Media data transaction method based on block chain

Abstract

The invention relates to the technical field of income management, and discloses a blockchain-based media data transaction method, which comprises the following steps: obtaining original media data, transmitting a generated feature representation to a curvature calculation module, obtaining modified media data, and transmitting the modified media data to the curvature calculation module; and transmitting the result of the geodesic distance and contribution data of the participants to an income distribution module, checking the security and data integrity of the transaction according to a dual verification mechanism after income distribution, finally optimizing the cooperative relationship among all participants through a hypergraph model, and returning an optimization result by a network optimization module. According to the invention, the data analysis module is adopted to carry out data deep analysis on the income management platform, the technical effect of timely discovering non-uniform income distribution is achieved, compared with a method which only depends on manual auditing in the prior art, the defects that the efficiency is low and errors are easy to occur are overcome, and the auditing accuracy is improved due to the characteristic of system automation.

Description

Technical Field

[0001] The present invention relates to the technical field of revenue management, and particularly to a media data trading method based on blockchain. Background Art

[0002] In the current revenue management field, traditional revenue distribution mechanisms often rely on manual review and management. Although this method can achieve revenue distribution to a certain extent, due to the limitations of manual operations, the review efficiency is often low. The revenue distribution of some participants may have errors, affecting the overall interests. This phenomenon is particularly obvious in large-scale transactions, and frequent errors reduce the trust of all parties in the system.

[0003] In addition, existing auditing technologies often only focus on a single dimension of transactions and lack a comprehensive analysis of data. This makes it impossible for management to evaluate the rationality and consistency of revenue distribution from multiple perspectives, ultimately leading to mistakes in overall decision-making. This single analysis mode limits the enterprise's ability to dig out the problems hidden behind complex data, and the risk management ability is limited; moreover, many existing technologies lack timeliness in dealing with abnormal transactions. The traditional mode often relies on the ex-post review of historical data and fails to monitor potential abnormal transaction situations in real time. This post-audit method allows potential risks to accumulate continuously without timely response, increasing the compliance risks faced by enterprises; finally, poor information communication is also one of the important deficiencies of existing technologies. Existing systems often fail to effectively integrate various types of data, resulting in information silos. This phenomenon causes managers to rely on fragmented information when making decisions, lacking the necessary background and global perspective, making the decision-making process long and inefficient, and further affecting the optimization ability of revenue distribution.

[0004] By analyzing the deficiencies of existing technologies, it can be concluded that traditional revenue management methods urgently need innovation and improvement to meet the needs of modern enterprises for efficient, transparent, and secure revenue distribution mechanisms. Summary of the Invention

[0005] Aiming at the deficiencies of existing technologies, the present invention provides a media data trading method based on blockchain, which solves the problems of low efficiency, high error rate, and insufficient risk management existing in the traditional revenue distribution process, and realizes the optimization of revenue distribution.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: A media data trading method based on blockchain includes the following steps: S1: Obtain the original media data, then extract its features, and use deep learning algorithms to generate feature representations in a high-dimensional manifold space; S2: Transmit the generated feature representation to the curvature calculation module, calculate the curvature of the manifold, generate the unique fingerprint of the data, and store the fingerprint in the blockchain to ensure the immutability of the data; S3: Obtain the modified media data, perform the same feature extraction as in step S1, and transmit the new feature representation to the geodesic distance calculation module between manifolds to calculate the geodesic distance between manifolds and quantify the degree of data change; S4: Transmit the result of the geodesic distance and the contribution data of the participating parties to the revenue distribution module; S5: After revenue distribution, check the security and data integrity of the transaction according to the double verification mechanism. First, compare the content-addressable ID and the hash value to ensure the physical locking of the data. Second, use zk-SNARKs to verify the access rights and verify the effectiveness of the logical locking; S6: Finally, optimize the collaboration relationship between all participating parties through the hypergraph model. The optimization result is returned by the network optimization module and recorded in the blockchain to achieve more efficient processing of subsequent transactions.

[0007] Preferably, the feature extraction step adopts a deep learning algorithm, specifically including a convolutional neural network or a Mel-frequency cepstral coefficient extraction algorithm. The feature vector is represented as a vector in a high-dimensional feature space: ; Where: represents crossing the eigenvalue of the data, is the dimension of the feature vector.

