] Next, the system integrates the content of the current rich media resource file and similar rich media resource files selected by the user according to the parsed editing instructions. The technical principle of this process is based on document structure analysis, content matching and intelligent merging algorithms. The system first needs to understand the structure and content of each file, which may involve document parsing technology, semantic analysis and topic modeling. Then, the system will determine the content parts that need to be integrated according to the user's editing instructions, and use intelligent merging algorithms to merge these contents.

[0081] This content integration technology solves the technical problem of how to effectively combine valuable information from multiple similar documents. It can help users quickly extract and combine the required information from multiple related resources, greatly improving the efficiency of content creation and knowledge integration. For example, when writing a research report, the system can intelligently integrate the content of multiple related documents to provide users with a comprehensive and concise information summary.

[0082] However, in the process of content integration, some conflicts may occur, such as content duplication, inconsistent information or structural conflicts. In order to solve these problems, the system adopts a conflict resolution mechanism. The technical principles of this mechanism may include version control algorithms, difference comparison technology and intelligent decision-making systems.

[0083] The conflict resolution mechanism first identifies and classifies different types of conflicts. For content duplication, the system may use text similarity algorithms to detect and merge duplicate content; for information inconsistency, the system may use rule-based logical reasoning or machine learning models to determine which information is more reliable or updated; for structural conflicts, the system may analyze the logical structure of the document and use heuristic algorithms to determine the most reasonable way to organize it.

[0084] When dealing with these conflicts, the system may make decisions automatically based on preset rules, or it may ask for the user's opinion at key points. This human-computer collaborative approach not only ensures the efficiency of processing, but also ensures that the final result meets the user's expectations. The application of the conflict resolution mechanism solves the technical problem of how to maintain consistency and logic in the complex content integration process, greatly improving the quality and usability of the integration results.

[0085] After content integration and conflict resolution, the system will display the processed file content in the preset user file editing interface. This display process involves information visualization and user interface design technology. The system may use highlights, annotations, or other visual effects to mark the integrated and processed parts so that users can clearly understand the changed content. At the same time, the system may also provide options for undoing, modifying, or further editing so that users can make fine adjustments or corrections.

[0086] This intelligent content display and interaction method solves the technical problem of how to enable users to intuitively understand and control complex editing results. It not only improves the user's operating efficiency, but also enhances the user's sense of control over the editing process, thereby improving the overall user experience.

[0087] In general, step S103 achieves several important technical effects. First, it realizes intelligent file editing and content integration, greatly improving the efficiency of users in processing complex rich media resources. Users can quickly extract valuable information from multiple similar files and integrate it into their current work, which has wide application value in research, writing, and content creation.

[0088] Secondly, through the conflict resolution mechanism, the system can intelligently handle various conflicts in the content integration process to ensure the consistency and rationality of the final result. This not only reduces the workload of users to manually resolve conflicts, but also improves the quality and reliability of the integration results.

[0089] Again, this step provides users with an intuitive and interactive editing environment, making the complex content integration process more transparent and controllable. Users can clearly see the results of the integration and make further adjustments as needed, which greatly enhances the user's operating experience and satisfaction.

[0090] From the above description, it can be seen that the rich media intelligent information editing method provided in the embodiment of the present application can receive and store rich media resource files uploaded by users, perform file identification on the rich media resource files, and perform file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file type, file content characteristics and preset size threshold to obtain one or more file fragments; perform multimodal content vectorization on the file fragments through a preset rich media vectorization model, extract the corresponding semantic meaning array, determine the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm based on the semantic meaning array, and display information whose similarity with the current rich media resource file exceeds the preset threshold in the preset user file editing interface, thereby more flexibly processing rich media information.

[0091] In one embodiment of the rich media intelligent information editing method of the present application, see Figure 2 , and can also include the following:

[0092] Step S201: reading the binary data stream of the rich media resource file, extracting magic number information and file extension in the file header, matching the extracted magic number information with a preset file type magic number database, and determining a preliminary file type;

[0093] Step S202: Parse the metadata structure of the file, extract key attribute information, and use a pre-trained deep learning model to extract feature vectors of the file content to obtain file content features. The preliminary file type, the key attribute information, and the content feature vector are comprehensively input into a decision tree classifier to obtain the final file type recognition result.

[0094] Optionally, in this embodiment, steps S201 and S202 constitute an efficient and accurate rich media resource file type identification process, combining traditional file structure analysis and modern machine learning technology. This process is intended to solve the technical problem of accurately identifying various complex rich media file types, especially when the file extension may be tampered with or lost.

[0095] In step S201, the system first reads the binary data stream of the file, which is a low-level file access method that can obtain the original information of the file. From this data stream, the system extracts the magic number information and file extension of the file header. The magic number is a specific byte sequence at the beginning of the file, which is usually used to quickly identify the file type. The system matches the extracted magic number information with a preset file type magic number database, which contains standard magic numbers for various file types. Through this matching, the system can preliminarily determine the file type. The advantage of this method is that it is fast and reliable, and can still correctly identify the file type even if the file extension is changed.

