Multi-modal data management system for intelligent media resources

Abstract

The invention discloses a multi-modal data management system for intelligent media resources, which relates to the field of data management, and comprises a media resource management center which comprises a data acquisition module, a data processing module, a data distribution module, a data retrieval module, a data analysis module and a data management module; the data acquisition module acquires multi-modal data information; the data processing module performs classification processing on the multi-modal data information; the data distribution mode is used for constructing a corresponding multi-mode data distribution storage library according to the classification processing result; the data retrieval module obtains personalized retrieval data and a personalized retrieval database according to user requirements; the data analysis module is used for acquiring user demand feedback information; the data management module performs layout processing on the multi-modal data distribution storage library again according to the corresponding data information, and updates the corresponding multi-modal data distribution storage library; according to the method, the flexibility and the management efficiency in the multi-modal data management process are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data management, and in particular to a multi-modal data management system for intelligent media resources. BACKGROUND

[0002] With the development of technologies such as the Internet and mobile devices, the amount of data corresponding to multi-modal data such as text, images, audio, and video is growing exponentially, and multi-modal data is widely used, and businesses are becoming increasingly diversified. However, traditional media data management has many limitations, limited processing capacity, difficulty in integrating multi-modal data, insufficient data security and reliability, and still cannot support business innovation. At the same time, with the continuous progress of technologies such as artificial intelligence, big data, cloud computing, and 5G, powerful support is provided for the understanding, storage, processing, and transmission of multi-modal data. Under this background, a multi-modal data management system for intelligent media resources has emerged.

[0003] After searching, the invention patent with Chinese patent number CN118503148B discloses a multi-modal data management system, method, electronic device, and storage medium. The multi-modal data management system includes a first cache module, a synchronization alignment module, a second cache module, a flow storage module, and a third cache module. The multi-modal data management system buffers the multi-modal data collected, the synchronized multi-modal data, and the converted data after format conversion based on the first cache module, the second cache module, and the third cache module to achieve three-level data buffering. The synchronized multi-modal data about flow cytometry is integrated, sorted, and stored in the storage hard disk based on the synchronization alignment module and the flow storage module, effectively matching the data flow rate gap in each link of data processing, and realizing real-time acquisition, synchronization alignment, and orderly storage of high-throughput multi-modal data flow.

[0004] Compared with the prior art, the invention patent with Chinese patent number CN118503148B can improve the management efficiency of multi-modal data by setting multiple cache modules to collect, synchronize, align, format convert, store, and schedule the corresponding multi-modal data. However, the above system in the actual use process, through the unified multi-modal data processing mode to the corresponding multi-modal data for acquisition, synchronization alignment and storage, in this process, without considering the flexibility of user demand, the obtained storage result cannot be updated and adjusted according to user demand, thereby reducing the flexibility and user use efficiency in the process of multi-modal data management. SUMMARY

[0005] The present application relates to the technical field of data management, and in particular to a multi-modal data management system for intelligent media resources.

[0006] In order to achieve the above object, the present application adopts the following technical solutions: A multi-modal data management system of intelligent media resources, comprising a media resource management center, wherein the media resource management center comprises a data acquisition module, a data processing module, a data distribution module, a data retrieval module, a data analysis module and a data management module; The data acquisition module is configured to acquire corresponding intelligent media resource data, and perform modal recognition on the intelligent media resource data to obtain multi-modal data information of corresponding types; The data processing module is configured to perform mapping processing on the obtained multi-modal data information respectively, perform classification processing according to the mapping processing results, and obtain corresponding resource distribution basis element information; The data distribution module is configured to analyze and process the corresponding resource distribution basis element information, set corresponding modal basic information and modal identification information, perform distribution processing on the corresponding multi-modal data information based on the blockchain technology, and construct a multi-modal data distribution storage library; The data retrieval module is configured to obtain corresponding personalized retrieval data according to user demand, and obtain a corresponding personalized retrieval database in the multi-modal data distribution storage library according to the personalized retrieval data; The data analysis module is configured to integrate demand according to the corresponding personalized retrieval database, generate personalized demand output information, feed back the personalized demand output information to a corresponding user account, and perform feedback processing by the user account to obtain user demand feedback information; The data management module performs redistribution processing on the corresponding multi-modal data distribution storage library according to each personalized retrieval database, personalized demand output information and user demand feedback information, and obtains a multi-modal data storage library.

