A data tracing analysis system and method

Abstract

The present invention proposes a data tracing and analysis system and method, including: collecting voice data through a user terminal device, obtaining a voice sample after digital processing; the voice recognition server processes the voice sample using artificial intelligence technology, extracts characteristic parameters and semantic content, and generates a digital signature based on the semantics; the parameters and signature are sent to the blockchain network node for consensus verification and record storage; the node stores the parameters and signature in different blocks to form a complete specification; the voice sample is uploaded to the distributed storage network; after the query device initiates a query request, multiple blockchain nodes perform matching based on the request and return characteristic parameters and digital signatures; by comparing the characteristic parameters and signatures, the original voice is obtained from the storage network to achieve secondary verification of the results. The scheme of the present invention uses blockchain and artificial intelligence technology to greatly improve the credibility, scope and efficiency of data tracing and analysis and application.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a data source tracing analysis system and method. Background Art

[0002] In today's digital information age, computer systems often face various network security risks and threats. Some network attackers may arbitrarily delete or tamper with key information records within the system in order to cover up their illegal activities, evade investigation and assume corresponding legal responsibilities. Such practices not only damage the integrity and reliability of the data itself, but also bring great hidden dangers to the overall security and stability of the system. Data security issues have become the basic security issues at present, and data traceability technology is the key technology to achieve data security; but the existing traceability analysis system is not smart enough, the work efficiency is not high, and there are problems with credibility. Summary of the invention

[0003] Based on the above problems, the present invention proposes a data tracing and analysis system and method, which collects voice information through the Internet of Things, relies on the advantages of artificial intelligence technology to achieve in-depth processing of voice content, and uses the tamper-proof nature of blockchain to trace the integrity of voice data, thereby improving the credibility, application scope, intelligence and work efficiency of data tracing and analysis.

[0004] In view of this, one aspect of the present invention proposes a data tracing and analysis system, including: a user terminal device, a speech recognition server, a blockchain network node and a query device; wherein,

[0005] The user terminal device is configured as follows:

[0006] Acquire the voice data to be analyzed, and perform digital sampling conversion to obtain voice digital samples;

[0007] Sending the digital voice sample to the voice recognition server;

[0008] The speech recognition server is configured to: process the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content, and generate a corresponding digital signature according to the semantic content;

[0009] The user terminal device is also configured as:

[0010] Sending the voice feature parameters and the digital signature to a plurality of blockchain network nodes for consensus verification and recording;

[0011] Sending the voice digital sample to a distributed storage network for storage;

[0012] The query device is configured to: send a voice tracing request to a plurality of blockchain network nodes;

[0013] The blockchain network node is configured as follows:

[0014] The voice feature parameters and the digital signature are stored in different blocks to form a complete voice data feature specification;

[0015] Perform a matching query based on the voice tracing request, and return a matching first feature parameter and a first digital signature;

[0016] The query device is further configured to:

[0017] Verify the authenticity of the voice data based on the returned results;

[0018] The original speech sample is obtained from the distributed storage network for verification.

[0019] Another aspect of the present invention provides a data source tracing analysis method, comprising:

[0020] The user end device obtains the voice data to be analyzed and performs digital sampling conversion to obtain voice digital samples;

[0021] The user terminal device sends the digital voice sample to the voice recognition server;

[0022] The speech recognition server processes the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content;

[0023] The speech recognition server generates a corresponding digital signature according to the semantic content;

[0024] The user terminal device sends the voice feature parameters and the digital signature to multiple blockchain network nodes for consensus verification and recording;

[0025] The blockchain network node stores the voice feature parameters and the digital signature in different blocks to form a complete voice data feature specification;

[0026] The user terminal device sends the voice digital sample to a distributed storage network for storage;

[0027] The query device sends a voice tracing request to a plurality of blockchain network nodes;

[0028] The plurality of blockchain network nodes execute a matching query based on the voice tracing request, and return a matching first feature parameter and a first digital signature;

[0029] The query device verifies the authenticity of the voice data according to the returned result;

[0030] The query device obtains the original voice sample from the distributed storage network for verification.

[0031] Optionally, the speech recognition server processes the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content, including:

[0032] The speech recognition server selects a pre-trained speech processing model according to the first attribute data of the speech digital sample;

[0033] The speech recognition server sequentially inputs the digital speech samples into the speech processing model according to speech subframes in the time domain;

[0034] The speech processing model extracts corresponding speech feature parameters according to the speech subframe;

[0035] In combination with the speech feature parameters, key features are extracted and semantic units are identified through an attention mechanism;

[0036] The speech recognition server performs grammatical analysis on the recognition result of the semantic unit to form a natural language description;

[0037] Obtaining semantic content according to the semantic unit and the natural language description;

[0038] The speech recognition server returns the speech feature parameters and the semantic content to the user terminal device.

[0039] Optionally, the step of the speech recognition server generating a corresponding digital signature according to the semantic content includes:

[0040] The speech recognition server analyzes the semantic content to extract first key information;

[0041] Based on the first key information, select a corresponding first digital signature algorithm to perform a measurement description on the content;

[0042] Selecting a corresponding first signature algorithm based on the measurement description result;

[0043] Hash the semantic content and corresponding information using the first signature algorithm to obtain a first signature;

[0044] Encrypting the first signature using an encryption algorithm to generate a digital signature;

[0045] The digital signature is returned to the user terminal device and recorded on the blockchain for use.