[0096] Step S202 further deepens the file type identification process. First, the system parses the metadata structure of the file and extracts key attribute information. These attributes may include information such as file creation time, author, version, etc., which can provide additional clues for file type judgment. Next, the system uses a pre-trained deep learning model to extract feature vectors from the file content. This process involves complex machine learning algorithms, such as convolutional neural networks (CNN) or recurrent neural networks (RNN), which can extract high-dimensional abstract features from file content.

[0097] Finally, the system inputs the preliminary file type, key attribute information and content feature vector into the decision tree classifier. The decision tree classifier is a machine learning model that can process multiple input features. It uses a series of judgment nodes to finally obtain the classification result of the file type. This comprehensive analysis method can make full use of various feature information of the file and greatly improve the accuracy of recognition.

[0098] This two-step file type identification process solves the technical problem of how to accurately identify complex rich media file types. It can effectively handle situations where file extensions may be tampered with or lost, and can also identify new or non-standard file formats. The technical effects of this method are mainly reflected in the following aspects: First, it improves the accuracy of file type identification and reduces the possibility of misjudgment. Second, it enhances the system's processing capabilities for various complex rich media files, laying the foundation for subsequent file processing and management. Furthermore, this method has good scalability and can adapt to newly emerging file types by updating the magic number database and retraining the machine learning model. Finally, this comprehensive analysis method improves the security of the system and can effectively identify disguised or maliciously changed files, which is of great significance for preventing security threats.

[0099] In one embodiment of the rich media intelligent information editing method of the present application, see Figure 3 , and can also include the following:

[0100] Step S301: According to the identified file type, a corresponding content feature extraction algorithm and segmentation strategy are selected to calculate the total size of the file, and the total size is compared with a preset size threshold to determine the initial number of segments;

[0101] Step S302: Analyze the file content through the selected feature extraction algorithm to generate a content feature sequence, and apply the adaptive segmentation algorithm based on dynamic programming to calculate the optimal segmentation point, perform the file segmentation operation at the calculated segmentation point, generate one or more file fragments, and create metadata for each of the file fragments including the original file information, segmentation position and fragment sequence number.

[0102] Optionally, in this embodiment, steps S301 and S302 constitute an intelligent file segmentation process, which combines file type identification, content feature extraction and adaptive segmentation algorithm to solve the problem of efficient processing and storage of large rich media files. This method can not only adapt to the characteristics of different types of files, but also perform intelligent segmentation according to the semantic structure of the file content, thereby improving the efficiency of subsequent processing and storage.

[0103] In step S301, the system first selects the corresponding content feature extraction algorithm and segmentation strategy based on the previously identified file type. This algorithm selection based on file type is very critical because different types of files (such as text, images, audio or video) have different content structures and feature distributions. For example, for text files, the system may choose a feature extraction algorithm based on natural language processing; while for image files, it may choose an algorithm based on computer vision. Next, the system calculates the total size of the file and compares it with a preset size threshold to determine the initial number of segments. The purpose of this step is to provide a preliminary reference framework for subsequent fine segmentation, while also ensuring that the size of the segmented file segments is suitable for storage and transmission.

[0104] Step S302 is the core of the entire segmentation process, which involves complex content analysis and optimization algorithms. First, the system uses the selected feature extraction algorithm to analyze the file content and generate a content feature sequence. This process may involve deep learning models, such as convolutional neural networks or recurrent neural networks, to capture the semantic features and structural information of the file content. Next, the system applies an adaptive segmentation algorithm based on dynamic programming to calculate the optimal segmentation point. The core idea of ​​this algorithm is to find the best segmentation position while maintaining the semantic integrity of the content, so that the segmented file fragments meet the size requirements and maintain the coherence of the content.

[0105] The dynamic programming algorithm considers various possible segmentation schemes through iterative calculations and selects the optimal solution based on predefined objective functions (such as minimizing the degree of content fragmentation and balancing the size of the fragments). This method can effectively process large-scale files and achieve a good balance between computational efficiency and segmentation quality. After determining the optimal segmentation point, the system performs the actual file segmentation operation to generate one or more file fragments.

[0106] Finally, the system creates metadata for each file fragment, including the original file information, segmentation position, and fragment sequence number. These metadata are crucial for subsequent file reorganization and management. They not only record the position and order of each fragment in the original file, but also retain the key information of the original file, ensuring the integrity and traceability of the file.

[0107] This two-step file segmentation process solves the technical problem of how to efficiently process and store large rich media files. Its main technical effects include: First, by intelligently selecting feature extraction algorithms and segmentation strategies, the adaptability of the segmentation process to different file types is improved. Second, adaptive segmentation based on content features ensures the semantic coherence of the segmented file fragments, and maximizes the retention of the structure and meaning of the original file. Furthermore, the application of dynamic programming algorithms improves the efficiency and quality of segmentation, allowing the system to quickly process large-scale files. Finally, detailed metadata creation facilitates subsequent file management, retrieval, and reorganization.