[0007] The above technical solution further comprises the following steps: setting a data acquisition unit and a data recognition unit; acquiring intelligent media resource information corresponding to the operation results of a corresponding user according to user demand and a user account in a corresponding intelligent media device terminal through the data acquisition unit; identifying and analyzing the acquired intelligent media resource information through the data recognition unit, setting a data recognition analysis node, inputting the intelligent media resource information into the data recognition analysis node, sequentially traversing the recognition analysis nodes of corresponding modal types, obtaining corresponding modal recognition information, and obtaining corresponding multi-modal data information according to the modal recognition information.

[0008] Further, the process of obtaining resource distribution basis element information comprises: setting a data processing unit and a data classification unit; The data processing unit is configured to obtain corresponding multi-modal data information, respectively perform mapping processing on the multi-modal data information, respectively perform analysis processing on the corresponding multi-modal data information, obtain corresponding semantic representation information, perform connection processing on the obtained semantic representation information, obtain a semantic mapping vector corresponding to the multi-modal data information, and perform marking processing on the semantic mapping vector. The data classification unit is configured to obtain a semantic mapping vector corresponding to corresponding multi-modal data information, respectively perform feature extraction on the corresponding semantic mapping vector according to mapping connection nodes, and obtain semantic feature information corresponding to the semantic mapping vector. The semantic feature information corresponding to the corresponding multi-modal data information in the media resource management center is set to a classification evaluation interval data, the semantic feature information corresponding to the corresponding semantic mapping vector is respectively compared and analyzed with the corresponding classification evaluation interval data, interval evaluation data is obtained, and the interval evaluation data corresponding to the multi-modal data information is respectively set to corresponding resource distribution basis element information.

[0009] Further, the process of setting the modal basic information and the modal identification information includes: A data identification unit and a data storage unit are set. The data identification unit is configured to set a basic storage space and an identification storage space, obtain multi-modal data information, store the obtained multi-modal data information in the basic storage space, set storage coding for the corresponding multi-modal data information according to the storage result, and set the corresponding modal basic information according to the corresponding storage coding and the corresponding multi-modal data information. The resource distribution basis element information corresponding to the corresponding multi-modal data information is obtained, the basic storage space is obtained according to the resource distribution basis element information, the identification coding is set in the basic storage space according to the corresponding modal basic information, the resource distribution basis element information, and the corresponding semantic mapping vector, the modal identification information is set according to the identification coding, and the modal identification information is stored in the identification storage space.

[0010] Further, the process of constructing a multi-modal data distribution storage library includes: The data storage unit is configured to obtain the modal basic information and the modal identification information stored in the basic storage space and the identification storage space, set the basic storage space as a block center node, and set each identification storage space as a corresponding block sub-node. The spatial layout processing is performed on the corresponding modal identification information in each block sub-node according to the correlation data between the corresponding semantic mapping vectors, and the position information of the block sub-node to which the corresponding modal identification information belongs is obtained. According to the position information of the corresponding block sub-node to which the corresponding modal identification information belongs, the corresponding associated storage information is set in the corresponding modal basic information in the corresponding block center node, the associated storage information corresponding to the corresponding modal basic information in the block center node and the block sub-node and the corresponding position information are set in a chain storage structure, and the corresponding multi-modal data distribution storage library is constructed according to the storage results in the block center node and the block sub-node and the corresponding chain storage structure.

[0011] Further, the process of obtaining the personalized retrieval database includes: setting a personalized collection unit and a personalized retrieval unit; The personalized collection unit is connected with the corresponding user account, and the personalized retrieval data corresponding to the user demand is obtained according to the corresponding user account, and the personalized retrieval data includes retrieval keyword information, user preference information and retrieval expansion information; The personalized retrieval unit obtains the corresponding personalized retrieval data and the multi-modal data distribution storage library, respectively processes the obtained personalized retrieval data, obtains the corresponding personalized semantic representation information, and sets a retrieval mapping vector according to the personalized semantic representation information; The obtained retrieval mapping vector is input into the multi-modal data distribution storage library, the associated modal basic information is obtained, the modal identification information corresponding to the obtained associated storage information is obtained according to the obtained modal basic information, the corresponding retrieval modal information is obtained according to the corresponding distribution in the modal identification information, the distribution of the corresponding multi-modal data distribution storage library to which the obtained retrieval modal information belongs is extracted in turn, and the corresponding personalized retrieval database is constructed.