[0046] Optionally, the step in which the user terminal device sends the voice feature parameters and the digital signature to multiple blockchain network nodes for consensus verification and recording includes:

[0047] The user terminal device selects a number of trusted nodes from the blockchain network nodes to form a temporary signature verification network;

[0048] The user terminal device packages the voice feature parameters and the digital signature in the form of a transaction and sends them to all signature verification nodes of the temporary signature verification network;

[0049] Each of the signature verification nodes verifies the digital signature according to the received transaction information;

[0050] Each of the signature verification nodes performs multiple rounds of exchanges through the Byzantine fault-tolerant algorithm to achieve a consistent comparison result;

[0051] After the consistency is reached, the transaction information is recorded in the local block of each of the signature verification nodes;

[0052] An enhanced block is formed within the blockchain network to complete the information consensus record.

[0053] Optionally, the blockchain network node stores the voice feature parameters and the digital signature in different blocks respectively to form a complete voice data feature specification, including:

[0054] The blockchain network node extracts the voice feature parameter and the digital signature from the enhanced block;

[0055] Recording the voice feature parameters in a new block A and recording the digital signature in another new block B;

[0056] The two new blocks each contain within themselves a hash value of the enhanced block as a precondition;

[0057] The blockchain network node broadcasts the two new blocks to the entire blockchain network. After other nodes reach a consensus through verification, a voice data feature specification distributed across all nodes in the blockchain network is formed.

[0058] Optionally, the step of the user terminal device sending the voice digital sample to a distributed storage network for storage includes:

[0059] The user terminal device packages the digital voice samples into a standard data format to obtain standard format voice data;

[0060] Comparing the first attribute data with the storage network topology corresponding to the preset different attribute ranges, and selecting a storage network template that meets the preset matching conditions;

[0061] Selecting a group of candidate storage nodes according to the storage network template to form a temporary distributed storage network;

[0062] Sending the standard format voice data to all the candidate storage nodes;

[0063] Each of the candidate storage nodes permanently stores the standard format voice data locally and returns a storage hash value receipt.

[0064] Optionally, the step of sending a voice tracing request to a plurality of blockchain network nodes by the query device includes:

[0065] The query device constructs a query parameter, wherein the query parameter includes information related to the voice sample to be traced;

[0066] The query device selects multiple blockchain network nodes to form a temporary query network;

[0067] The query device broadcasts the query parameters to all query network nodes in the form of a transaction order;

[0068] After receiving the request, each query network node searches for transaction records corresponding to the query parameters from the local block;

[0069] The query network node returns the search result to the query device;

[0070] The query device integrates the search results to determine whether there is a unique match;

[0071] If there is no match, the query device initiates a secondary query to clarify the details of the processing.

[0072] Optionally, the plurality of blockchain network nodes perform a matching query based on the voice tracing request, and return a matching first feature parameter and a first digital signature, comprising:

[0073] The query network node queries the corresponding fields in the feature parameter block and the digital signature block in the blockchain network according to the voice tracing request;

[0074] After the query network node matches the corresponding record, the corresponding first characteristic parameter and first digital signature are extracted;

[0075] Each of the query network nodes returns a matching result to the query device.

[0076] Optionally, the step of the query device acquiring an original voice sample from the distributed storage network for verification includes:

[0077] The query device queries the distributed storage network according to the first characteristic parameter and the first digital signature to locate the storage node storing the original voice sample;

[0078] The query device sends an acquisition request to one or more of the storage nodes, where the request parameter is a unique identifier of the voice sample to be queried;

[0079] The storage node retrieves the corresponding original voice file locally according to the request parameter;

[0080] The storage node returns the original voice file to the query device;

[0081] The query device compares the voice sample to be queried with the original voice file to extract features and compare them for verification.

[0082] According to the technical solution of the present invention, the data tracing and analysis method includes: the user terminal device obtains the voice data to be analyzed, and performs digital sampling conversion to obtain a voice digital sample; the user terminal device sends the voice digital sample to a voice recognition server; the voice recognition server processes the voice digital sample based on a preset voice processing model to obtain voice feature parameters and semantic content; the voice recognition server generates a corresponding digital signature according to the semantic content; the user terminal device sends the voice feature parameters and the digital signature to multiple blockchain network nodes for consensus verification and recording; the blockchain network nodes store the voice feature parameters and the digital signature in different blocks respectively to form a complete voice data feature specification; the user terminal device sends the voice digital sample to a distributed storage network for storage; the query device sends a voice tracing request to multiple blockchain network nodes; multiple blockchain network nodes perform matching queries based on the voice tracing request and return the matching first feature parameters and first digital signature; the query device verifies the authenticity of the voice data according to the returned results; the query device obtains the original voice sample from the distributed storage network for verification. By collecting voice information through the Internet of Things, relying on the advantages of artificial intelligence technology to achieve in-depth processing of voice content, and using the immutability of blockchain to trace the integrity of voice data, the credibility, application scope, intelligence and work efficiency of data traceability analysis are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 is a schematic block diagram of a data tracing and analysis system provided by an embodiment of the present invention;

[0084] Figure 2 It is a flow chart of a data source tracing and analysis method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0085] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0086] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0087] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0088] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0089] Refer to the following Figure 1 to Figure 2 To describe a data tracing analysis system and method provided according to some embodiments of the present invention.

