Method for jewelry transaction public service platform based on 5g fusion application

Abstract

The application relates to the technical field of information indexing and data processing, and discloses a jewelry transaction public service platform method based on 5G fusion application, which comprises the following steps: receiving a recalibration query request of a query party for a verified information pair in a search engine index library, issuing a new dynamic challenge token to a holder of the information pair in response to the request, receiving a 5G challenge response data stream uploaded by the holder and containing a real object image and the new token, dynamically refreshing a context verification state of the information pair in the search engine index library after verifying the data stream, and returning the refreshed state to the query party, the application establishes a closed-loop information updating mechanism triggered by a retrieval behavior, solves the problem that high-value information in the index library is caused to be in a trust state decay due to time elapse, and guarantees the timeliness of information service.

Description

Technical Field

[0001] This invention relates to a method for a public service platform for jewelry transactions based on 5G converged applications, belonging to the field of information indexing and data processing technology. Background Technology

[0002] With technological advancements, the information objects indexed by platforms are becoming increasingly diverse, expanding from existing static information such as authoritative certificates and documents to high-definition, dynamic, real-time video streams enabled by technologies like 5G. However, existing information service methods inherit an inherent limitation in their architecture: platforms only focus on the relevance of information content, making it difficult to technically verify the contextual consistency between different information objects, such as whether static certificate data and dynamic video streams point to the same physical entity. This lack of consistency creates opportunities for fraudulent activities that separate information objects from physical entities, especially in specific application scenarios involving the transfer of high-value assets. Simply introducing 5G or ultra-high-definition video technology cannot solve the aforementioned contextual consistency problem; it essentially only increases the quantity and fidelity of aggregated information without building a trustworthy connection between information objects. This leaves information service platforms in a passive, post-verification state when indexing high-value information, unable to fundamentally ensure the authenticity of the indexed information.

[0003] When addressing the challenges brought by 5G converged applications, existing technologies often focus on the security of data transmission and storage, rather than the authenticity or timeliness of the information itself. This also fails to solve the fundamental problem of contextual consistency. For example, Chinese invention patent CN120471621A discloses a method for data security and privacy protection in jewelry transactions based on 5G converged applications. This solution aims to solve the risk of data leakage in jewelry transactions. It achieves encryption and desensitization protection of transaction data by constructing a complex first asymmetric encryption neural network model, key derivation function KDF, and distributed database. However, the core of this method is to ensure the confidentiality of data during the flow and storage process. Its premise is that the information is authentic and consistent at the source of entry. This solution does not provide any technical means to actively verify whether an information object, digital certificate, and the physical entity it claims to represent, such as the jewelry, are still consistent in context at the current moment. Therefore, it cannot prevent fraudulent behavior of evidence separation. A highly encrypted and authentic certificate information may have already been replaced or transferred to the corresponding physical object, resulting in the information in the index being secure but inaccurate.

[0004] Therefore, the technical problem to be solved by this invention is how to transform existing information services and indexing methods so that they are no longer passive information aggregators, but can be transformed into active verifiers that can proactively create and adjudicate contextual binding relationships between different information objects in a reliable and engineering-simple manner before information pairs are publicly retrieved. Summary of the Invention

[0005] This invention provides a method for a public service platform for jewelry transactions based on 5G converged applications. Its main purpose is to solve the problem of actively creating and adjudicating the contextual binding relationship between different information objects before information indexing.

[0006] To achieve the above objectives, this invention provides a method for a public service platform for jewelry transactions based on 5G converged applications. This method is executed by the server of the public service platform. The server's database stores information pairs associated with jewelry assets that have been bound through contextual association. These information pairs have an initial state of context verification in the platform's search engine index and are publicly searchable. The initial state is generated in response to the verification of a first dynamic challenge token. The method includes:

[0007] Step 101: Receive a context recalibration query request initiated by the querying user based on information in the search engine index.

[0008] Step 102: In response to the context recalibration query request, the server sends a recalibration instruction to the holder user terminal associated with the information pair, and generates and issues a second dynamic challenge token. The second dynamic challenge token has a preset time limit and is different from the first dynamic challenge token corresponding to the information pair when it was generated in the initial state.