[0012] Further, the process of obtaining the user demand feedback information includes: setting a demand analysis unit and a demand feedback unit; The demand analysis unit is used to obtain the corresponding personalized retrieval database, and the corresponding multi-modal data information in the personalized retrieval database is fused and processed; The historical personalized retrieval database is analyzed and processed based on a deep learning algorithm, a multi-modal fusion model is constructed, the corresponding personalized retrieval database is input into the multi-modal fusion model for analysis and processing, the corresponding personalized demand output information is output, and the corresponding user account is sent to the corresponding user account; The demand feedback unit is used for the user to obtain the personalized demand output information through the corresponding user account, to perform multi-angle feedback evaluation on the obtained personalized demand output information, to quantitatively process the multi-angle feedback evaluation in sequence, and to generate the user demand feedback information according to the quantitatively processed results.

[0013] Further, the process of redistributing the multi-modal data distribution repository includes: setting a statistical analysis unit and a data management unit; The statistical analysis unit is used to obtain personalized search databases, personalized demand output information and user demand feedback information corresponding to the corresponding user account, and to statistically analyze the data information obtained by the corresponding user account to obtain distribution statistical data corresponding to the corresponding multi-modal data information. The data management unit obtains distribution statistical data corresponding to the corresponding multi-modal data information, and redistributes the current multi-modal data distribution repository according to the corresponding distribution statistical data, and re-arranges the corresponding multi-modal data information to obtain an updated multi-modal data distribution repository and update the mark.

[0014] The present application has the following advantages: 0、In the present application, the personalized search data corresponding to the user demand is obtained, the multi-modal data information in the multi-modal data distribution repository is mapped according to the personalized search data, the personalized search database is obtained, and the data analysis and processing is carried out according to the user demand through the personalized search database, thereby improving the user flexibility demand in the multi-modal data acquisition process, and improving the efficiency in the data processing process to a certain extent.

[0015] 1、In the present application, the personalized search data corresponding to the user demand is analyzed and processed through the personalized search database, the personalized demand output information and the corresponding user demand feedback information are obtained, and the statistical analysis is carried out on the corresponding data information, so as to re-arrange the corresponding multi-modal data distribution repository, thereby realizing the update management of the multi-modal data distribution repository according to the different flexibility of the user demand, improving the management efficiency of the multi-modal data analysis and management, and improving the flexibility of the multi-modal data update management process to a certain extent.

[0016] 2、In the present application, the multi-modal data information is connected by semantic representation information, the semantic mapping vector is set according to the connection processing result, the multi-modal data is extracted according to the semantic mapping vector, the resource distribution basis element information is set according to the feature extraction result, the multi-modal distribution repository is set according to the corresponding resource distribution basis element information, and the multi-modal distribution repository is stored according to the modal basic information and the modal identification information, thereby improving the correlation in the multi-modal data information storage management process to a certain extent, and saving the data resources to a certain extent. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1This is a schematic diagram of the structure of a multimodal data management system for intelligent media resources proposed in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1 like Figure 1 As shown, the present invention proposes a multimodal data management system for intelligent media resources, including a media resource management center, which includes a data acquisition module, a data processing module, a data distribution module, a data retrieval module, a data analysis module, and a data management module.