[0090] like Figure 1 As shown, an embodiment of the present invention provides a data tracing and analysis system, including: a user terminal device, a speech recognition server, a blockchain network node and a query device; wherein,

[0091] The user terminal device is configured as follows:

[0092] Acquire the voice data to be analyzed, and perform digital sampling conversion to obtain voice digital samples;

[0093] Sending the digital voice sample to the voice recognition server;

[0094] The speech recognition server is configured to: process the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content, and generate a corresponding digital signature according to the semantic content;

[0095] The user terminal device is also configured as:

[0096] Sending the voice feature parameters and the digital signature to a plurality of blockchain network nodes for consensus verification and recording;

[0097] Sending the voice digital sample to a distributed storage network for storage;

[0098] The query device is configured to: send a voice tracing request to a plurality of blockchain network nodes;

[0099] The blockchain network node is configured as follows:

[0100] The voice feature parameters and the digital signature are stored in different blocks to form a complete voice data feature specification;

[0101] Perform a matching query based on the voice tracing request, and return a matching first feature parameter and a first digital signature;

[0102] The query device is further configured to:

[0103] Verify the authenticity of the voice data based on the returned results;

[0104] The original speech sample is obtained from the distributed storage network for verification.

[0105] It should be known that Figure 1 The block diagram of the data source tracing and analysis system shown is for illustration only, and the number of modules shown therein does not limit the protection scope of the present invention.

[0106] See also Figure 2 Another embodiment of the present invention provides a data tracing analysis method, comprising:

[0107] The user end device obtains the voice data to be analyzed and performs digital sampling conversion to obtain voice digital samples;

[0108] The user terminal device sends the digital voice sample to the voice recognition server;

[0109] The speech recognition server processes the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content (the speech recognition server directly extracts speech feature parameters from the speech sample, and extracts semantic content from the speech sample through semantic recognition);

[0110] The speech recognition server generates a corresponding digital signature according to the semantic content;

[0111] The user terminal device sends the voice feature parameters and the digital signature to multiple blockchain network nodes for consensus verification and recording;

[0112] The blockchain network node stores the voice feature parameters and the digital signature in different blocks to form a complete voice data feature specification;

[0113] The user terminal device sends the voice digital sample to a distributed storage network for storage;

[0114] The query device sends a voice tracing request to a plurality of blockchain network nodes;

[0115] The plurality of blockchain network nodes execute a matching query based on the voice tracing request, and return a matching first feature parameter and a first digital signature;

[0116] The query device verifies the authenticity of the voice data according to the returned result;

[0117] The query device obtains the original voice sample from the distributed storage network for verification.

[0118] The technical solution of this embodiment includes: the user terminal device obtains the voice data to be analyzed, and performs digital sampling conversion to obtain a voice digital sample; the user terminal device sends the voice digital sample to a voice recognition server; the voice recognition server processes the voice digital sample based on a preset voice processing model to obtain voice feature parameters and semantic content; the voice recognition server generates a corresponding digital signature according to the semantic content; the user terminal device sends the voice feature parameters and the digital signature to multiple blockchain network nodes for consensus verification and recording; the blockchain network nodes store the voice feature parameters and the digital signature in different blocks respectively to form a complete voice data feature specification; the user terminal device sends the voice digital sample to a distributed storage network for storage; the query device sends a voice tracing request to multiple blockchain network nodes; multiple blockchain network nodes perform matching queries based on the voice tracing request and return the matching first feature parameters and first digital signature; the query device verifies the authenticity of the voice data according to the returned results; the query device obtains the original voice sample from the distributed storage network for verification. By collecting voice information through the Internet of Things, relying on the advantages of artificial intelligence technology to achieve in-depth processing of voice content, and using the immutability of blockchain to trace the integrity of voice data, the credibility, application scope, intelligence and work efficiency of data traceability analysis are improved.

[0119] In some possible implementations of the present invention, the user terminal device acquires the voice data to be analyzed, performs digital sampling conversion, and obtains the voice digital sample, including:

[0120] The user-end device collects voice signals (i.e., voice data to be analyzed) through hardware such as a recorder;

[0121] The user-end device digitizes the analog voice signal through an A / D converter to obtain a discrete digital voice signal;

[0122] The user-side equipment pre-processes the digital voice signal, such as filtering to remove noise;

[0123] The user end device samples the pre-processed digital voice signal and selects a suitable sampling frequency for sampling conversion;

[0124] The user end device packages the digital sampling result to obtain a voice digital sample file.

[0125] This embodiment digitizes the analog voice signal to facilitate subsequent digital processing and transmission. The discreteness of the voice signal is reduced by sampling, and the processing efficiency is improved. Filtering and noise reduction can improve the quality and recognition effect of the voice signal, and a standardized voice digital sample file is obtained, which lays the foundation for subsequent storage, transmission and processing; the initial processing is simple and efficient, laying a good foundation for subsequent complex algorithm processing. This embodiment converts the analog voice into a standard file through digital sampling, which can standardize the representation of the voice signal and is conducive to subsequent distributed processing.