[0009] Step 103: Within a preset time limit, receive a new challenge response data stream containing physical images of the jewelry assets and the second dynamic challenge token uploaded by the holder's user terminal via the 5G network;

[0010] Step 104: Verify the new challenge response data stream to determine the validity of the second dynamic challenge token;

[0011] Step 105: Based on the verification results of the new challenge response data stream, dynamically refresh the initial state of the information pair in the search engine index to generate a refreshed context verification state.

[0012] Step 106: The refreshed context verification status is returned to the querying user as the sole response to the querying user's context recalibration query request.

[0013] Preferably, the context-verified initial state is generated through the following steps: receiving the main information object associated with the jewelry asset and placing it in a pending verification state; generating and issuing a first dynamic challenge token to the holder's user terminal; within the validity period of the first dynamic challenge token, receiving a challenge response data stream containing a physical image of the jewelry asset and the first dynamic challenge token uploaded by the holder's user terminal; and, only when the first dynamic challenge token is deemed valid, binding the main information object and the challenge response data stream with context to generate the initial state.

[0014] Preferably, when determining that the first dynamic challenge token is valid, the method further includes: analyzing an accompanying data feature of the challenge response data stream, the accompanying data feature being distinct from the content of the first dynamic challenge token and the content of the physical image of the jewelry asset; identifying whether the challenge response data stream is a real-time collected data stream; wherein the execution of the context association binding is further premised on the identification result being yes.

[0015] Preferably, the accompanying data features include ambient background noise features in the audio track of the challenge response data stream and network transmission jitter features of the challenge response data stream.

[0016] Preferably, the main information object is a composite information object, which includes an asset list listing multiple sub-asset identifiers; the first dynamic challenge token is a session token that remains unchanged within the validity period; the step of receiving a challenge response data stream uploaded by the holder's user terminal, which includes a physical image of the jewelry asset and the first dynamic challenge token, specifically includes: within the validity period, for each sub-asset identifier in the asset list, receiving a challenge response data stream that simultaneously contains a session token and a physical image of the sub-asset; context association binding specifically includes: only when all sub-asset identifiers in the asset list have passed verification, the composite information object and all received challenge response data streams are treated as a whole and context-associated.

[0017] Preferably, the step of determining the validity of the first dynamic challenge token further includes: generating a verification confidence level based on the automated recognition quality of the first dynamic challenge token and the automated recognition quality of the physical image; the step of context association binding to generate an initial state specifically includes: when the verification confidence level is higher than a high threshold, updating the information pair to the search engine index with standard weights; and when the verification confidence level is lower than the high threshold, applying a pending review mark to the information pair and updating it to the search engine index with a reduced weight lower than the standard weights.

[0018] Preferably, the confidence level of verification Determined in the following ways: ,in, The confidence level for the automated identification of the first dynamic challenge token. Confidence level for automated recognition of real-object images, and The preset weighting coefficients, and .

[0019] Preferably, the step of determining the validity of the first dynamic challenge token further includes: extracting the physical feature identifier of the jewelry asset from the challenge response data stream; and determining whether the physical feature identifier is consistent with the feature identifier contained in the main information object.

[0020] Preferably, the main information object is a digital certificate, and the physical feature is a laser engraving on the jewelry asset.

[0021] Preferably, the step of context association binding to generate the initial state specifically includes: applying a context-verified tag to the information pair and updating it together with the information to the platform's search engine index.

[0022] Compared with the prior art, the beneficial effects of the present invention are:

[0023] 1. This method places the main information object in a state of pending verification and uses a time-sensitive dynamic challenge token as a trigger, forcing it to appear together with the physical image in the challenge response data stream. This achieves a one-time binding of the information object and its physical entity in the spatiotemporal context. This process makes context verification a rigid prerequisite for the information pair to be publicly searchable, reconstructing the indexing logic of the information service. This prevents the information index from including context-separated or inaccurate information from the source, ensuring the basic credibility of the information provided by the platform's search service. The content verification and activity identification of the information stream are coupled in a coordinated manner. While verifying the correctness of the information content through the dynamic challenge token, the method further analyzes the accompanying data characteristics of the data stream, such as background noise in the audio or network transmission jitter, to determine whether the data stream is collected in real time. This dual verification mechanism requires that the premise of information indexing not only be consistent with the content, but also that the physical process of the information source is real-time. This effectively prevents attacks such as pre-recorded data injection and ensures the authenticity of the context binding process on which the index release is based.