[0020] In this embodiment, the media resource management center is used to manage and analyze the multimodal data corresponding to the intelligent media resources. Based on the management and analysis results, it sequentially stores and manages each piece of multimodal data, and sets corresponding data management processes for the storage management results. This improves the management efficiency, flexibility, and utilization rate of the corresponding multimodal data within the intelligent media resources. The specific implementation process includes: The data acquisition module is used to collect relevant intelligent media resource data and perform modal recognition on the intelligent media resource data to obtain multimodal data information of the corresponding type. Its specific implementation process includes: Set up a data acquisition unit and a data recognition unit; The data acquisition unit is used to collect corresponding intelligent media resource data, and the process includes: Obtain the corresponding smart media device terminal, and set up corresponding media resource acquisition nodes within the smart media device terminal. Collect the smart media resource information corresponding to the smart media device terminal through the media resource acquisition nodes, including: Intelligent media device terminals are media terminals used by users to perform operations such as intelligent media resource production, exchange, access, or acquisition based on user needs and user accounts. The media resource acquisition node is the acquisition and monitoring node corresponding to the corresponding operation process within the corresponding smart media device terminal; Intelligent media resource information refers to the data collected by the corresponding user during the corresponding operation process when collecting and accepting orders for the corresponding media resources. The data identification unit is used to classify the acquired intelligent media resource data and obtain corresponding types of multimodal data information. The process includes: The data recognition analysis node in the data recognition unit is set according to the corresponding multi-modal data type, and the data recognition analysis node includes an image recognition analysis node, a text recognition analysis node, an audio recognition analysis node, and a video recognition analysis node; The corresponding intelligent media resource data is input into the corresponding data recognition analysis node, and the different types of recognition analysis nodes in the corresponding data recognition analysis node are sequentially traversed and analyzed to output corresponding modal recognition information, wherein: If only a single recognition analysis node in the corresponding data recognition analysis node recognizes the corresponding intelligent media resource data, the corresponding intelligent media resource data is marked as single modal data information; If two or more recognition analysis nodes in the corresponding data recognition analysis node recognize the corresponding intelligent media resource data, the corresponding intelligent media resource data is marked as mixed modal data information; The recognition results of each intelligent media resource data obtained by the data recognition analysis node are marked and processed, the corresponding single modal data information and mixed modal data information are marked and processed according to the corresponding modal type, and the corresponding multi-modal data information is obtained.

[0021] The data processing module is used for mapping processing of the obtained multi-modal data information, classification processing according to the mapping processing result, and obtaining of corresponding resource distribution basis element information, and the specific implementation process includes: The data processing unit and the data classification unit are set; The data processing unit is used for obtaining the corresponding multi-modal data information, mapping processing of the multi-modal data information, analysis processing of the corresponding multi-modal data information, and obtaining of corresponding semantic representation information, and the process includes: The multi-modal data information corresponding to the image type is analyzed visually, the corresponding multi-modal data information is analyzed and processed based on the CNN model and the corresponding target detection model, the target region information is obtained according to the analysis processing result, the obtained target region information is analyzed semantically, and the corresponding semantic representation information is obtained according to the semantic analysis result of each target region information; The multi-modal data information corresponding to the text type is analyzed textually, the corresponding keyword information and context semantic information are obtained, the obtained keyword information and context semantic information are analyzed and processed, and the semantic representation information corresponding to the corresponding multi-modal data information is obtained; The acoustic feature processing is performed on the multi-modal data information corresponding to the audio type, the semantic sentiment information, the semantic keyword information and the front and rear segment semantic information are respectively obtained according to the acoustic feature processing result, the obtained acoustic information is subjected to semantic integration analysis, and the semantic representation information corresponding to the corresponding multi-modal data information is obtained; The frame segment processing and the audio processing are performed on the multi-modal data information corresponding to the video type, the multi-modal data information of the image type corresponding to the corresponding frame segment and the multi-modal data information of the audio type are obtained, the frame segment semantic information is obtained according to the data processing method of the corresponding type, the dynamic deviation processing is performed on the corresponding frame segment semantic information according to the frame segment sorting, the joint feature between the frame segments is obtained, the corresponding joint feature is subjected to semantic integration analysis, and the semantic representation information corresponding to the corresponding multi-modal data information is obtained; It should be further explained that, in the specific implementation process, the semantic representation information corresponding to the multi-modal data information corresponding to the image recognition analysis node, the text recognition analysis node, the audio recognition analysis node and the video recognition analysis node includes meaningful data information extracted from the corresponding data information for expression understanding; The obtained semantic representation information is subjected to connection processing, the semantic mapping vector corresponding to the multi-modal data information is obtained, and the process includes: The corresponding semantic mapping space is set according to the corresponding semantic keyword information and the semantic sentiment information, and the distribution of the corresponding semantic keyword information and the semantic sentiment information in the semantic mapping space is included; The distribution of the semantic keyword information and the semantic sentiment information corresponding to the semantic representation information corresponding to the corresponding multi-modal data information in the semantic mapping space is obtained, the corresponding semantic mapping vector is obtained according to the corresponding distribution, the position information corresponding to the corresponding semantic keyword information and the semantic sentiment information is included in the semantic mapping vector, the corresponding information and the connection relationship are marked as a mapping connection node, and the semantic mapping vector is subjected to marking processing; The data classification unit is used for obtaining the semantic mapping vector corresponding to the corresponding multi-modal data information, performing feature analysis and extraction processing on the corresponding mapping connection node in the corresponding semantic mapping vector based on the principal component analysis method, and obtaining the semantic feature information corresponding to the semantic mapping vector; The classification evaluation interval data is set for the semantic feature information corresponding to the corresponding multi-modal data information in the media resource management center, the classification evaluation interval data includes evaluation data corresponding to a plurality of hierarchical classification intervals, the plurality of hierarchical classification intervals respectively include a plurality of word-level feature intervals, a plurality of semantic feature intervals and a plurality of scene feature intervals, and the corresponding classification evaluation interval data is set for the corresponding feature intervals; The corresponding semantic feature information is compared and analyzed with the corresponding classification evaluation interval data, the classification evaluation interval data corresponding to the multi-level classification interval is obtained in sequence, it is judged that the semantic feature information corresponding to the corresponding multi-modal data information belongs to the classification interval, the comparison result is set as the corresponding interval evaluation data, and the interval evaluation data respectively includes the word-level feature interval, the semantic feature interval and the scene feature interval in the multi-level classification interval; The interval evaluation data corresponding to the corresponding multi-modal data information is obtained, and the corresponding resource distribution basis element information is set according to the corresponding feature interval.