[0126] In some possible implementations of the present invention, the step of the user terminal device sending the voice digital sample to the voice recognition server includes:

[0127] The user terminal device performs a preliminary analysis on the voice digital sample to extract first attribute data (such as language, duration, volume, etc.) of the voice sample;

[0128] The user terminal device is pre-installed with a speech recognition server resource library (the speech recognition server resource library stores data such as the processing capabilities, speech type expertise, function information, attribute information, and work abnormality rate of multiple servers);

[0129] The user terminal device selects, according to the first attribute data, a plurality of candidate servers having a matching degree exceeding a preset first matching degree threshold in the speech recognition server resource library;

[0130] The user terminal device sends a test request to the candidate server to obtain asynchronous processing time evaluation data of each candidate server;

[0131] The user terminal device evaluates the comprehensive matching degree of each of the candidate servers according to the asynchronous processing time evaluation data, and selects the server with the highest comprehensive matching degree as the best speech recognition server;

[0132] The user terminal device sends the digital voice sample to the voice recognition server.

[0133] This embodiment can accurately match the server expertise according to the characteristics of the voice, and the recognition effect is better. At the same time, considering the asynchronous processing time of the server, the recognition efficiency is higher. By evaluating and avoiding wasting resources by sending to unmatched servers, the task and server are dynamically matched, and the resource capacity is fully utilized, making the subsequent processing chain of recognition more efficient and smooth, and improving the efficiency and cost-effectiveness of voice sample distribution and recognition as a whole.

[0134] In some possible implementations of the present invention, the speech recognition server processes the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content, including:

[0135] The speech recognition server selects a pre-trained speech processing model according to the first attribute data of the speech digital sample;

[0136] The speech recognition server sequentially inputs the digital speech samples into the speech processing model according to speech subframes in the time domain;

[0137] The speech processing model (each hidden layer of the model) extracts corresponding speech feature parameters (such as frequency characteristics, acoustic contour, etc.) according to the speech subframe;

[0138] In combination with the speech feature parameters, key features are extracted and semantic units are identified through an attention mechanism;

[0139] In this embodiment, the speech processing model (deep learning model) includes multiple hidden layers, and each hidden layer will perform feature extraction. In this step, the last hidden layer will combine the feature parameter information extracted by each previous hidden layer (such as low-level features such as frequency and pronunciation), the contextual relationship information after the features are extracted by each hidden layer, and the comprehensive information of the overall perspective of the input speech features, etc., and perform high-level feature mining on these combined information through the attention mechanism, and then identify the semantic unit, that is, the semantic unit is the semantic key information extracted by the last hidden layer from the training set knowledge through the attention mechanism.

[0140] The speech recognition server performs grammatical analysis on the recognition result of the semantic unit to form a natural language description;

[0141] In this step, the natural language description is a semantic content sentence formed by the speech recognition server matching the semantic units through grammatical rules.

[0142] Obtaining semantic content according to the semantic unit and the natural language description;

[0143] In this embodiment, the last hidden layer extracts keywords and key phrases as semantic units from the training set knowledge through the attention mechanism; based on the extracted semantic units, the server combines them through pre-defined grammatical rules to form preliminary natural language sentences (i.e., natural language descriptions); the formed natural language sentences (i.e., natural language descriptions) are deeply semantically parsed to extract semantic roles such as sentence subjects, verbs, and objects; semantic corrections and completions are performed based on the semantic roles and context to eliminate doubts; the resulting combination of natural sentences and semantic roles is the semantic content.

[0144] The speech recognition server returns the speech feature parameters and the semantic content to the user terminal device.

[0145] This embodiment realizes automatic speech feature extraction and semantic understanding based on deep learning. It extracts speech features in detail for each subframe to describe the overall picture of the speech. The semantic recognition accuracy is improved through large model training. The overall execution efficiency is achieved through server deployment. The extracted parameters and content carry all elements of the speech, which is conducive to traceability analysis. It can efficiently and accurately extract speech features and semantic content, providing strong support for subsequent traceability.

[0146] In some possible implementations of the present invention, the step of the speech recognition server generating a corresponding digital signature according to the semantic content includes:

[0147] The speech recognition server analyzes the semantic content and extracts first key information (such as theme, emotion, etc.);

[0148] Based on the first key information, select a corresponding first digital signature algorithm to perform a measurement description on the content;

[0149] In this step, the digital signature algorithm is used to quantitatively analyze the semantic content (i.e., the first key information extracted therefrom), and some measurement values ​​are extracted to reflect the content characteristics (for example: statistics on the frequency of keywords as a content hotspot measurement; measurement of sentence length and structural complexity as content richness; quantification of sentiment polarity as content sentiment preference, etc.). This measurement description has the following functions: using quantitative parameters to fully express the characteristics of various aspects of the semantic content; providing a reference basis for subsequent signature algorithm selection and signature generation; and facilitating feature comparison and authentication in the content matching process.