[0108] In one embodiment of the rich media intelligent information editing method of the present application, see Figure 4 , and can also include the following:

[0109] Step S401: selecting a corresponding pre-trained model according to the type of the file segment to perform format conversion, noise reduction and standardization pre-processing, inputting the pre-processed data into the selected pre-trained model, and extracting a high-dimensional feature vector representation;

[0110] Step S402: input the high-dimensional feature vector representation into a pre-trained semantic understanding model, extract key semantic features in the high-dimensional feature vector through the multi-layer self-attention mechanism of the semantic understanding model, cluster the extracted semantic features using a hierarchical clustering algorithm to form semantic concepts at different levels, map the clustering results to a predefined semantic ontology library, obtain standardized semantic labels and sort them according to importance, and construct a structured semantic meaning array, wherein the elements in the semantic meaning array include semantic labels, importance scores and related metrics.

[0111] Optionally, in this embodiment, steps S401 and S402 constitute a complex and powerful file content semantic analysis process, which combines advanced machine learning technology and semantic understanding methods to extract deep semantic information from file fragments and convert it into a structured, easy-to-understand and easy-to-process form. This method can not only process various types of file content, but also capture the semantic hierarchy and importance of the content, providing a strong foundation for subsequent content analysis, retrieval and management.

[0112] In step S401, the system first selects the corresponding pre-trained model according to the type of file fragment to perform a series of pre-processing operations. These operations include format conversion, noise reduction and standardization. Format conversion ensures that different types of file content can be processed uniformly; the noise reduction process removes noise information that may affect semantic extraction; standardization makes data from different sources comparable. The technical principles of these pre-processing steps involve technologies in multiple fields such as signal processing, image processing and natural language processing. The pre-processed data is input into the selected pre-trained model, which may be a deep neural network, such as a convolutional neural network (CNN) or a transformer model. These models can extract high-dimensional feature vector representations from the input data, which contain rich semantic information about the file content.

[0113] Step S402 is the core of the whole process, which involves complex semantic understanding and knowledge organization processes. First, the system inputs the high-dimensional feature vector into a pre-trained semantic understanding model. This model may be based on advanced natural language processing models such as BERT or GPT. Through the model's multi-layer self-attention mechanism, the system is able to capture the key semantic features in the feature vector. The self-attention mechanism allows the model to focus on different parts of the input and understand the relationship between them, which is crucial for understanding complex semantic structures.

[0114] Next, the system uses a hierarchical clustering algorithm to cluster the extracted semantic features. Hierarchical clustering can discover the hierarchical relationship between features and form semantic concepts at different levels. The advantage of this method is that it can capture the multi-level structure of semantics, thereby more comprehensively understanding the meaning of the content. The clustering results are then mapped to a predefined semantic ontology library. The semantic ontology library is a knowledge base that contains standardized semantic concepts and the relationships between them. Through this mapping, the system is able to convert the extracted semantic features into standardized semantic labels.

[0115] Finally, the system sorts the semantic tags according to their importance and builds a structured array of semantic meanings. Each element in this array contains a semantic tag, an importance score, and a related metric. This structured representation not only retains the key semantic information of the original content, but also provides an evaluation of the importance and relevance of this information, providing rich semantic context for subsequent applications.

[0116] This two-step semantic analysis process solves the technical problem of how to extract deep semantic information from complex rich media files and transform it into a structured and quantifiable form. Its main technical effects include: First, through the application of preprocessing and pre-training models, the processing capabilities of various types of file content and the accuracy of feature extraction are improved. Secondly, the use of multi-layer self-attention mechanism enables the system to understand complex semantic relationships and contextual information. Furthermore, hierarchical clustering and semantic ontology mapping realize the standardization and structuring of semantic information, which is convenient for subsequent retrieval and analysis. Finally, the structured semantic meaning array provides strong support for intelligent understanding and management of content.

[0117] In one embodiment of the rich media intelligent information editing method of the present application, see Figure 5 , and can also include the following:

[0118] Step S501: Map each semantic tag in the semantic meaning array to a predefined feature space, assign a weight to each dimension according to the importance score of the semantic tag, and generate a weighted feature vector as an input for similarity calculation;

[0119] Step S502: Calculate the similarity of the three dimensions of content, structure and semantics respectively, determine the weight of each dimension according to the file type, and combine the weighted similarity results of the three dimensions to obtain a final similarity score.

[0120] Optionally, in this embodiment, steps S501 and S502 constitute a comprehensive and sophisticated file similarity calculation process, which integrates semantic understanding, feature mapping and multi-dimensional similarity analysis technology, aiming to solve the technical difficulties of similarity evaluation of complex rich media files. This method not only considers the surface features of the file, but also deeply analyzes its inherent semantic structure, thereby providing a more accurate and meaningful similarity measurement.