[0022] The data distribution module is used for analyzing and processing according to the corresponding resource distribution basis element information, setting the corresponding modal basic information and modal identification information, distributing the corresponding multi-modal data information based on the blockchain technology, and constructing a multi-modal data distribution storage library, and the specific implementation process includes: A data identification unit and a data storage unit are set; The data identification unit is provided with a basic storage space and an identification storage space, multi-modal data information is obtained, the basic storage space is used for storing the obtained multi-modal data information, the storage code is set according to the corresponding semantic feature information of the corresponding multi-modal data information and the corresponding semantic feature information of the corresponding multi-modal data information; A basic code library is set, the encoding scrambling characters and the encoding scrambling algorithms are respectively set according to the corresponding modal type and the feature interval, the corresponding encoding characters are extracted from the multi-modal data information and the semantic feature information according to the encoding scrambling characters and the encoding scrambling algorithms, the extracted encoding characters are operated and processed according to the encoding scrambling algorithms, the operation processing results are integrated respectively, and the corresponding storage code is obtained; The corresponding multi-modal data information is marked by the corresponding storage code, and the corresponding modal basic information is obtained; The resource distribution basis element information corresponding to the corresponding multi-modal data information is obtained, the basic storage space is obtained according to the resource distribution basis element information, and the basic storage space is obtained according to the combination result of the corresponding word-level feature interval, the semantic feature interval and the scene feature interval in the resource distribution basis element information; The identification code is set in the basic storage space according to the corresponding modal basic information, the resource distribution basis element information and the corresponding semantic mapping vector; The identification code library is set according to the corresponding modal type, the semantic feature information in the feature interval and the semantic mapping vector, the modal basic information corresponding to the corresponding multi-modal data information is encoded based on the identification code library, the corresponding identification code is obtained, and the modal identification information is set; Store the obtained modal identification information into the identification storage space; The data storage unit is configured to obtain modal base information and modal identification information stored in the base storage space and the identification storage space respectively, set the base storage space as a block center node, and set each identification storage space as a corresponding block sub-node; The correlation data between the corresponding semantic mapping vectors is used to process the spatial layout of the corresponding modal identification information in each block sub-node, and the position information of the block sub-node to which the corresponding modal identification information belongs is obtained. According to the position information of the corresponding block sub-node to which the corresponding modal identification information belongs, the corresponding associated storage information in the corresponding modal base information in the corresponding block center node is set, the associated storage information corresponding to the corresponding modal base information in the block center node and the block sub-node and the corresponding position information are set in a chain storage structure, and the corresponding multi-modal data distribution storage library is constructed according to the storage results in the block center node and the block sub-node and the corresponding chain storage structure.