[0008] Preferably, the curvature calculation method is based on Riemannian geometry, and the Riemannian curvature tensor is used to evaluate the uniqueness and difference of the manifold. The specific calculation formula is: ; Where: is the curvature of the manifold ; is the Riemannian curvature tensor, representing the degree of bending of the tangent vector in the manifold; and are the tangent vectors on the manifold ; represents the inner product operation; and represent the norms of the vectors and respectively.

[0009] Preferably, the calculation of the geodesic distance is based on the metric tensor on the manifold, ensuring the continuity and traceability of the rheological characteristics in the high-dimensional space. The specific calculation method is as follows: ; Where: is the manifold and the geodesic distance between; is the path function connecting the two manifolds, representing the curve between the manifolds; is the metric tensor along the path ; is the tangent vector of the path.

[0010] Preferably, in the process of dynamically adjusting the revenue weight, the Lorenz equation is used to describe the interaction relationship between the participants. The Lorenz equation is: ; Where: are the different weight values of the participants in the dynamic game; is the strength parameter of the system, indicating the speed of weight change; is the parameter, representing the feedback of the system; is the attenuation rate of the system, reflecting the attenuation characteristics of the weight over time; represents the variable changes over time and is proportional to the difference between and . That is, when is greater than , the growth rate is positive, otherwise it is negative. Here is a positive constant, representing the rate of this growth; represents the rate of change of the variable . It consists of two parts: the first part represents multiplied by , indicating that affects grows over time. When is less than the constant , the growth rate is positive; when is greater than , The rate of change is negative. The second part indicates itself is decreasing its value, forming a self-regulating effect; represents the variable The rate of change over time consists of two parts: indicates and the product of indicates that their interaction promotes the increase of. While is decreasing the value of is a positive constant, representing the competition or decay effect.

[0011] Preferably, the dual verification mechanism includes comparing the content-addressable ID and the hash value to ensure the effectiveness of physical and logical locking of data, wherein the logical locking uses zk-SNARKs to verify the effectiveness of access rights, and the zk-SNARKs generator includes a specific calculation process for evidence to ensure the persistence and security of data.

[0012] Preferably, the hypergraph model models the multi-party collaboration relationship by defining hyperedges and nodes, and the hypergraph Laplacian matrix calculation formula is: ; where: is the Laplacian matrix of the hypergraph; is the degree matrix, and its diagonal elements represent the degrees of multiple nodes; is the adjacency matrix, representing the connection relationship between nodes.

[0013] Preferably, a device for implementing the method according to any one of claims 1 to 7, the device includes: A feature extraction module for extracting features from the original media data and generating a high-dimensional manifold space representation; A curvature calculation module for calculating the curvature of the manifold and generating a unique fingerprint; A revenue distribution module for automatically calculating and distributing revenue based on the contributions and rheological characteristics of the participating parties, and the revenue distribution module realizes automated intelligent contract management; A verification module for performing integrity checks on transactions through quantum entanglement states and anti-quantum hash algorithms;[[ID=...]] A network optimization module for optimizing the multi-party collaboration relationship and data exchange process based on the hypergraph model.

[0014] Preferably, the feature extraction module adopts deep learning algorithms to improve the accuracy and efficiency of feature extraction. The algorithms include convolutional neural network and Mel-frequency cepstral coefficient extraction algorithm.

[0015] Preferably, the revenue distribution module is connected to the blockchain to adjust the revenue weights of participants in real time, so as to ensure the completion of revenue distribution immediately after the transaction. The blockchain adopts a decentralized method to ensure the transparency and security of data.

[0016] The present invention provides a media data trading method based on blockchain, having the following beneficial effects: 1. The present invention uses the data analysis module to deeply analyze the data of the revenue management platform, achieving the technical effect of timely discovering uneven revenue distribution. Compared with the method that only relies on manual review in the prior art, it solves the problems of low efficiency and easy error. The automated feature of the system improves the accuracy of review.

[0017] 2. The present invention integrates the auditing and analysis functions, and can comprehensively evaluate historical transaction data to achieve comprehensive and transparent operation decisions. Compared with the solution that processes transaction records in isolation in the prior art, it solves the problem of information silos, ensures the integrity and consistency of data, and reduces potential compliance risks.

[0018] 3. The present invention introduces the generation of visual reports, improving the intuitiveness of data presentation and facilitating the management to quickly understand the analysis results. Different from the complex data arrangement method in the prior art, it solves the problem that information is not easy to convey, promotes efficient decision-making, and shortens the response time.