[0024] 2. The deterministic problem of context verification is transformed into a weighted problem of information retrieval ranking. By generating a verification confidence level based on the quality of automated identification during the verification process, and directly linking this confidence level to the ranking weight in the search engine index, the information service platform can no longer simply reject low-confidence results caused by real-world interference such as video stream jitter and reflections. Instead, it can apply a verification mark and assign a reduced weight to include the result in the index in a more flexible way. This approach ensures the credibility of public search results while also taking into account the completeness of information collection, enriching the control methods of information services. The verification logic of atomic operations is extended to a transactional batch processing method. By using a session token that remains unchanged within a preset time limit and requiring each sub-asset in the asset list to pass the context verification of the session token, the overall binding of a batch of composite information objects is achieved. Only when all sub-asset passes verification is the composite information body released to the index as a single indivisible entry, ensuring the integrity and consistency of batch asset information in the index.

[0025] 3. This method transforms the static verification of information services into a dynamic and maintainable approach. By allowing users who retrieve information pairs to initiate context recalibration query requests, and using these requests as triggers, the information holder is forced to re-execute a complete challenge-response verification process. Based on the results of this re-verification, the platform server dynamically refreshes the verification status of the information pair in the index. This closed-loop information update mechanism triggered by retrieval behavior solves the problem of the decay of the trust status of high-value information in the index over time, ensuring the timeliness of information services. Attached Figure Description

[0026] Figure 1 This is a flowchart of the context recalibration process triggered by the query in this invention;

[0027] Figure 2 This is a graph showing the relationship between the batch processing verification probability and the number of assets in this invention.

[0028] Figure 3 This is a schematic diagram of the public service platform system architecture of the present invention. Detailed Implementation

[0029] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention is further described below in conjunction with specific embodiments. It should be noted that this part is intended to provide further explanation of the invention and is not intended to limit the scope of protection of the invention.

[0030] The present invention provides a method for a public service platform for jewelry transactions based on 5G converged applications. This method is executed by a public service platform server deployed in the cloud. This server manages a public database and a publicly accessible search engine index, and securely communicates with registered holder user terminals (an application for merchants) and query user terminals (an application for buyers or regulators) via 5G or other networks. The complete operation flow of this method includes an initial context verification phase for information pairs, followed by a dynamic, on-demand recalibration phase for maintaining index timeliness. The initial context verification phase addresses the problem of inconsistent static information and physical entity context in this field. When a holder... When a user terminal intends to include their jewelry assets in the platform's search engine index, it uploads a key information object associated with the jewelry asset to the platform server. This key information object can be a digital certificate issued by an authoritative institution, containing a unique physical identifier for the jewelry. A specific example is a laser inscription on the girdle of a diamond, such as GIA12345678. Upon receiving this key information object, the server does not immediately release it into the publicly searchable search engine index. Instead, it places it in a temporary database and marks it as pending verification—this is the pre-indexing suspension. Subsequently, in response to this upload, the server immediately generates a unique, time-limited identifier on the server side. The first dynamic challenge token can be either a dynamically refreshed QR code image or a highly recognizable one-time random string. The preset validity period can be set to 180 seconds, a time setting used to balance user operation time with the risk of token theft and fraudulent attacks. The server then sends the first dynamic challenge token to the holder's user terminal via a secure channel. Upon receiving the first dynamic challenge token, the holder's user terminal must, within the preset validity period of 180 seconds, activate its 5G terminal's camera to capture a high-definition real-time video stream and upload it as a challenge response data stream to the platform server's designated listening port. This challenge response data stream is required to be clear and stable within the same frame. The data stream simultaneously includes a physical image of the jewelry asset and the first dynamic challenge token displayed on the user's terminal screen. After receiving the challenge response data stream, the server initiates a concurrent verification module. One verification step involves the server performing liveness detection on the data stream to prevent pre-recorded video injection attacks. This detection step is achieved by analyzing an accompanying data feature of the data stream, which is distinct from content information such as the token or physical image. In one embodiment, the accompanying data feature is the ambient background noise feature in the audio track. The server analyzes the audio track of the data stream, and if its data is zero or lower than a preset extremely low amplitude, it is determined to be non-real-time collected silent data, and the detection fails.In another implementation, the accompanying data feature is the network transmission jitter characteristic of the data stream. The server analyzes the arrival timestamps of the data packets in the 5G data stream. If its jitter model is too smooth or exhibits regular local file injection characteristics, rather than the random jitter model of real-time public network transmission, the identification fails. Only when the liveness identification result is positive, i.e., it is determined to be a real-time collected data stream, does the server continue with subsequent content verification.