[0121] In step S501, the system first maps each semantic tag in the semantic meaning array to a predefined feature space. This feature space is a multidimensional vector space, in which each dimension represents a specific semantic concept or attribute. The mapping process actually converts abstract semantic concepts into quantifiable numerical features. The technical principle of this mapping is based on the theory of distributed semantic representation, such as models such as word2vec or GloVe. These models are able to encode semantic concepts into dense vector representations, preserving the semantic relationships between concepts.

[0122] Next, the system assigns weights to the corresponding dimensions in the feature space based on the importance score of each semantic tag. The purpose of this step is to highlight important semantic features while reducing the influence of less relevant or less important features. The process of weight assignment may involve normalization to ensure that the weight distribution between different files is comparable. Through this weighting process, the system generates a weighted feature vector, which not only contains the original semantic information, but also reflects the relative importance of each semantic concept.

[0123] Step S502 is the core of similarity calculation, which adopts a multi-dimensional analysis method. The system calculates the similarity of content, structure and semantics respectively. Content similarity may involve text matching, image feature comparison and other technologies; structural similarity may consider the organization of the file, format features, etc.; and semantic similarity is based on the weighted feature vector generated previously, using cosine similarity or other vector similarity measurement methods for calculation.

[0124] When calculating the similarity of these three dimensions, the system determines the weight of each dimension based on the file type. This is because the importance of these three dimensions may be different for different types of files. For example, for text files, semantic similarity may be more important; while for image files, content similarity may have a greater weight. Through this dynamic weight allocation, the system can flexibly adapt to the characteristics of different types of files and provide more accurate similarity evaluation.

[0125] Finally, the system combines the weighted similarity results of the three dimensions to obtain the final similarity score. This combination process may involve simple linear combinations or more complex nonlinear models such as support vector machines (SVMs) or neural networks to capture the complex interactions between dimensions.

[0126] This two-step similarity calculation process solves the technical problem of how to comprehensively and accurately evaluate the similarity of complex rich media files. Its main technical effects include: First, by mapping semantic tags to feature space and weighting them, the quantification and importance distinction of semantic information are achieved. Second, the multi-dimensional similarity calculation takes into account the content, structure and semantic features of the file, providing a more comprehensive similarity evaluation. Furthermore, the method of dynamically adjusting the dimension weights according to the file type improves the flexibility and adaptability of the similarity calculation. Finally, by combining the similarity results of multiple dimensions, a comprehensive and accurate final similarity score is obtained.

[0127] In one embodiment of the rich media intelligent information editing method of the present application, see Figure 6 , and can also include the following:

[0128] Step S601: parsing the content of the file editing instruction sent by the user, identifying the instruction type and operation parameters, and converting the parsed instruction into an operation sequence executable by the system;

[0129] Step S602: Load the contents of the current rich media resource file and similar rich media resource files selected by the user, select the corresponding content integration algorithm according to the file type to perform the content integration operation, generate a new integrated rich media resource file, and retain the version information of the original file, wherein the content integration operation includes content merging, replacement, deletion or insertion.

[0130] Optionally, in this embodiment, steps S601 and S602 constitute an intelligent rich media resource file editing and integration process, which combines natural language processing, command parsing and content integration technology to solve the problem of efficient editing and intelligent integration of complex rich media files. This method can not only understand and execute the user's editing intention, but also select the most suitable integration strategy according to the file type, thereby improving editing efficiency and integration quality.

[0131] In step S601, the system first parses the content of the file editing instruction sent by the user. This process involves natural language processing and command parsing technology. The system needs to identify the type of instruction (such as merge, replace, delete or insert) and related operation parameters (such as target location, content range, etc.). This parsing process may use rule-based methods or more advanced machine learning models, such as sequence labeling models or intent recognition models. These models can extract structured operation information from natural language instructions.

[0132] After parsing, the system converts these instructions into a sequence of operations that the system can execute. This conversion process involves command mapping and operation optimization. The system needs to convert high-level editing instructions into low-level file operations, while considering the order and dependencies of operations to ensure the correctness and efficiency of execution. The technical principle of this step is based on the compilation principle and the command interpreter design theory in the operating system.

[0133] Step S602 is the core of the whole process, which involves a complex file content integration process. First, the system loads the content of the current rich media resource file and similar rich media resource files selected by the user. This loading process needs to consider the characteristics of different types of rich media files, and may involve technologies such as file decoding and content parsing.

[0134] Next, the system selects the corresponding content integration algorithm based on the file type. This algorithm selection based on file type is very critical because different types of files (such as text, images, audio or video) require different integration strategies. For example, for text files, the system may use a paragraph merging algorithm based on semantic understanding; for image files, it may use image fusion technology based on computer vision; for audio and video files, it may involve complex multimedia editing algorithms.