[0019] 4. The present invention provides an abnormal transaction identification mechanism to monitor potential risks in real time. Compared with the method of retrospective tracing after the event in the prior art, the present invention can give early warnings before problems occur, solve the problem of insufficient risk management, and thus enhance the security and stability of the platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic flowchart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0022] Please refer to the attach Figure 1 , the embodiments of the present invention provide a media data trading method based on blockchain, including: S1: Obtain the original media data, then perform feature extraction on it, and use a deep learning algorithm to generate a feature representation in a high-dimensional manifold space; Specifically, first, establish a data acquisition module for collecting the original media data. The original media data may include text, audio, video, and image formats. The data sources can be multiple channels, such as social media platforms, news websites, online video services, or user-uploaded content. The design of the data acquisition module should ensure that a large amount of data can be obtained from these channels in real time, stably, and efficiently to ensure the extensiveness and representativeness of the data.

[0023] Then, process the obtained original media data through a feature extraction module. According to different types of data, different feature extraction algorithms are adopted: For image data, the feature extraction module preferably uses a convolutional neural network (CNN). The CNN structure usually includes multiple convolutional layers, activation function layers (such as RelU), pooling layers, and fully connected layers. In the feature extraction process, the following process is used to calculate the feature vector: ; Where: is the extracted image feature vector; is the input image data.

[0024] CNN realizes the gradual abstraction of features through a multi-layer design, and the finally generated feature vector is a -dimensional vector: ; Where: is the dimension of the image feature vector; represents the th image feature value.

[0025] For audio data, the feature extraction module uses Mel-frequency cepstral coefficients (MFCC) for processing. This algorithm obtains the audio feature vector by performing a short-time Fourier transform (STFT) on the audio signal and a linear transform in the Mel-frequency domain: ; Where: is the extracted audio feature vector; is the input audio signal.

[0026] The feature vector generated by MFCC is a One-dimensional vector: ; Wherein: is the dimension of the audio feature vector; represents the th audio eigenvalue.

[0027] For text data, the feature extraction module uses a long short-term memory network (LSTM) to capture the sequential information of the text. The LSTM architecture includes an input layer, a hidden layer (including multiple LSTM cells), and an output layer. The feature extraction process is as follows: ; Wherein: is the extracted text feature vector; is the input text sequence.

[0028] The feature vector generated by the LSTM is a -dimensional vector: ; Wherein: is the dimension of the text feature vector; represents the th text eigenvalue.

[0029] Subsequently, the extracted feature vectors must be uniformly processed into a single form to adapt to subsequent analysis. To this end, the three feature vectors are concatenated or fused to obtain the final comprehensive feature vector : ; Wherein: is the final feature vector, containing the comprehensive features of all media data, with a dimension of .

[0030] The next step is the transfer of the feature vector. The feature extraction module and the curvature calculation module are effectively connected through an interface to ensure that the high-dimensional features can be correctly transmitted. In a specific implementation, the feature vector will be used as an input and passed to the subsequent curvature calculation module to enter the calculation process of step S2.

[0031] In this embodiment, through a reasonable structure design and deep learning algorithm, efficient feature extraction of various types of media data is achieved. The generated feature vector , which can accurately reflect the internal characteristics of the data and has good subsequent analysis capabilities. The technical effect of this step is to achieve a deep understanding of the original media content and lay a solid foundation for subsequent data verification and revenue calculation. Therefore, this embodiment can effectively solve the problem of information loss in the feature extraction process and ensure the accuracy and effectiveness of subsequent operations.

[0032] S2: Transmit the generated feature representation to the curvature calculation module, calculate the curvature of the manifold, generate a unique fingerprint of the data, and store the fingerprint in the blockchain to ensure the immutability of the data; Specifically, first, establish a curvature calculation module to receive the high-dimensional feature vector from step S1 , expressed as: ; Where: represents the total dimension of the feature vector; represents the th eigenvalue in the feature vector, .

[0033] Next, the curvature calculation module constructs a feature manifold , and the manifold is defined by the continuous data points in the feature vector . To calculate the curvature of the manifold, appropriate tangent vectors and need to be selected from within the manifold for subsequent curvature calculation.

[0034] Set the tangent vector as the local change of the feature vector, defined as: ; Where: represents the dimension of the tangent vector ; represents the dimension of the tangent vector ; and are both tangent vectors on the manifold , representing the change in the local direction.