[0031] Before the platform server's activity identification module was put into operation, its key criterion for distinguishing between real-time and non-real-time data was the energy threshold of the environmental background noise. Statistical models of network transmission jitter All of these must be determined through a standardized pre-calibration procedure. This procedure utilizes a test environment covering different 5G terminal models, network conditions, signal-to-noise ratios, and base station loads to collect a benchmark dataset containing at least 10,000 labeled data stream samples. This dataset is clearly distinguished into real-time acquired samples and pre-recorded injected samples. Subsequently, for the audio tracks of all real-time acquired samples in the dataset, the statistical distribution of their signal energy is calculated, and the 5th percentile of this distribution is determined and solidified as the environmental background noise discrimination threshold for the active identification module. Simultaneously, for the Inter-ArrivalTimestamps of data packet arrival timestamp sequences of real-time acquired samples and pre-recorded injected samples, their non-Gaussianity and short-term autocorrelation were extracted as jitter features, and their respective probability density functions (PDFs) were established. These two functions were then solidified into the statistical model of the activity identification module. The module calculates the jitter characteristics of the new data stream during actual runtime. The fit difference is determined using the KS test to determine whether it is a real-time acquired data stream. In content verification, the server uses OCR or QR code decoding to determine whether the token in the video stream is consistent with the first dynamic challenge token issued by the server and is within the validity period. Simultaneously, the server performs extraction and comparison of physical feature identifiers. That is, using OCR technology, it extracts the laser engraving code (GIA12345678) from the jewelry image contained in the video stream and determines whether the extracted physical feature identifier is consistent with the feature identifier contained in the main information object to be verified, i.e., the digital certificate. To address the issue of decreased automatic recognition quality caused by real-world interference such as jitter and reflection in 5G video stream transmission, the server's verification module is designed to output a continuous verification confidence level, rather than a rigid pass / fail binary result. This verification confidence level (VC) can be determined using a weighted formula. In the formula, CT represents the automated recognition confidence level of the first dynamic challenge token, CF represents the physical image (an example of which is a laser-engraved code), and CF represents the automated recognition confidence level of the object. and The preset weighting coefficients, and To illustrate with an example, if the token recognition clarity is average (CT score 0.80), while the laser-engraved code recognition is clear (CF score 0.95), then... Then, the overall verification confidence level .

[0032] The server then employs a hierarchical indexing and dynamic ranking weighting method. The server presets a high threshold of 0.85. When the first dynamic challenge token is deemed valid (validity criteria include successful activity verification, content consistency, and feature matching, and a calculated VC ≥ 0.85), the server performs contextual association binding, establishing a strong association between the main information object and the challenge response data stream in the database, generating an information pair. A context-verified flag is applied to this information pair, and finally, the information pair is updated to the platform's search engine index with standard weights, making it publicly searchable. This generates the initial context-verified state. However, when VC < 0.85, indicating automated verification failure or high suspicion, to avoid false negatives leading to the rejection of legitimate information, the server still performs contextual association binding on the information pair, but applies a pending verification flag to the index and updates it with a weight significantly lower than the standard weight (which can be set to 0.1 or a negative value). This causes the entry to be ranked at the end or not displayed in regular searches, ensuring the credibility of public search results and maintaining the integrity of information assets. In addition to the above, this method also provides a transactional batch processing approach for handling bulk asset transactions such as B2B transactions. In this case, the main information object is a composite information object containing an asset list with multiple sub-asset identifiers, for example, 20 different laser-engraved codes. The first dynamic challenge token issued by the server is defined as a session token that remains unchanged for a preset validity period of 180 seconds. Within this validity period, the server cyclically receives challenge response data streams uploaded by the holder for each sub-asset in the list. Each data stream must simultaneously contain the unchanged session token and a corresponding physical image of a sub-asset. The server maintains a verification status list corresponding to the asset list and makes a transaction decision when the validity period expires: the server only binds the composite information object and all received challenge response data streams as an inseparable whole, as a context association, and releases it to the search engine index if and only if all sub-asset identifiers in the asset list pass verification. If any sub-asset fails verification, the entire transaction is rolled back, and the composite information object remains in a pending verification state and is not indexed.