[0135] The system performs content integration operations, including content merging, replacement, deletion, or insertion. The specific implementation of these operations depends on the selected integration algorithm. For example, content merging may involve the intelligent splicing of text paragraphs, the seamless fusion of images, or the smooth transition of audio and video clips. Replacement operations may need to consider contextual consistency to ensure that the replaced content is coordinated with the surrounding environment. Deletion operations may require reorganization of the remaining content to maintain the coherence of the file. Insertion operations need to consider how to naturally integrate new content into the existing structure.

[0136] After the integration operation is completed, the system generates a new integrated rich media resource file. At the same time, the system retains the version information of the original file, which is very important for file management, version control, and possible rollback operations. The preservation of version information may use incremental storage or snapshot technology to save storage space and support efficient version comparison and recovery.

[0137] This two-step file editing and integration process solves the technical problem of how to efficiently process and integrate complex rich media files. Its main technical effects include: First, by intelligently parsing user instructions, the accuracy and efficiency of editing operations are improved. Second, the integration algorithm selection based on file type ensures the adaptability of the integration process to different types of files. Third, complex content integration operations achieve high-quality file merging and editing. Finally, the retention of version information provides support for file management and version control.

[0138] In one embodiment of the rich media intelligent information editing method of the present application, see Figure 7 , and can also include the following:

[0139] Step S701: Compare the content structures of the current rich media resource file and similar rich media resource files to identify repeated, contradictory or inconsistent content segments, and mark the conflicting positions and types;

[0140] Step S702: Process some conflicts according to the conflict type and the preset resolution strategy. For conflicts that cannot be automatically resolved, generate corresponding user options, receive and execute the user's conflict resolution selection, update the integrated rich media resource file content, and eliminate the conflict.

[0141] Optionally, in this embodiment, steps S701 and S702 constitute an intelligent rich media resource file conflict detection and resolution process, which combines content comparison, conflict identification and intelligent decision-making technology to solve content conflict problems in the process of complex rich media file integration. This method can not only accurately identify various types of conflicts, but also flexibly handle these conflicts according to preset strategies and user intentions, thereby improving the quality and efficiency of file integration.

[0142] In step S701, the system first compares the content structure of the current rich media resource file and similar rich media resource files. This comparison process involves complex content analysis and structure matching technology. For text content, the system may use text similarity algorithms, such as cosine similarity or Jaccard similarity; for image content, it may use image feature extraction and matching technology; for audio and video content, it may involve time series analysis and content fingerprint technology. The comprehensive application of these technologies enables the system to identify repeated, contradictory or inconsistent parts in different types of rich media content.

[0143] During the recognition process, the system needs to consider different levels of content units, from word paragraphs to image areas, and then to audio and video clips. The system identifies potential conflicts by setting thresholds or using machine learning models to determine the relationship between content units. The technical principle of this process is based on information retrieval and pattern recognition theory, which can effectively filter out the parts that need special attention from massive content.

[0144] Once a conflict is identified, the system will mark the location and type of the conflict. The conflict type may include content duplication, information contradiction, inconsistent format, etc. This mark includes not only the location information of the conflict, but also metadata such as the severity of the conflict and the scope of impact, providing important reference for subsequent conflict handling.

[0145] Step S702 is the core of conflict resolution, which adopts an intelligent processing method. First, the system handles some conflicts according to the conflict type and the preset resolution strategy. These preset strategies may be based on statistical laws, domain knowledge or previous user selections. For example, for simple duplicate content, the system may automatically choose to keep one of the versions; for minor format inconsistencies, the system may automatically unify the format.

[0146] However, for complex or high-risk conflicts, the system cannot make decisions automatically. In this case, the system generates corresponding user options. These options need to be presented to the user in a clear and understandable way, which may include a comparative display of the conflict content, resolution suggestions, etc. The system receives the user's selection and performs the corresponding conflict handling operation. This process embodies the idea of ​​human-computer collaboration, combining the system's efficient processing capabilities with human judgment.

[0147] After conflict resolution, the system updates the integrated rich media resource file content to eliminate conflicts. This update process needs to consider the coherence and consistency of the content, which may involve operations such as content re-arrangement and format adjustment. The system also needs to ensure that these changes do not introduce new conflicts or problems.

[0148] This two-step conflict detection and resolution process solves the problem of how to efficiently handle content conflicts in the process of integrating complex rich media files. Its main technical effects include: First, through intelligent comparison and conflict identification, the accuracy and comprehensiveness of conflict detection are improved. Second, the application of preset strategies realizes the automatic processing of some conflicts and improves efficiency. Third, the human-machine collaborative conflict resolution method ensures the accuracy of key decisions. Finally, intelligent conflict handling and content updates ensure the quality and consistency of the final file.

[0149] This method not only improves the quality and efficiency of rich media resource file integration, but also provides strong support for applications such as intelligent content management, version control, and collaborative editing. It has broad application prospects in document management systems, content creation platforms, knowledge base integration, and other fields, and can significantly improve the intelligence level of content processing and management, providing users with more efficient and accurate content integration services.