[0035] Curvature calculation formula and parameter definition The specific formula for curvature calculation is: ; Where: is the curvature of the manifold , representing the degree of bending of the manifold; is the Riemann curvature tensor, representing the curvature characteristics of the manifold at a certain point; represents the inner product operation, used to calculate the interaction between tangent vectors; : calculates the tangent vectors in the manifold and The result of the action reflects the curvature change in these two directions; : represents the tangent vector The square of the norm; : represents the tangent vector The square of the norm; : represents the tangent vector and The inner product; Here, the curvature value of the manifold will be used to generate a data uniqueness fingerprint later to ensure the security and uniqueness of the data.

[0036] Based on the calculated curvature , generate a uniqueness fingerprint. This fingerprint can be defined as follows: Fingerprint = Hash ; Where: The fingerprint is a value generated by a hash function, used to uniquely identify the data; Hash( ) represents the hash process, which combines the input curvature value and feature vector into a fixed-length hash value to ensure the unique identification of the data.

[0037] In this step, the curvature calculation module and the feature extraction module are connected through a standard data interface to ensure that the feature vector can be transferred smoothly. The feature vector needs to be processed by the curvature calculation module, and the finally generated curvature value and the feature vector together form a uniqueness fingerprint, which is then output to the blockchain module for storage and confirmation.

[0038] Through this series of mathematical models and processes, step S2 can accurately describe the geometric characteristics of the data manifold, achieving the purpose of generating a uniqueness fingerprint. This process effectively ensures the secure identification and immutability of the data content, providing a solid foundation for subsequent revenue distribution and rights confirmation. Therefore, this implementation method has important practicality and technical effects in the entire data transaction process.

[0039] S3: Obtain the modified media data, perform the same feature extraction as in step (1), and pass the new feature representation to the geodesic distance calculation module between manifolds to calculate the geodesic distance between manifolds and quantify the degree of data change; Specifically, first, establish a data verification module. The main function of this module is to receive the feature vector FF generated from step S1 and the unique fingerprint generated from step S2. The verification module verifies the consistency and integrity of the input media data. This module consists of a data storage part, a verification algorithm part, and a result feedback part.

[0040] The data storage part is responsible for saving all previously generated feature vectors and their corresponding unique fingerprints for subsequent comparison. When verifying the data, first retrieve the feature vector corresponding to the current input data and its associated unique fingerprint.

[0041] ; Where: represents the th eigenvalue in the modified feature vector, ; is the total dimension of the feature vector.

[0042] Next, perform feature comparison and judgment. Calculate the Euclidean distance between the input data and the generated feature vector , which is expressed as: ; Where: represents the Euclidean distance between the feature vectors, reflecting the similarity between the two; is the th eigenvalue in the input feature vector, ; is the th eigenvalue in the modified feature vector; represents the square of the component difference between the calculated vectors and in the th dimension. This reflects the difference between the two vectors in each dimension; represents the sum of the squares of the differences for all dimensions to obtain an overall difference measure; It means that finally the square root of the sum is taken to obtain the actual distance between the two vectors. This is because the definition of distance needs to be non - negative, and the square root ensures the unit consistency of the result.

[0043] To ensure the validity of the data, a distance threshold ε is set. This value is a preset tolerance used to judge the consistency of the data. If the calculated distance satisfies the following conditions, the data is considered valid: ; Where: In this formula, ε represents the threshold parameter.

[0044] The revenue calculation module is used to allocate revenue according to the feature vectors and their similarities. The revenue function can be expressed as: ; Where: represents the revenue of the participant ; represents the dynamic weight related to the participant ; represents the vector and The distance between them, here it may represent the difference between the feature vector of the possible trading party and a certain reference vector; represents the weighted sum of the weights and distances of all trading parties . This normalization process ensures that the calculated revenue has relative fairness, enabling the revenues of different trading parties to be compared under the same standard; The total transaction amount represents the total amount jointly recognized by all participants in the current transaction.

[0045] The dynamic weight is calculated based on the contribution degree of the participant in previous transactions and can be dynamically adjusted through the following formula: ; Where: represents the contribution degree of the participant in the current transaction, based on the quantity and quality of the resources it submits; represents the total revenue obtained by the participant in past transactions, reflecting its historical performance in the system; represents the participant The level of activity on the platform can be measured by the frequency or quantity of transactions participated in; α, β, and γ are adjustment coefficients used to control the influence of different factors on the dynamic weight.