[0033] After the information pair completes initial context verification and obtains its initial state, thus entering the search engine index, the dynamic on-demand recalibration phase of this method is initiated. This phase addresses the problem of information timeliness decay in information services; that is, an asset may be verified at time T1, but evidence separation may have occurred at time T2, causing the verified state in the index to become a false trust endorsement. To address this, this method establishes a closed-loop information update mechanism triggered by retrieval behavior. When a querying user, who may be a potential buyer, retrieves the information pair with the initial context verification state on the platform, the querying user can initiate a context recalibration query request for that information pair. The platform server... In step 101, the request is received; in step 102, in response to the context recalibration query request, the server does not immediately return the initial state stored in the database, but immediately sends a recalibration instruction to the associated holder user terminal; simultaneously, the server generates and issues a second dynamic challenge token with a preset time limit, which is different in value or form from the first dynamic challenge token corresponding to the initial state, to ensure the real-time nature of this re-verification; in step 103, the server starts listening, waiting for the holder user terminal to upload a new challenge response data stream via the 5G network within a preset time limit, such as 180 seconds. The data stream must include a physical image of the jewelry asset and a second dynamic challenge token. In step 104, the server verifies the new challenge response data stream. This verification process can reuse all or part of the modules from the initial context verification phase, including activity identification, token content verification, and physical feature identification, such as the extraction and comparison of laser-engraved codes. In step 105, based on the re-verification result of step 104, the server dynamically refreshes the initial state of the information pair in the search engine index. If the re-verification is successful, i.e., the activity, token, and features all pass, the server updates the context-verified state of the information pair and refreshes its verification timestamp to the current time. If the re-verification fails, this failure... Failure states include the holder failing to respond within a timeout period or the uploaded data stream failing verification. The server then dynamically modifies the state of the information pair to a negative state, indicating that the context has expired or the holder has not responded to the verification. Finally, in step 106, the server returns the refreshed context verification state generated in step 105, whether it is a verified (new) state or an expired state, as the sole response to the query user's context recalibration query request. In this way, the passive query action is transformed into an active, closed-loop, on-demand triggered information timeliness guarantee mechanism to ensure the dynamic and authentic credibility of high-value information in the index.

[0034] Example 1: A user, in this example a potential buyer, uses the search engine index of a public service platform—an information service—to retrieve a high-priced, specifically numbered jewelry asset. The user finds a matching information pair in the search results. This pair contains the digital certificate of the jewelry asset as its primary information object, along with a context-verified initial state. However, the timestamp of this initial state indicates it was generated 60 days ago. For high-value asset transactions, this time delay leads to a decay in the timeliness of information in the index, making the user unable to be certain that the physical entity corresponding to this information pair has not been the target of fraudulent segregation of evidence within the past 60 days. This is a typical trust challenge faced by internet information services in high-value sectors. To avoid this trust impasse caused by information lag, the querying user does not directly contact the information holder. Instead, they initiate a context recalibration query request for the entry in the search engine index. Upon receiving the request, the platform server does not return the decayed initial state in the database. Instead, it immediately generates a second dynamic challenge token with a validity period of 180 seconds. The token's content can be "Blue Ocean-771-Jupiter." The server then sends a recalibration command and the second dynamic challenge token to the associated holder's user terminal. Upon receiving the command, the holder's user terminal, within the 180-second validity period, takes and uploads a new challenge response data stream containing an image of the jewelry asset and the second dynamic challenge token displaying the words "Blue Ocean-771-Jupiter."

[0035] The platform server verifies the new challenge response data stream. This verification performs liveness detection by analyzing the presence of persistent ambient background noise in the audio track of the 5G data stream and whether its network transmission jitter characteristics conform to the 5G public network real-time transmission model, thus determining that the data stream was collected in real time. Subsequently, the verification module confirms that the token "Blue Ocean-771-Jupiter" in the video stream is consistent with the one issued by the server and is within the time limit. It also extracts the laser markings on the physical image using OCR, and the results are consistent with the feature identifiers stored in the main information object. Based on this successful re-verification, the server dynamically refreshes the context verification status of this information in the search engine index, updating it to "Context Verified" and marking the current timestamp. Finally... The platform server returns this newly refreshed context-verified status with the current timestamp as the sole response to the query user's context recalibration query request. This process transforms the nature of information services from a passive information aggregation and display model to a dynamic, on-demand reverification method triggered by the query user, enforced by the platform, and collaboratively participated in by the holder. Instead of attempting to solve the problem of how to trust the holder, it addresses the core timeliness issue of ensuring the credibility of information in the platform's search engine index at the time of query by reconstructing the information index maintenance mechanism. This makes the credibility of information services no longer a static value that decays over time, but a dynamic value that can be recalibrated in real time.