[0023] The data retrieval module is configured to obtain corresponding personalized retrieval data according to user demand, and obtain a corresponding personalized retrieval database in the multi-modal data distribution storage library according to the personalized retrieval data, and the specific implementation process includes: setting a personal collection unit and a personal retrieval unit; The personal collection unit is connected with the corresponding user account, and the personalized retrieval data corresponding to the user demand is obtained according to the corresponding user account, and the personalized retrieval data includes retrieval keyword information, user preference information, and retrieval expansion information; The personal retrieval unit obtains the corresponding personalized retrieval data and the multi-modal data distribution storage library, respectively processes the obtained personalized retrieval data in terms of semantics and mapping, obtains corresponding personalized semantic representation information and retrieval mapping vectors, and the retrieval mapping vectors include position information corresponding to semantic keyword information and semantic sentiment information in the corresponding personalized retrieval data; The position information corresponding to the semantic keyword information and the semantic sentiment information in the obtained retrieval mapping vector is subjected to feature analysis and extraction processing based on a principal component analysis method, and retrieval feature information of the corresponding retrieval mapping vector is obtained; The obtained retrieval feature information is input into the multi-modal data distribution storage library, the modal base information existing in association is obtained, the modal identification information corresponding to the obtained associated storage information is obtained according to the obtained modal base information, the corresponding retrieval modal information is obtained according to the distribution of the modal identification information, and the distribution of the corresponding multi-modal data distribution storage library to which the obtained retrieval modal information belongs is extracted in sequence to construct a corresponding personalized retrieval database.

[0024] The data analysis module is used for demand integration according to the corresponding personalized search database, generates personalized demand output information feedback to the corresponding user account, and is processed by the user account to obtain user demand feedback information, and the specific implementation process includes: Set up a demand analysis unit and a demand feedback unit; The demand analysis unit is used to obtain the corresponding personalized search database, and fuse the corresponding multi-modal data information in the personalized search database; Based on deep learning algorithm, the historical personalized search database is analyzed and processed, a multi-modal fusion model is constructed, the corresponding personalized search database is input into the multi-modal fusion model for analysis and processing, and the corresponding personalized demand output information is output and sent to the corresponding user account; The demand feedback unit is used for the user to obtain personalized demand output information through the corresponding user account, to obtain the personalized demand output information, to carry out multi-angle feedback evaluation, and to obtain the corresponding angle evaluation result. The quantization processing result is obtained, and the corresponding user demand feedback information is obtained.

[0025] The data management module re-distributes the corresponding multi-modal data distribution storage according to the individual personalized search database, the personalized demand output information and the user demand feedback information, and obtains the multi-modal data storage, and the specific implementation process includes: Set up a statistical analysis unit and a data management unit; The statistical analysis unit is used to obtain the corresponding personalized search database, personalized demand output information and user demand feedback information of the corresponding user account, and to analyze the data information obtained by the corresponding user account, including: Respectively obtain the quantization processing result of the data quality angle, the semantic understanding angle, the demand matching angle and the data fusion angle; The quantization processing results corresponding to the semantic understanding angle and the demand matching angle are statistically analyzed, the quantization processing results are classified into intervals, the statistical results are obtained according to the interval classification processing results, and the statistical proportion data corresponding to each classification interval is obtained; Set up a distribution management reference curve, compare the statistical proportion data with the corresponding parameter information in the corresponding distribution management reference curve, judge whether distribution management is needed, and mark the judgment result; Statistical analysis is performed on the quantization results corresponding to the data quality and data fusion angles, interval classification processing is performed on the quantization results, and statistical analysis is performed on the interval classification processing results to obtain statistical proportion data corresponding to each classification interval; A model management reference curve is set, the statistical proportion data is compared and analyzed with the corresponding parameter information in the model management reference curve, it is judged whether model management is needed, and marking processing is performed according to the judgment result; According to the statistical analysis result of the user demand feedback information, distribution statistical data corresponding to the corresponding multi-modal data information is obtained, and the distribution statistical data includes distribution management and model management marking between the corresponding multi-modal data information in the corresponding personalized retrieval database; The data management unit obtains the distribution statistical data corresponding to the corresponding multi-modal data information; If the corresponding distribution statistical data includes the marking result of the distribution management between the corresponding multi-modal data information, the corresponding multi-modal data distribution storage library is redistributed, the corresponding multi-modal data information is rearranged, an updated multi-modal data distribution storage library is obtained, and an update mark is performed; If the corresponding distribution statistical data includes the marking result of the model management between the corresponding multi-modal data information, the corresponding multi-modal fusion model is reanalyzed, an updated multi-modal fusion model is obtained, and an update mark is performed.