[0150] In order to process rich media information more flexibly, the present application provides an embodiment of a rich media intelligent information editing and reviewing device for implementing all or part of the contents of the rich media intelligent information editing and reviewing method, see Figure 8 The rich media intelligent information editing and reviewing device specifically includes the following contents:

[0151] The file segmentation module 10 is used to receive and store the rich media resource files uploaded by the user, perform file identification on the rich media resource files, and perform file segmentation on the rich media resource files by an adaptive segmentation algorithm according to the identified file types, file content characteristics and a preset size threshold to obtain one or more file segments;

[0152] A similarity display module 20 is used to perform multimodal content vectorization on the file fragment through a preset rich media vectorization model, extract a corresponding semantic meaning array, determine the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and display information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type;

[0153] The similarity merging module 30 is used to receive the file editing instruction sent by the user, integrate the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handle the conflict of the content of similar files through a conflict resolution mechanism, and display the file content after the content integration and conflict handling in the preset user file editing interface.

[0154] From the above description, it can be seen that the rich media intelligent information editing device provided in the embodiment of the present application can receive and store rich media resource files uploaded by users, perform file identification on the rich media resource files, and perform file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file type, file content characteristics and preset size threshold to obtain one or more file fragments; perform multimodal content vectorization on the file fragments through a preset rich media vectorization model, extract the corresponding semantic meaning array, determine the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm based on the semantic meaning array, and display the information whose similarity with the current rich media resource file exceeds the preset threshold in the preset user file editing interface, thereby processing rich media information more flexibly.

[0155] From the hardware level, in order to be able to process rich media information more flexibly, the present application provides an embodiment of an electronic device for implementing all or part of the content in the rich media intelligent information editing method, and the electronic device specifically includes the following content:

[0156] Processor, memory, communication interface and bus; wherein the processor, memory and communication interface communicate with each other through the bus; the communication interface is used to realize information transmission between the rich media intelligent information editing and reviewing device and related devices such as the core business system, user terminal and related database; the logic controller can be a desktop computer, a tablet computer and a mobile terminal, etc., but the present embodiment is not limited thereto. In the present embodiment, the logic controller can be implemented with reference to the embodiment of the rich media intelligent information editing and reviewing method and the embodiment of the rich media intelligent information editing and reviewing device in the embodiment, and the contents thereof are incorporated herein, and the repeated parts are not repeated.

[0157] It is understandable that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0158] In practical applications, part of the rich media intelligent information editing method can be executed on the electronic device side as described above, or all operations can be completed in the client device. The specific selection can be based on the processing capability of the client device and the limitations of the user's usage scenario. This application does not limit this. If all operations are completed in the client device, the client device may also include a processor.

[0159] The client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side, and other implementation scenarios may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, or a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0160] Fig. 9 FIG. 9 is a schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Fig. 9 As shown, the electronic device 9600 may include a central processor 9100 and a memory 9140; the memory 9140 is coupled to the central processor 9100. It is worth noting that Fig. 9 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0161] In one embodiment, the function of the rich media intelligent information editing method can be integrated into the central processing unit 9100. The central processing unit 9100 can be configured to perform the following control:

[0162] Step S101: receiving and storing a rich media resource file uploaded by a user, performing file identification on the rich media resource file, and performing file segmentation on the rich media resource file by an adaptive segmentation algorithm according to the identified file type, file content characteristics, and a preset size threshold, to obtain one or more file segments;

[0163] Step S102: multimodal content vectorization is performed on the file fragment through a preset rich media vectorization model to extract a corresponding semantic meaning array, and similarity between other rich media resource files and the current rich media resource file is determined through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold is displayed in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type;

[0164] Step S103: receiving the file editing instruction sent by the user, integrating the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handling the conflict of the similar file contents through a conflict resolution mechanism, and displaying the file content after the content integration and conflict handling in the preset user file editing interface.

[0165] From the above description, it can be seen that the electronic device provided in the embodiment of the present application receives and stores rich media resource files uploaded by users, performs file identification on the rich media resource files, and performs file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file type, file content characteristics and preset size threshold, so as to obtain one or more file fragments; multimodal content vectorization is performed on the file fragments through a preset rich media vectorization model, and the corresponding semantic meaning array is extracted, and the similarity between other rich media resource files and the current rich media resource file is determined through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and information whose similarity with the current rich media resource file exceeds the preset threshold is displayed in a preset user file editing interface, thereby processing rich media information more flexibly.

[0166] In another embodiment, the rich media intelligent information editing device can be configured separately from the central processing unit 9100. For example, the rich media intelligent information editing device can be configured as a chip connected to the central processing unit 9100, and the functions of the rich media intelligent information editing method are implemented through the control of the central processing unit.

[0167] like Fig. 9 As shown, the electronic device 9600 may also include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily have to include Fig. 9 In addition, the electronic device 9600 may also include Fig. 9 For components not shown, reference may be made to the prior art.