[0150] Selecting a corresponding first signature algorithm based on the measurement description result;

[0151] In this step, define the application scenario descriptions of multiple digital signature algorithms (such as content length range, amount of information, etc.), classify and count the measurement description results, and obtain the values ​​of various content parameters; compare the application scenario descriptions of each algorithm to find the best matching one or several candidate algorithms; test the candidate algorithms, such as security performance, signature length, etc., to evaluate the pros and cons; comprehensively measure the description requirements and multiple rounds of testing to determine the most suitable algorithm. Specific operations include: if the content has a large amount of information, choose an algorithm that considers more security and reliability; if the content is short, choose an algorithm with a more concise and efficient signature; perform multiple hash tests on the stability of each candidate algorithm; and perform weighted scoring according to advantages and difficulties. The above process can reduce the signature generation cost based on content requirements by matching and selecting algorithm characteristics with content features.

[0152] The semantic content and corresponding information (e.g., key information extracted from the semantic content (e.g., theme, emotion, etc.), the algorithm name or parameters used for the digital signature, the acquisition timestamp of the voice sample, the model name of the voice recognition, the name of the signature generation server and other identification information, etc.) are hashed using the first signature algorithm (e.g., a cryptographic hash function) to obtain a first signature;

[0153] Encrypting the first signature using an encryption algorithm to generate a digital signature;

[0154] The digital signature is returned to the user terminal device and recorded on the blockchain for use.

[0155] The digital signature in this embodiment has the protection function of non-tampering and content integrity. The signature algorithm guarantees the unique correspondence of the digital signature, and the encryption process improves the security and tamper-proof ability of the signature. Digital signatures are more convenient for network transmission, storage and traceability verification, and shared signatures are conducive to cross-system interpretation and verification of semantic content. This embodiment can provide secure and reliable digital signature support for semantic content and improve its application efficiency in a distributed environment.

[0156] In this embodiment, for voice data source tracing analysis, the content of the digital signature may include: semantic type (recognized semantic type, such as intent, emotion, object, etc.), keywords (some representative keywords extracted from semantic understanding), pragmatic relationship (recognized pragmatic relationship between sentences, such as causality, timing, etc.), pronunciation features (some extracted pronunciation features, such as tone trend, pronunciation time, etc.), speaking features (such as speaking speed, pause time, volume and other speaking behavior features), environmental background (environmental background identified by acoustic features, such as indoor and outdoor, traffic, etc.), sampling time (collection timestamp of voice sample), collection device (device type and configuration used for voice collection), recognition model (deep learning model structure for semantic recognition), recognition result (overview of the main content of semantic understanding), voice feature summary (refinement of representative voice features), digital fingerprint (hash value of recognition result), etc.; the acquisition of these contents is based on the recognition, analysis, processing, extraction and other work of the aforementioned language processing model, and can be used as a voice digital signature for subsequent source tracing analysis and matching.

[0157] In some possible implementations of the present invention, the step of the user terminal device sending the voice feature parameters and the digital signature to multiple blockchain network nodes for consensus verification and recording includes:

[0158] The user terminal device selects a number of trusted nodes from the blockchain network nodes to form a temporary signature verification network;

[0159] The user terminal device packages the voice feature parameters and the digital signature in the form of a transaction and sends them to all signature verification nodes of the temporary signature verification network;

[0160] Each of the signature verification nodes verifies the digital signature (such as replay attack verification, etc.) according to the received transaction information (the voice feature parameters and the digital signature);

[0161] Each of the signature verification nodes performs multiple rounds of exchanges through the Byzantine fault-tolerant algorithm to achieve a consistent comparison result;

[0162] In this step, each verification node verifies the received characteristic parameters and digital signatures separately; the verification node broadcasts its own verification results to all other verification nodes; each verification node compares the verification results of the other party and finds inconsistencies; multiple rounds of voting are conducted to reach a consensus through the main number; the consensus result can be fully passed or partially passed and requires further investigation, etc. The "multi-round exchange" here means that each verification node repeatedly broadcasts and compares the results, votes, and finally reaches a consensus through the iterative consensus algorithm.

[0163] After the consistency is reached, the transaction information is recorded in the local block of each of the signature verification nodes;

[0164] An enhanced block is formed within the blockchain network to complete the information consensus record.

[0165] In this embodiment, after multiple rounds of exchanges and comparisons between nodes, each node reaches a consensus result; this consensus result will be recorded in a new block generated locally by each node; compared with traditional blocks, the content of this block is expanded, and feature parameters and signature information are added; this enhanced block completes the decentralized accounting of this batch of traceability information; the main functions of this process include: forming a decentralized ledger record that is authenticated by all nodes and contains special traceability data; compared with individual node records, the information is tamper-proof and credible; it facilitates long-term archiving of traceability information, while allowing distributed query and verification; and provides a reliable digital foundation for subsequent deeper traceability analysis and auditing. Through the generation of enhanced blocks, distributed decentralized secure and reliable information accounting processing is achieved

[0166] This embodiment can improve the immutability and integrity of digital evidence; the consensus mechanism prevents single-point node network anomalies from affecting the traceability results; and achieves decentralized records and trusted operations, which is conducive to voice service transactions and traceability audits. The above implementation uses blockchain technology to record and verify voice traceability digital evidence in real time, improving the overall traceability credibility.