[0026] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and modifications can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A multi-modal data management system for intelligent media assets, comprising a media asset management center, characterized in that, The media resource management center comprises a data collection module, a data processing module, a data distribution module, a data retrieval module, a data analysis module and a data management module; The data collection module is configured to collect corresponding intelligent media resource data, perform modal recognition on the intelligent media resource data, and obtain multi-modal data information of a corresponding type; The data processing module is configured to perform mapping processing on the obtained multi-modal data information respectively, perform classification processing according to the mapping processing result, and obtain corresponding resource distribution basis element information; The data distribution module is configured to analyze and process the resource distribution basis element information, set corresponding modal basic information and modal identification information, perform distribution processing on the corresponding multi-modal data information based on a blockchain technology, and construct a multi-modal data distribution storage library; The data retrieval module is configured to obtain corresponding personalized retrieval data according to user demand, and obtain a corresponding personalized retrieval database in the multi-modal data distribution storage library according to the personalized retrieval data; The data analysis module is configured to integrate demand according to the corresponding personalized retrieval database, generate personalized demand output information, feed back the personalized demand output information to a corresponding user account, and obtain user demand feedback information through feedback processing by the user account; The data management module performs redistribution processing on the corresponding multi-modal data distribution storage library according to each personalized retrieval database, personalized demand output information and user demand feedback information, and obtains a multi-modal data storage library.

2. The multi-modal data management system of intelligent media resources according to claim 1, wherein, The process of obtaining multi-modal data information of a corresponding type includes: setting a data collection unit and a data recognition unit; collecting intelligent media resource information corresponding to operation results of a corresponding user according to user demand and a user account in a corresponding intelligent media equipment terminal through the data collection unit; identifying and analyzing the collected intelligent media resource information through the data recognition unit, setting a data identification analysis node, inputting the intelligent media resource information into the data identification analysis node, sequentially traversing the identification analysis nodes of corresponding modal types, obtaining corresponding modal identification information, and obtaining corresponding multi-modal data information according to the modal identification information.

3. The multi-modal data management system of intelligent media resources according to claim 2, wherein, The process of obtaining resource distribution basis element information includes: setting a data processing unit and a data classification unit; The data processing unit is configured to obtain corresponding multi-modal data information, perform mapping processing on the multi-modal data information, analyze and process the corresponding multi-modal data information respectively, obtain semantic representation information, connect the obtained semantic representation information, obtain a semantic mapping vector corresponding to the multi-modal data information, and mark the semantic mapping vector; The data classification unit is configured to obtain a semantic mapping vector corresponding to the corresponding multi-modal data information, sequentially extract features from the corresponding semantic mapping vector according to a mapping connection node, and obtain semantic feature information corresponding to the semantic mapping vector; The semantic feature information corresponding to the corresponding multi-modal data information in the media resource management center is set with classification evaluation interval data, the semantic feature information corresponding to the corresponding semantic mapping vector is compared and analyzed with the corresponding classification evaluation interval data, interval evaluation data is obtained, and the corresponding resource distribution basis element information is set according to the interval evaluation data corresponding to the multi-modal data information.