[0168] like Fig. 9 As shown, the central processing unit 9100 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0169] The memory 9140 may be, for example, one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The above-mentioned information related to the failure may be stored, and a program for executing the relevant information may also be stored. The CPU 9100 may execute the program stored in the memory 9140 to implement information storage or processing, etc.

[0170] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0171] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 9140 may also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142, which is used to store application programs and function programs or processes for executing the operation of the electronic device 9600 through the central processor 9100.

[0172] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0173] The communication module 9110 is a transmitter / receiver that sends and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processor 9100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0174] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless LAN module, etc. The communication module 9110 (transmitter / receiver) is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby realizing a common telecommunication function. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local machine through the microphone 9132, and the sound stored on the local machine can be played through the speaker 9131.

[0175] The embodiments of the present application also provide a computer-readable storage medium capable of implementing all the steps of the rich media intelligent information editing and review method in the above embodiment, where the execution subject is a server or a client. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, all the steps of the rich media intelligent information editing and review method in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented:

[0176] Step S101: receiving and storing a rich media resource file uploaded by a user, performing file identification on the rich media resource file, and performing file segmentation on the rich media resource file by an adaptive segmentation algorithm according to the identified file type, file content characteristics, and a preset size threshold, to obtain one or more file segments;

[0177] Step S102: multimodal content vectorization is performed on the file fragment through a preset rich media vectorization model to extract a corresponding semantic meaning array, and similarity between other rich media resource files and the current rich media resource file is determined through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold is displayed in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type;

[0178] Step S103: receiving the file editing instruction sent by the user, integrating the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handling the conflict of the similar file contents through a conflict resolution mechanism, and displaying the file content after the content integration and conflict handling in the preset user file editing interface.

[0179] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application receives and stores rich media resource files uploaded by users, performs file identification on the rich media resource files, and performs file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file type, file content characteristics and preset size threshold, so as to obtain one or more file fragments; performs multimodal content vectorization on the file fragments through a preset rich media vectorization model, extracts the corresponding semantic meaning array, determines the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and displays information whose similarity with the current rich media resource file exceeds the preset threshold in a preset user file editing interface, thereby processing rich media information more flexibly.

[0180] The embodiments of the present application also provide a computer program product capable of implementing all the steps of the rich media intelligent information editing and review method in the above embodiments, where the execution subject is a server or a client. When the computer program / instruction is executed by a processor, the steps of the rich media intelligent information editing and review method are implemented. For example, the computer program / instruction implements the following steps:

[0181] Step S101: receiving and storing a rich media resource file uploaded by a user, performing file identification on the rich media resource file, and performing file segmentation on the rich media resource file by an adaptive segmentation algorithm according to the identified file type, file content characteristics, and a preset size threshold, to obtain one or more file segments;

[0182] Step S102: multimodal content vectorization is performed on the file fragment through a preset rich media vectorization model to extract a corresponding semantic meaning array, and similarity between other rich media resource files and the current rich media resource file is determined through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold is displayed in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type;

[0183] Step S103: receiving the file editing instruction sent by the user, integrating the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handling the conflict of the similar file contents through a conflict resolution mechanism, and displaying the file content after the content integration and conflict handling in the preset user file editing interface.

[0184] From the above description, it can be seen that the computer program product provided by the embodiment of the present application receives and stores rich media resource files uploaded by users, performs file identification on the rich media resource files, and performs file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file type, file content characteristics and preset size threshold, so as to obtain one or more file fragments; performs multimodal content vectorization on the file fragments through a preset rich media vectorization model, extracts the corresponding semantic meaning array, determines the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and displays information whose similarity with the current rich media resource file exceeds the preset threshold in a preset user file editing interface, thereby processing rich media information more flexibly.

[0185] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0186] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0187] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0188] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0189] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, 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 the present invention.

Claims

1. A rich media intelligent information editing method, characterized in that: The method comprises: Receive and store rich media resource files uploaded by users, perform file identification on the rich media resource files, and perform file segmentation on the rich media resource files using an adaptive segmentation algorithm according to the identified file types, file content features, and a preset size threshold to obtain one or more file segments; The file fragment is multimodally vectorized by a preset rich media vectorization model, and a corresponding semantic meaning array is extracted; the similarity between other rich media resource files and the current rich media resource file is determined by a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold is displayed in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type; Receive the file editing instruction sent by the user, integrate the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handle the conflict of the similar file contents through a conflict resolution mechanism, and display the file content after the content integration and conflict handling in the preset user file editing interface.

2. The method for editing and reviewing rich media intelligent information according to claim 1, characterized in that: The performing file identification on the rich media resource file includes: Read the binary data stream of the rich media resource file, extract the magic number information and the file extension in the file header, match the extracted magic number information with a preset file type magic number database, and determine a preliminary file type; Parse the metadata structure of the file, extract key attribute information, and use the pre-trained deep learning model to extract feature vectors of the file content to obtain file content features. The preliminary file type, the key attribute information and the content feature vector are comprehensively input into the decision tree classifier to obtain the final file type recognition result.