[0036] Example 2: This example objectively verifies the effectiveness of a 5G-based public service platform method for jewelry transactions in addressing pre-recorded video injection attacks in an information service scenario, as well as the engineering rationality of its hierarchical indexing and dynamic sorting weighting mechanism. The test platform consists of a central server simulating a public service platform server and two 5G test terminals simulating the holder's user terminal and the querying user terminal, respectively. The server and terminals communicate via a standard 5G commercial network channel. All verification modules are deployed on the server side, including an activity identification module, which in this experiment is configured to analyze the background noise characteristics of the audio track, and a verification confidence module for calculating VC, with its weight coefficient set as follows: and The high threshold was set to 0.85. Four test groups were set up to isolate and verify the effects of different technical features: Control Group 1, Control Group 2, Sample Group 1 of this invention, and Sample Group 2 of this invention. Control Group 1 was used to simulate the index logic without the activity identification step in this invention. It tested a fraud attack, which obtained a valid first dynamic challenge token through technical means and pre-recorded a video file containing the token and the corresponding physical image of the jewelry asset. The audio track of the video file was processed to a digital zero-silence state, and then the video file was directly injected as a challenge response data stream and uploaded to the server. Control Group 2 was set up the same as Control Group 1, both using pre-recorded video injection attacks, but the difference was that the activity identification module of this invention was enabled on the server side. Sample Group 1 of this invention was used to test the initial context verification process under good working conditions. A holder user, under good indoor lighting conditions, used a 5G terminal to shoot and upload the challenge response data stream in real time. Sample Group 2 of this invention was used to test the process under non-ideal working conditions. The holder user was in a state of light reflection and slight shaking when shooting, thus simulating common real-world interference in information collection.

[0037] The execution process and data recording of the experiment are shown in Table 1. In control group 1, since no activity identification was performed, the server only verified the video content. Its OCR module recognized the token and laser marking, with a CT score of 0.99 and a CF score of 0.98. The overall verification confidence level was calculated. The value is higher than the preset threshold of 0.85. The server determines that the information is context-verified and releases it to the search engine index with standard weight. In control group 2, the server performs activity identification. Its audio analysis module detects that the audio track of the data stream is in a digital zero-silence state, which is lower than the preset environmental background noise threshold. Therefore, it immediately determines that the data stream is a non-real-time data stream, and the identification fails. The server then refuses to index the data stream. In sample group 1 of this invention, the activity identification module detects continuous environmental background noise that is higher than the threshold and determines that it is a real-time data stream. The content verification module obtains clear recognition results under good working conditions, with a CT score of 0.97 and a CF score of 0.95. Since VC ≥ 0.85, the server determines that the information pair has been verified as context and updates it to the search engine index with standard weights; in sample group 2 of this invention, the activity identification also passes; however, during content verification, due to video stream jitter and reflection interference, the OCR recognition quality decreases, with a CT score of 0.72 and a CF score of 0.81, calculated as follows: Since VC < 0.85, the server executes hierarchical indexing logic based on this result, applies a pending review flag to the information, and updates the index with a reduced weight that is lower than the standard weight.

[0038] Table 1: Test Data Table of Contextual Verification Mechanism Effectiveness

[0039]

[0040] Example 3: This example combines Figures 1 to 3 The methodological description of the public service platform for jewelry transactions based on 5G converged applications, such as... Figure 1 As shown, the process is triggered by the querying user's query action, and specifically includes step 101, receiving the context recalibration query request; step 102, issuing the recalibration command and the second dynamic challenge token; step 103, uploading the challenge response data stream; step 104, verifying the new challenge response data stream. This verification can enable the key modules: content verification and activity identification. Depending on whether the verification passes or fails, the process proceeds to step 105, which is dynamically refreshed to indicate that the context has been verified if successful, or to step 105, which is dynamically refreshed to indicate that the context has expired if it fails. Finally, in step 106, the refreshed context verification status is returned to the querying user.