[0046] After verification and profit calculation, the results need to be recorded. The final output results include: All participating parties and their corresponding profits , which will be recorded in the current transaction record.

[0047] The generated profit information will be stored through the blockchain module to ensure the transparency and verifiability of all transaction records. The recorded content includes: The unique fingerprint of the participating party and its corresponding profit information .

[0048] Through these mechanisms, step S3 effectively realizes the consistency verification of media data and reasonable profit distribution, ensuring the protection of the rights and interests of participating parties. The whole process guarantees the security and transparency of data transactions and ensures the immutability of data through blockchain technology. Such a design not only improves the efficiency of data transactions but also provides a fair platform for creators and sharers to ensure that their contributions are reasonably rewarded.

[0049] S4: Transfer the results of the geodesic distance and the contribution data of the participating parties to the profit distribution module; Specifically, first, establish a profit distribution module. This module is responsible for receiving the profit calculation results and participating party information from step S3 and implementing profit distribution. This module can be divided into a profit input part, a distribution algorithm part, and a confirmation output part.

[0050] The profit input part first receives the information of the participating party and its corresponding profit, expressed as: ; Where: represents the participating party 's profit value, ; represents the total number of participating parties; This profit vector is generated from the profit calculation results in step S3 to ensure that the profit information of all participating parties is accurately recorded.

[0051] During the profit distribution process, the smart contract executes the distribution according to the preset profit distribution rules. The profit distribution rules can be expressed as: Total profit; Where: Represents the benefit value of the participant ; Represents the dynamic weight related to the participant , covering its contribution degree in the benefit calculation; The total benefit represents the total benefit of all participants.

[0052] Dynamic weight The calculation is based on the contribution degree of the participant in the current transaction and is determined by the following formula: ; Where: Represents the contribution degree of the participant , usually reflecting the comprehensive value of the content, resources or services provided by it in the transaction; Represents the contribution degree of the participant ;

[0053] Through the above formula, the standardized weight enables the contribution degree of each participant to be relatively fairly reflected in the benefit distribution. After the benefit distribution is completed, the actual transfer amount of each participant is determined : ; Where: Represents the transfer amount allocated to the participant ;

[0054] The smart contract will be responsible for executing the following steps: Read the benefit information of the participants ; Calculate the benefit and transfer amount of each participant And conduct the transfer; Ensure the success and security of the transfer, and all transfer records will be recorded on the blockchain.

[0055] The format of the recorded content is as follows: Record = ; Where: The fingerprint is the unique identification information of the participant, coming from the unique fingerprint generation in step S2; Is the benefit value of the participant ; Is the amount transferred to the participant ; The timestamp records the specific time of creation to ensure the traceability of the time of each record.

[0056] During the entire revenue distribution process, the system should have a risk control mechanism to ensure the security and effectiveness of transactions. This mechanism includes: Identity authentication: Ensure that the unique fingerprint of the participating party is consistent with its identity information to prevent forgery or fraud.

[0057] Anomaly monitoring: Timely monitor abnormal situations during the revenue distribution process, such as excessive revenue fluctuations of a certain participating party, trigger a preset alarm and conduct manual review. Such records will also be saved to the blockchain for auditing.

[0058] The implementation process of step S4 includes: First, establish a revenue distribution module to receive the revenue information of the participating parties, then execute the revenue distribution according to the preset rules through a smart contract, determine the specific transfer operation by calculating the transfer amount, and finally feedback the revenue record to the blockchain to ensure transparency and immutability.

[0059] This implementation effectively combines the output information of steps S1, S2, and S3, achieving the efficiency and security of revenue distribution, ensuring that the rights and interests of each participating party can be fully protected. At the same time, the transparency and traceability of the entire process promote the fairness and security of transactions, helping to establish a transparent and trustworthy media data trading platform, improving transaction efficiency and reducing the risk of possible disputes. S5: After revenue distribution, check the security and data integrity of the transaction according to the dual verification mechanism. First, compare the content-addressable ID and the hash value to ensure the physical locking of the data. Second, use zk-SNARKs to verify the access rights and verify the effectiveness of the logical locking; Specifically, first, establish an audit module. The main function of this module is to receive the revenue distribution results and transaction records from step S4 and conduct auditing and retrospective analysis. This module can be divided into a data input section, an audit algorithm section, and a result feedback section.