3. The rich media intelligent information editing method according to claim 1, characterized in that: The process of segmenting the rich media resource file by an adaptive segmentation algorithm according to the identified file type, file content features and a preset size threshold to obtain one or more file segments includes: According to the identified file type, the corresponding content feature extraction algorithm and segmentation strategy are selected to calculate the total size of the file, and compared with the preset size threshold to determine the initial number of segments; The file content is analyzed by a selected feature extraction algorithm to generate a content feature sequence, and an adaptive segmentation algorithm based on dynamic programming is applied to calculate the optimal segmentation point. The file segmentation operation is performed at the calculated segmentation point to generate one or more file fragments, and metadata containing the original file information, segmentation position and fragment sequence number is created for each of the file fragments.

4. The rich media intelligent information editing method according to claim 1, characterized in that: The multimodal content vectorization of the file segment is performed by using a preset rich media vectorization model to extract a corresponding semantic meaning array, including: Selecting a corresponding pre-trained model according to the type of the file segment to perform format conversion, noise reduction and standardization pre-processing, inputting the pre-processed data into the selected pre-trained model, and extracting a high-dimensional feature vector representation; The high-dimensional feature vector representation is input into a pre-trained semantic understanding model, the key semantic features in the high-dimensional feature vector are extracted through the multi-layer self-attention mechanism of the semantic understanding model, the extracted semantic features are clustered using a hierarchical clustering algorithm to form semantic concepts at different levels, the clustering results are mapped to a predefined semantic ontology library, normalized semantic labels are obtained and sorted according to importance, and a structured semantic meaning array is constructed, wherein the elements in the semantic meaning array include semantic labels, importance scores and related metrics.

5. The rich media intelligent information editing method according to claim 1, characterized in that: Determining the similarity between other rich media resource files and the current rich media resource file by using a multi-dimensional weighted similarity algorithm according to the semantic meaning array includes: Map each semantic label in the semantic meaning array to a predefined feature space, assign a weight to each dimension according to the importance score of the semantic label, and generate a weighted feature vector as the input for similarity calculation; The similarities of the three dimensions of content, structure and semantics are calculated separately, the weight of each dimension is determined according to the file type, and the weighted similarity results of the three dimensions are combined to obtain the final similarity score.

6. The rich media intelligent information editing method according to claim 1, characterized in that: The receiving the file editing instruction sent by the user, and integrating the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, includes: Parse the contents of the file editing instructions sent by the user, identify the instruction type and operation parameters, and convert the parsed instructions into an operation sequence that can be executed by the system; Load the contents of the current rich media resource file and similar rich media resource files selected by the user, select the corresponding content integration algorithm according to the file type to perform the content integration operation, generate a new integrated rich media resource file, and retain the version information of the original file, wherein the content integration operation includes content merging, replacing, deleting or inserting.

7. The rich media intelligent information editing method according to claim 1, characterized in that: The conflict resolution mechanism is used to handle the conflict of similar file contents, including: Compare the content structures of the current rich media resource file and similar rich media resource files to identify repeated, contradictory or inconsistent content segments, and mark conflict locations and types; Handle some conflicts according to the conflict type and preset resolution strategy. For conflicts that cannot be resolved automatically, generate corresponding user options, receive and execute the user's conflict resolution selection, update the integrated rich media resource file content, and eliminate the conflict.

8. A rich media intelligent information editing and reviewing device, characterized in that: The device comprises: A file segmentation module is used to receive and store rich media resource files uploaded by users, perform file identification on the rich media resource files, and perform file segmentation on the rich media resource files through an adaptive segmentation algorithm according to the identified file types, file content characteristics, and a preset size threshold to obtain one or more file segments; A similarity display module, used to perform multimodal content vectorization on the file fragment through a preset rich media vectorization model, extract a corresponding semantic meaning array, determine the similarity between other rich media resource files and the current rich media resource file through a multi-dimensional weighted similarity algorithm according to the semantic meaning array, and display information of similar rich media resource files whose similarity with the current rich media resource file exceeds a preset threshold in a preset user file editing interface, wherein the multi-dimensional weighted similarity algorithm includes content similarity calculation, structural similarity calculation and semantic similarity calculation, and the similarity calculation weight of each dimension is determined by the file type; The similarity merging module is used to receive the file editing instruction sent by the user, integrate the content of the corresponding current rich media resource file and the similar rich media resource file selected by the user according to the file editing instruction, and handle the conflict of the content of similar files through the conflict resolution mechanism, and display the file content after the content integration and conflict handling in the preset user file editing interface.

9. An electronic 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 program, the steps of the rich media intelligent information editing method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the rich media intelligent information editing method described in any one of claims 1 to 7 are implemented.

Images