4. The multi-modal data management system of intelligent media resources according to claim 3, wherein, The process of setting the modal basic information and the modal identification information includes: setting a data identification unit and a data storage unit; The basic storage space and the identification storage space are set in the data identification unit, the multi-modal data information is obtained, the obtained multi-modal data information is stored in the basic storage space, the storage code is set for the corresponding multi-modal data information according to the storage result, and the modal basic information is set according to the corresponding storage code and the corresponding multi-modal data information; The resource distribution basis element information corresponding to the corresponding multi-modal data information is obtained, the basic storage space is obtained according to the resource distribution basis element information, the identification code is set in the basic storage space according to the corresponding modal basic information, the resource distribution basis element information and the corresponding semantic mapping vector, the modal identification information is set according to the identification code, and the modal identification information is stored in the identification storage space.

5. The multi-modal data management system of intelligent media resources according to claim 4, characterized in that, The process of constructing the multi-modal data distribution storage library includes: The data storage unit is used to obtain the modal basic information and the modal identification information stored in the basic storage space and the identification storage space, the basic storage space is set as a block center node, and each identification storage space is set as a corresponding block sub-node; The spatial layout processing of the correlation data between the corresponding semantic mapping vectors of the corresponding modal identification information in each block sub-node is obtained, and the position information of the block sub-node to which the corresponding modal identification information belongs is obtained; The corresponding associated storage information is set in the corresponding modal basic information in the corresponding block center node according to the position information of the corresponding block sub-node to which the corresponding modal identification information belongs, the associated storage information corresponding to the corresponding modal basic information in the block center node and the block sub-node and the corresponding position information are set as a chain storage structure, and the multi-modal data distribution storage library is constructed according to the storage results in the block center node and the block sub-node and the corresponding chain storage structure.

6. The multi-modal data management system of intelligent media resources according to claim 5, wherein, The process of obtaining the personalized retrieval database includes: setting a personalized collection unit and a personalized retrieval unit; The personalized collection unit is connected with the corresponding user account, the personalized retrieval data corresponding to the corresponding user demand is obtained according to the corresponding user account, and the personalized retrieval data includes retrieval keyword information, user preference information and retrieval expansion information; The personalized retrieval unit obtains the corresponding personalized retrieval data and the multi-modal data distribution storage library, maps the obtained personalized retrieval data respectively, obtains the corresponding personalized semantic representation information, and sets the retrieval mapping vector according to the personalized semantic representation information; The obtained search mapping vector is input into the multi-modal data distribution repository, the associated modal basic information is obtained, the corresponding modal identification information corresponding to the obtained associated storage information is obtained according to the obtained modal basic information, the corresponding search modal information is obtained according to the corresponding distribution in the modal identification information, the distribution of the corresponding multi-modal data distribution repository to which the obtained search modal information belongs is extracted in turn, and a corresponding personalized search database is constructed.

7. The multi-modal data management system of intelligent media resources according to claim 6, wherein, The process of obtaining user demand feedback information includes: setting a demand analysis unit and a demand feedback unit; the demand analysis unit is used for obtaining a corresponding personalized search database, and fusing corresponding multi-modal data information in the personalized search database; based on a deep learning algorithm, the historical personalized search database is analyzed and processed to construct a multi-modal fusion model, the corresponding personalized search database is input into the multi-modal fusion model for analysis and processing, the corresponding personalized demand output information is output, and the personalized demand output information is sent to the corresponding user account; the demand feedback unit is used for a user to obtain personalized demand output information through a corresponding user account, to perform multi-angle feedback evaluation on the obtained personalized demand output information, to quantitatively process the multi-angle feedback evaluation in sequence, and to generate user demand feedback information according to the quantitatively processed results.

8. The multi-modal data management system of intelligent media resources according to claim 7, wherein, The process of redistributing the multi-modal data distribution repository includes: setting a statistical analysis unit and a data management unit; the statistical analysis unit is used for obtaining the personalized search database, the personalized demand output information and the user demand feedback information corresponding to the corresponding user account, respectively performing statistical analysis on the data information obtained by the corresponding user account, and respectively obtaining the distribution statistical data corresponding to the corresponding multi-modal data information; the data management unit obtains the distribution statistical data corresponding to the corresponding multi-modal data information, performs redistribution processing on the current multi-modal data distribution repository according to the corresponding distribution statistical data, performs re-layout on the corresponding multi-modal data information, obtains the multi-modal data distribution repository after the layout is updated, and performs update marking.

Images