[0060] The data input section first receives the transaction record and the corresponding revenue data, expressed as: ; Where: represents the th transaction record, , usually including information such as the participating party identifier, transaction amount, transaction time, etc.; represents the total number of transaction records.

[0061] The audit module conducts a review of data integrity and consistency through the following algorithm. First, it performs a validity check on the input transaction records to ensure that these records exist on the blockchain. The validity check can be achieved through the following formula: Validity = ; The set of valid transaction records can be defined as: ; Next, for all valid transaction records, income consistency confirmation is carried out. The specific steps are as follows: Obtain the due income of the participating parties from the blockchain and the actual transfer amount ; The following calculations are performed on each transaction record, which can be expressed as: ; If they are not equal, it is marked as an anomaly.

[0062] Where: represents the participating party 's due income, based on the income calculation result of step S3; represents the actual transfer amount allocated to the participating party from the transfer execution result of step S4.

[0063] During the audit process, if it is found that the income of a certain participating party does not match the actual transfer amount, the record will be marked as an anomaly. Anomaly tracking can be achieved through the following definition: Anomaly set , anomaly type, timestamp ; Where: represents the abnormal transaction record; is the income value that the participating party i should receive; is the actual transfer amount of the participating party i; The anomaly type indicates the type of problem found during the audit, such as "income mismatch", "missing record", etc.; The timestamp is the specific time when the anomaly is marked for tracking.

[0064] All abnormal records will be saved for subsequent analysis and rectification. The record format is: ; After the audit is completed, the Result Feedback Department will generate an audit report. The content of this report includes the audit results, detailed records of abnormal situations, rationalization suggestions, and rectification measures. The report structure can be expressed as: ; Among them: The audit time is the specific time when the audit is completed; The audit status indicates the audit results, such as "passed" or "not passed"; The number of abnormal transactions represents the quantity of abnormal transactions discovered during the audit process; The recommended rectification measures are specific rectification suggestions for abnormal situations, including measures and responsible departments.

[0065] Based on the audit results, the system will provide decision-making support for the management. The audit report is marked in the storage system for subsequent query and management, and can be used to improve transaction and audit processes. For recurring problems, the system will recommend revising the participant agreement or optimizing the revenue calculation model.

[0066] The design of the audit module should also consider data traceability and transparency to ensure the integrity and security of all information. All audit information and rectification measures will be recorded on the blockchain to maintain its immutability.

[0067] The implementation process of step S5 is as follows: First, establish an audit module to receive the revenue distribution results and transaction records. Subsequently, verify the validity of the transaction records and the consistency of the revenue through an audit algorithm. Finally, generate an audit report and record abnormal situations and rectification measures.

[0068] This implementation method combines the outputs of steps S1, S2, S3, and S4, effectively achieving a comprehensive audit and traceability of transaction data. Through a systematic audit mechanism, it can ensure that the rights and interests of participants are effectively protected, and the accuracy and integrity of transaction data are improved. The design of this step not only promotes the transparency and trust of the platform but also helps to improve the efficiency and security of the overall transaction process.

[0069] S6: Finally, optimize the collaboration relationship among all participants through a hypergraph model. The optimization results are returned by the network optimization module and recorded on the blockchain to achieve more efficient processing of subsequent transactions; Specifically, first, establish a data analysis module. The main function of this module is to receive the audit report from step S5 and related transaction data, conduct comprehensive analysis, and generate a visual report. This module can be divided into a data input section, a data analysis section, and a report generation section.

[0070] The data input section first receives the audit report and transaction data, expressed as: ; Wherein: represents the th audit report or transaction data, ; represents the total number of received audit reports and transaction data; represents that the vector belongs to , indicating that all components of this vector are real numbers, and it has components, emphasizing that is a -dimensional real vector.

[0071] After the data input is completed, the data analysis module will perform the following steps to deeply analyze the data and form decision support. First, the module will use statistical analysis methods to evaluate the efficiency of income distribution, calculate the average income and standard deviation to track the fund flow, which can be expressed as: ; Wherein: represents the average value of the participating party's income; represents the th income obtained by the participating party; represents the total number of participating parties.