[0167] In some possible implementations of the present invention, the blockchain network node stores the voice feature parameters and the digital signature in different blocks respectively to form a complete voice data feature specification step, including:

[0168] The blockchain network node extracts the voice feature parameter and the digital signature from the enhanced block;

[0169] Recording the voice feature parameters in a new block A and recording the digital signature in another new block B;

[0170] The two new blocks each contain within themselves a hash value of the enhanced block as a precondition;

[0171] The blockchain network node broadcasts the two new blocks to the entire blockchain network. After other nodes reach a consensus through verification, a voice data feature specification distributed across all nodes in the blockchain network is formed.

[0172] In this embodiment, the contents of the voice data feature specification include: feature parameter block (Block A, storing various feature parameters extracted from voice digital samples, such as frequency features, acoustic features, etc.), digital signature block (Block B, storing digital signatures generated for semantic content and voice samples), original enhanced consensus block (as a prerequisite for Block A and Block B, containing original information), connection relationship (Block A and Block B refer to each other's hash in the block to form a connection), metadata (recording metadata such as voice sample collection time, recognition model, signature generation mechanism, etc.), block relationship (the three blocks form a point-to-point relationship network in the above manner), distributed archive (each node together forms a decentralized complete feature specification distributed ledger), etc. Through the organic use and integration of these parts, a "voice data feature specification" digital archive that fully describes the detailed attributes and processing flow of a voice sample is formed; the data in each part has the characteristics of being tamper-proof and collaboratively verified.

[0173] This embodiment efficiently utilizes storage after data diversion, reducing storage costs; each part is associated together through hash values, the overall relationship is clear and reliable, and a structured storage presentation similar to a spreadsheet is realized, which is convenient for subsequent interpretation and query. Distributed archiving strengthens data integrity and non-tamperability, and constitutes a complete digital feature specification containing all traceability data. The above implementation uses blockchain technology to organize and store all feature information of voice samples, providing protection for traceability applications.

[0174] In some possible implementations of the present invention, the step of the user terminal device sending the voice digital sample to a distributed storage network for storage includes:

[0175] The user terminal device packages the digital voice sample into a standard data format (such as IPFS format) to obtain standard format voice data;

[0176] Comparing the first attribute data with the storage network topology corresponding to the preset different attribute ranges, and selecting a storage network template that meets the preset matching conditions;

[0177] Selecting a group of candidate storage nodes according to the storage network template to form a temporary distributed storage network;

[0178] Sending the standard format voice data to all the candidate storage nodes;

[0179] Each of the candidate storage nodes permanently stores the standard format voice data locally and returns a storage hash value receipt.

[0180] In this embodiment, a redundancy and replication mechanism may also be used to store multiple copies of each file to improve reliability; any node in the network may query the actual voice data file through hashing.

[0181] This embodiment makes full use of the attribute characteristics of data and effectively guides it to the appropriate network to improve the efficiency of use; randomly selects nodes to prevent malicious attacks, improves the stability of distributed storage, and realizes decentralized, high-reliability storage and access of voice digital samples; multi-node storage improves data security and reliability, and reduces the risk of single-point loss; adopts point-to-point transmission mode to enhance access efficiency and speed; unifies hash indexing methods to facilitate large-scale voice data management and sharing. Through the distributed network, digital voice samples can be stored permanently and safely to provide support for subsequent semantic processing and traceability.

[0182] In some possible implementations of the present invention, the step of the query device sending a voice tracing request to a plurality of blockchain network nodes includes:

[0183] The query device constructs a query parameter, wherein the query parameter includes information related to the voice sample to be traced;

[0184] The query device selects multiple blockchain network nodes to form a temporary query network;

[0185] The query device broadcasts the query parameters to all query network nodes in the form of a transaction order;

[0186] After receiving the request, each query network node searches for transaction records corresponding to the query parameters from the local block;

[0187] The query network node returns the search result to the query device;

[0188] The query device integrates the search results to determine whether there is a unique match;

[0189] If there is no match, the query device initiates a secondary query to clarify the details of the processing.

[0190] In this embodiment, the voice sample related information includes one or more of the following: voice sample data to be queried, a voice unique identification code (used to uniquely identify the voice sample), voice collection time (recording the specific time point when the voice is collected), feature overview (briefly describing the main features of the voice sample, such as length, format, etc.), collection device information (recording metadata such as the device type model for collecting voice), recognition model information (such as the language model used, voice recognition algorithm, etc.), initial processing node (recording the node organization for the first upload and processing of the voice), related transaction records (important transaction record IDs involved in the original voice), query purpose (indicating the purpose of this query, such as whether it is used for transaction verification, etc.), etc.; these key information can summarize the properties and change trajectories of voice resources throughout the entire processing life cycle, and effectively constitute a parameter system for traceability query.

[0191] In this step, if there is a mismatch in the initial traceability query results, a secondary query can be initiated for a specific node. For example: if the initial query result is not unique, it is necessary to further confirm the feature details of the matching voice sample; if there is a potential verification node processing anomaly, it is necessary to understand the cause of the anomaly; if it is necessary to track deeper details, such as sampling processing parameters. At this time, the query device can send a secondary query request to one or several blockchain nodes involved, requiring the node to provide more matching details or processing logs to support analysis. Through secondary targeted queries, the deficiencies in the initial results can be supplemented to help the device obtain a more refined and definite traceability conclusion.

[0192] In this embodiment, a point-to-point broadcast mode is adopted to realize efficient query initiation, and multiple node conclusions are achieved to achieve no single point of failure, thereby improving the query success rate; the returned results can be used for subsequent analysis to improve the preliminary tracing results; the blockchain network is used to realize a timeout and high-reliability voice tracing query function.

[0193] In some possible implementations of the present invention, the plurality of blockchain network nodes perform a matching query based on the voice tracing request, and return a matching first feature parameter and a first digital signature, including:

[0194] The query network node queries the corresponding fields in the feature parameter block and the digital signature block in the blockchain network according to the voice tracing request;

[0195] After the query network node matches the corresponding record, the corresponding first characteristic parameter and first digital signature are extracted;

[0196] Each of the query network nodes returns a matching result to the query device.

[0197] In this embodiment, the distributed characteristics of blockchain are fully exploited to achieve efficient and accurate query; multi-node participation reduces single-point misjudgment and improves the query positive sample rate; extracting primary features facilitates preliminary verification and simplifies subsequent processing; effectively responds to traceability needs and provides digital support for voice transactions. Distributed traceability matching is jointly performed through the blockchain network to provide reliable and efficient technical services for query requests.

[0198] In some possible implementations of the present invention, the step of the query device verifying the authenticity of the voice data according to the returned result includes:

[0199] The query device compares the returned first characteristic parameter with the locally stored sample for matching and identification;

[0200] Give a preliminary authenticity judgment based on the degree of match;

[0201] The query device regenerates a signature based on the matched characteristic parameters using the same algorithm and compares the signature with the returned first digital signature;

[0202] The secondary recognition result is given through the signature consistency judgment result;

[0203] If in doubt, the querying device may ask the node to provide more features or signature details for multiple rounds of verification;

[0204] At the same time, the query device can send a request to other nodes to verify the consistency of the results;

[0205] Based on multiple rounds of authentication results, a final conclusion on the authenticity of the voice data is given.

[0206] This embodiment conducts comprehensive review from different dimensions to accurately identify the authenticity of voice data.

[0207] In some possible implementations of the present invention, the step of the query device acquiring the original voice sample from the distributed storage network for verification includes:

[0208] The query device queries the distributed storage network according to the first characteristic parameter and the first digital signature to locate the storage node storing the original voice sample;

[0209] The query device sends an acquisition request to one or more of the storage nodes, where the request parameter is a unique identifier of the voice sample to be queried;

[0210] The storage node retrieves the corresponding original voice file locally according to the request parameter;

[0211] The storage node returns the original voice file to the query device;

[0212] The query device compares the voice sample to be queried with the original voice file to extract features and compare them for verification.

[0213] In this embodiment, the powerful distributed storage network can efficiently locate the original voice resources, obtain the original voice samples from the source for a second time for deep matching, and achieve better recognition results. At the same time, it solves the limitation of one-way verification through parameters, makes full use of blockchain and distributed networks, and realizes the complete voice traceability function under the global perspective. Through the second deep collection and verification, the preliminary results can be effectively judged, greatly improving the overall traceability capability.

[0214] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0215] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0216] In the several embodiments provided in the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above-mentioned units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0217] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0218] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0219] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the above-mentioned methods of each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or CD-ROM and other media that can store program codes.

[0220] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, and the memory can include: a flash drive, a read-only memory (English: Read-Only Memory, abbreviated as: ROM), a random access memory (English: Random Access Memory, abbreviated as: RAM), a magnetic disk or an optical disk, etc.

[0221] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

[0222] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions without departing from the spirit and scope of the present invention, and can make various changes and modifications, including the combination of the above-mentioned different functions and implementation steps, including software and hardware implementation methods, all of which are within the scope of protection of the present invention.

Claims

1. A data tracing and analysis system, characterized in that: Applied to the traceability analysis of voice data, including: user terminal equipment, voice recognition server, blockchain network node and query device; among them, The user terminal device is configured as follows: Acquire the voice data to be analyzed, and perform digital sampling conversion to obtain voice digital samples; Sending the digital voice sample to the voice recognition server; The speech recognition server is configured to: process the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content, and generate a corresponding digital signature according to the semantic content; The user terminal device is also configured as: Sending the voice feature parameters and the digital signature to a plurality of blockchain network nodes for consensus verification and recording; Sending the voice digital sample to a distributed storage network for storage; The query device is configured to: send a voice tracing request to a plurality of blockchain network nodes; The blockchain network node is configured as follows: The voice feature parameters and the digital signature are stored in different blocks to form a complete voice data feature specification; Perform a matching query based on the voice tracing request, and return a matching first feature parameter and a first digital signature; The query device is further configured to: Verify the authenticity of the voice data based on the returned results; Acquire voice digital samples from the distributed storage network for verification.

2. A data tracing analysis method, characterized in that: Applied to the source analysis of voice data, including: The user end device obtains the voice data to be analyzed and performs digital sampling conversion to obtain voice digital samples; The user terminal device sends the digital voice sample to the voice recognition server; The speech recognition server processes the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content; The speech recognition server generates a corresponding digital signature according to the semantic content; The user terminal device sends the voice feature parameters and the digital signature to multiple blockchain network nodes for consensus verification and recording; The blockchain network node stores the voice feature parameters and the digital signature in different blocks to form a complete voice data feature specification; The user terminal device sends the voice digital sample to a distributed storage network for storage; The query device sends a voice tracing request to the plurality of blockchain network nodes; The plurality of blockchain network nodes execute a matching query based on the voice tracing request, and return a matching first feature parameter and a first digital signature; The query device verifies the authenticity of the voice data according to the returned result; The query device obtains a voice digital sample from the distributed storage network for verification.