[0072] Calculate the standard deviation of the income to evaluate the degree of dispersion of the income, which can be expressed as: ; Wherein: represents the standard deviation of the participating party's income, reflecting the consistency of income distribution; represents the th income of the participating party; represents the average value of the participating party's income; represents the total number of participating parties; represents calculating the square of the difference between each income and the average income , and this square term ensures that regardless of the direction of the difference (positive or negative), a non-negative value can be obtained, thus avoiding the cancellation effect; represents summing up the squared deviations of all incomes, and this step integrates the overall dispersion of all transaction parties' incomes from the average value; Indicates the sample correction factor used in the standard deviation calculation, which is applied to the sample data to make the estimate of the standard deviation more accurate. rather than It can better reflect the overall volatility.

[0073] By analyzing the data, the system can also calculate the proportion of abnormal transactions to determine whether there is potential risk. The abnormal transaction computer can be expressed as: ; in: is the proportion of abnormal transactions, expressed as a percentage; Indicates the number of transactions marked as abnormal, from the abnormal records in step S5; Indicates the total number of transactions.

[0074] This analysis step can promptly identify potential issues and provide effective reference for decision-making. Furthermore, the system can provide improvement suggestions based on the abnormal ratio, achieving quantitative management of abnormal transactions.

[0075] After completing the data analysis, the data analysis module will generate a visual report to display the analysis results. The report structure can be expressed as: ; in: Review time indicates the specific time when data analysis was completed; It represents the average value of the participants’ benefits obtained through calculation; represents the standard deviation of participants’ returns obtained through statistical analysis; The results indicate the proportion of abnormal transactions; Visual charts refer to charts that display information such as return distribution and abnormal trading trends; Operational recommendations include suggestions for improving the profit distribution model and risk control measures.

[0076] Visual charts are presented using data graphical tools, including bar charts, pie charts or line charts, to help management quickly understand data analysis results and support intuitive decision-making processes.

[0077] Based on the audit and analysis results, the system will provide decision-making support for management. The audit reports and data are presented in a unified management interface for easy subsequent querying, comparison, and analysis. The system will support filtering, sorting, and real-time updating of the analysis results to enhance the timeliness and accuracy of the output reports.

[0078] If significant deficiencies or data anomalies are found in the revenue distribution model, the system will recommend revising the contract terms of the participants or optimizing the algorithm model and evaluate potential risk factors. The system will establish a feedback mechanism. For the implemented recommendations, changes in subsequent data will be tracked and recorded to verify their effectiveness.

[0079] Ultimately, the setup of this module takes into account data traceability, transparency, and compliance, ensuring that all information is accurately recorded and auditable. All audit information and rectification measures will be recorded on the blockchain to maintain its immutability and provide support for future research and improvement.

[0080] The implementation process of step S6 is as follows: First, establish a data analysis module to receive the audit reports and transaction data from step S5. Subsequently, evaluate the revenue distribution efficiency through data analysis algorithms, identify abnormal transactions, and finally generate a data analysis report and put forward operation suggestions.

[0081] This implementation mode combines the outputs of the foregoing steps S1, S2, S3, S4, and S5, and can effectively achieve in-depth analysis of revenue data and optimized decision-making. The design of this step not only improves the intelligent level of platform data management, increases the scientific nature of operation decisions, but also can promptly discover and correct potential risks, further ensuring the interests of all participants, enhancing the trust and sustainable development capabilities of the platform. Through this step, dynamic optimization of revenue distribution can be achieved, and transaction efficiency can be improved.

[0082] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A media data trading method based on blockchain, characterized in that It includes the following steps: S1: Obtain the original media data, then perform feature extraction on it, and use deep learning algorithms to generate feature representations in the high-dimensional manifold space; S2: Transmit the generated feature representations to the curvature calculation module, calculate the curvature of the manifold, generate the unique fingerprint of the data, and store the fingerprint in the blockchain to ensure the immutability of the data; S3: Obtain the modified media data, perform the same feature extraction as in step S1, and transmit the new feature representations to the geodesic distance calculation module between manifolds to calculate the geodesic distance between manifolds and quantify the degree of data change; S4: Transmit the results of the geodesic distance and the contribution data of the participating parties to the revenue distribution module; S5: After revenue distribution, check the security and data integrity of the transaction according to the double verification mechanism. First, compare the content-addressing ID and the hash value to ensure the physical locking of the data. Second, use zk-SNARKs to verify the access rights and verify the effectiveness of the logical locking; S6: Finally, optimize the cooperation relationship between all participating parties through the hypergraph model. The optimization results are returned by the network optimization module and recorded in the blockchain to achieve more efficient processing of subsequent transactions.