3. The data tracing and analysis method according to claim 2, characterized in that: The speech recognition server processes the speech digital sample based on a preset speech processing model to obtain speech feature parameters and semantic content, including: The speech recognition server selects a pre-trained speech processing model according to the first attribute data of the speech digital sample; The speech recognition server sequentially inputs the speech digital samples into the speech processing model according to speech subframes in the time domain; The speech processing model extracts corresponding speech feature parameters according to the speech subframe; In combination with the speech feature parameters, key features are extracted and semantic units are identified through an attention mechanism; The speech recognition server performs grammatical analysis on the recognition result of the semantic unit to form a natural language description; Obtaining semantic content according to the semantic unit and the natural language description; The speech recognition server returns the speech feature parameters and the semantic content to the user terminal device.

4. The data tracing and analysis method according to claim 3 is characterized in that: The step of the speech recognition server generating a corresponding digital signature according to the semantic content comprises: The speech recognition server analyzes the semantic content to extract first key information; Based on the first key information, select a corresponding first digital signature algorithm to perform a measurement description on the content; Selecting a corresponding first signature algorithm based on the measurement description result; Hash the semantic content and corresponding information using the first signature algorithm to obtain a first signature; Encrypting the first signature using an encryption algorithm to generate a digital signature; The digital signature is returned to the user terminal device and recorded on the blockchain for use.

5. The data tracing and analysis method according to claim 4, characterized in that: The step of the user terminal device sending the voice feature parameters and the digital signature to multiple blockchain network nodes for consensus verification and recording includes: The user terminal device selects a number of trusted nodes from the blockchain network nodes to form a temporary signature verification network; The user terminal device packages the voice feature parameters and the digital signature in the form of a transaction and sends them to all signature verification nodes of the temporary signature verification network; Each of the signature verification nodes verifies the digital signature according to the received transaction information; Each of the signature verification nodes performs multiple rounds of exchanges through the Byzantine fault-tolerant algorithm to achieve a consistent comparison result; After the consistency is reached, the transaction information is recorded in the local block of each of the signature verification nodes; An enhanced block is formed within the blockchain network to complete the information consensus record.

6. The data tracing and analysis method according to claim 5, characterized in that: The blockchain network node stores the voice feature parameters and the digital signature in different blocks respectively to form a complete voice data feature specification, including: The blockchain network node extracts the voice feature parameter and the digital signature from the enhanced block; Recording the voice feature parameters in a new block A and recording the digital signature in another new block B; The two new blocks each contain within themselves a hash value of the enhanced block as a precondition; The blockchain network node broadcasts the two new blocks to the entire blockchain network. After other nodes reach a consensus through verification, a voice data feature specification distributed across all nodes in the blockchain network is formed.

7. The data tracing and analysis method according to claim 6, characterized in that: The step of the user terminal device sending the voice digital sample to the distributed storage network for storage includes: The user terminal device packages the voice digital samples into a standard data format to obtain standard format voice data; Comparing the first attribute data with the storage network topology corresponding to the preset different attribute ranges, and selecting a storage network template that meets the preset matching conditions; Selecting a group of candidate storage nodes according to the storage network template to form a temporary distributed storage network; Sending the standard format voice data to all the candidate storage nodes; Each of the candidate storage nodes permanently stores the standard format voice data locally and returns a storage hash value receipt.

8. The data tracing and analysis method according to claim 7, characterized in that: The step of the query device sending a voice tracing request to a plurality of blockchain network nodes comprises: The query device constructs a query parameter, wherein the query parameter includes information related to the voice sample to be traced; The query device selects multiple blockchain network nodes to form a temporary query network; The query device broadcasts the query parameters to all query network nodes in the form of a transaction order; After receiving the request, each query network node searches for transaction records corresponding to the query parameters from the local block; The query network node returns the search result to the query device; The query device integrates the search results to determine whether there is a unique match; If there is no match, the query device initiates a secondary query to clarify the details of the processing.

9. The data tracing and analysis method according to claim 8, characterized in that: The plurality of blockchain network nodes perform a matching query based on the voice tracing request and return a matching first feature parameter and a first digital signature, comprising: The query network node queries the corresponding fields in the feature parameter block and the digital signature block in the blockchain network according to the voice tracing request; After the query network node matches the corresponding record, the corresponding first characteristic parameter and first digital signature are extracted; Each of the query network nodes returns a matching result to the query device.

10. The data tracing and analysis method according to claim 9, characterized in that: The step of the query device acquiring a voice digital sample from the distributed storage network for verification comprises: The query device queries the distributed storage network according to the first characteristic parameter and the first digital signature to locate the storage node storing the voice digital sample; The query device sends an acquisition request to one or more of the storage nodes, where the request parameter is a unique identifier of the voice sample to be queried; The storage node retrieves the corresponding original voice file locally according to the request parameter; The storage node returns the original voice file to the query device; The query device compares the voice sample to be queried with the original voice file to extract features and compare them for verification.

Images