2. The media data trading method based on blockchain according to claim 1, wherein The feature extraction step uses deep learning algorithms, specifically including convolutional neural networks or Mel Frequency Cepstral Coefficient extraction algorithms. The feature vectors are represented as vectors in the high-dimensional feature space: ； Where: Indicating crossing the eigenvalue, is the dimension of the feature vector.

3. A media data trading method based on blockchain according to claim 1, characterized in that The curvature calculation method is based on Riemannian geometry. The Riemannian curvature tensor is used to evaluate the uniqueness and difference of the manifold. The specific calculation formula is: ； Where: is the curvature of the manifold ; is the Riemann curvature tensor, representing the degree of curvature of tangent vectors in a manifold; and is a tangent vector to the manifold ; Denotes the inner product operation; and denote the vector and norms of.

4. A media data trading method based on blockchain according to claim 1, characterized in that, The calculation of the geodesic distance is based on the metric tensor on the manifold to ensure the continuity and traceability of the rheological characteristics in the high-dimensional space. The specific calculation method is: ； Where: is the geodesic distance between and ; A path function for connecting two manifolds, representing a curve between the manifolds; is the metric tensor along the path ; is the tangent vector of the path.

5. A media data trading method based on blockchain according to claim 1, characterized in that, During the dynamic adjustment of the revenue weight, the Lorenz equation is used to describe the interaction relationship between the participating parties. The Lorenz equation is: ； Where: are the different weight values of the participants in the dynamic game; is the strength parameter of the system, indicating the speed of weight change; is a parameter representing the feedback of the system; is the attenuation rate of the system, reflecting the attenuation characteristics of the weight over time; Represents a variable The rate of change with respect to time is proportional to the difference between and That is, when is greater than , the growth rate of is positive, and vice versa. Here, is a positive constant representing the rate of this growth; Represents the variable The rate of change. It consists of two parts: The first part Represents The product of Indicates that Affects The growth over time. When Is less than the constant , The growth rate of Is positive; when Is greater than , Represents Itself is reducing its value, forming a self-regulating effect; Represents a variable The rate of change over time, which consists of two parts: Represents And The product of, indicating that their interaction promotes The increase of. While Is decreasing The value of, Is a positive constant representing the competitive or decay effect.

6. A media data trading method based on blockchain according to claim 1, characterized in that, The double verification mechanism includes comparing the content-addressing ID and the hash value to ensure the effectiveness of the physical locking and logical locking of the data. Among them, the logical locking uses zk-SNARKs to verify the effectiveness of the access rights. The zk-SNARKs generator includes the specific calculation process of the evidence to ensure the persistence and security of the data.

7. A media data trading method based on blockchain according to claim 1, characterized in that The hypergraph model models the multi-party cooperation relationship by defining hyperedges and nodes. The hypergraph Laplacian matrix calculation formula is: ； Where: is the Laplacian matrix of the hypergraph; is the degree matrix, and the diagonal elements thereof represent the degrees of a plurality of nodes; is an adjacency matrix, representing the connection relationship between nodes.

8. A media data trading method based on blockchain according to claim 1, characterized in that, An apparatus for implementing the method according to any one of claims 1 to 7, the apparatus comprising: A feature extraction module for extracting features from the original media data and generating a high-dimensional manifold space representation; A curvature calculation module for calculating the curvature of the manifold and generating a unique fingerprint; A revenue distribution module for automatically calculating and distributing revenues based on the contributions of the participating parties and the rheological characteristics. The revenue distribution module realizes automated intelligent contract management; A verification module for checking the integrity of the transaction through quantum entanglement states and anti-quantum hash algorithms; A network optimization module for optimizing the multi-party cooperation relationship and data exchange process based on the hypergraph model.

9. The method for trading media data based on blockchain according to claim 8, wherein Wherein, The feature extraction module adopts deep learning algorithms to improve the accuracy and efficiency of feature extraction. The algorithms include convolutional neural network and Mel-frequency cepstral coefficient extraction algorithm.

10. A media data trading method based on blockchain according to claim 8, characterized in that, Among them, The revenue distribution module is connected to the blockchain to adjust the revenue weights of participants in real time to ensure that revenue distribution is completed immediately after transactions. The blockchain adopts a decentralized method to ensure the transparency and security of data.

Images