[0041] like Figure 2 As shown, the horizontal axis represents the batch size and the number of sub-assets, and the vertical axis represents the probability (%). The graph reveals that as the batch size increases, the probability of complete verification decreases, the probability of partial verification increases, while the probability of complete failure remains at an extremely low level. Figure 3 As shown, the architecture includes query user terminals, holder user terminals, and a cloud-based public service platform that communicate via a 5G converged network. The public service platform server within the cloud-based public service platform integrates a concurrent verification module including an activity identification module and a content verification module, a dynamic challenge token generation module, an indexing and sorting module, a transactional batch processing module, and a verification status dynamic refresh module, and is connected to the search engine index library and the public database.

[0042] Example 4: This example discloses a standardized engineering calibration procedure for determining internal verification parameters of an information service platform. This procedure is executed during platform deployment or periodic maintenance and is used to set the verification logic parameters of the search engine index. The initial state of this procedure is defined as follows: technicians have prepared a real-world dataset containing 2000 challenge response data stream samples pre-labeled by technicians. These samples are divided into three categories: 1200 Category A samples: high-confidence valid data with clear images and well-defined features; 500 Category B samples: low-confidence valid data with interference such as jitter and reflection; and 300 Category C samples: invalid data containing faulty tokens or counterfeit items. This procedure calibrates the activity identification threshold. The server analyzes all Category A and Category B samples in the dataset, totaling 1700 challenge response data streams, and extracts their accompanying data features, specifically the average energy of the environmental background noise in the audio track. Simultaneously, the server analyzes a set of 100 known pre-recorded video injection samples. All were below 0.01; the server calculated 1700 valid samples. The statistical distribution of the values ​​is used to obtain their mean. and standard deviation ; Threshold for activity identification pass It was calculated that this threshold was set in the parameter configuration of the activity identification module.

[0043] Subsequently, the procedure calibrates the relevant parameters of the verification confidence score (VC) and clarifies the generation paths of CT and CF: the CT value is determined to be the normalized confidence score in the range of 0.0 to 1.0 directly output by the platform's built-in OCR engine when recognizing dynamic challenge tokens; the CF value is determined to be the normalized confidence score in the range of 0.0 to 1.0 output by a deep learning-based laser code reading model built into the platform when recognizing physical feature identifiers; next, the procedure performs adjustments to the weighting coefficients. And optimization with high thresholds; server with To constrain the process, a grid search is performed, and multiple test cases are conducted. Combinations; for each weight combination, the server uses the formula Calculate the VC value for all 2000 samples in the ground-based real dataset; the server plots the receiver operating characteristic (ROC) curves for valid data (class A and class B) and invalid data (class C), and calculates the area under the curve (AUC); when When the AUC value reached 0.972, the weighting coefficients for that group were set in the platform's verification module; Under fixed conditions, the server further analyzed the VC distribution of 1700 valid data points to distinguish between class A and class B. By analyzing the histogram of this distribution, a distinguishing point was found at VC=0.85, and this value was set as a high threshold. This ensures that 98.8% of the 1200 Class A samples, or 1186 samples, fall into the category of... The interval was assigned a standard weight, ensuring that 95.2% of the 500 Class B samples, or 476 samples, fell within it. The interval is then assigned a checkmark and a reduced weight; The value was therefore determined as the high threshold for the tiered index of the search engine's index library.

[0044] Example 5: A holder user uploads a composite information object for batch transactions to the platform server. This object contains an asset list with 20 sub-asset identifiers, i.e., 20 different diamond laser engravings, to perform initial context verification. In response to this request, the server generates and issues a session token that remains unchanged for a preset validity period of 300 seconds, with the content "Batch-A91-Mars". The holder user terminal must, within the validity period, upload a challenge response data stream via the 5G network for each sub-asset identifier in the asset list. Each data stream contains the unchanged session token and a corresponding physical image of a sub-asset.

[0045] At the deadline, the server's verification module confirmed that it had received 20 challenge response data streams, of which 19 data streams passed activity identification and content verification. However, the physical feature identifier in the video of the challenge response data stream corresponding to the 20th sub-asset identifier was inconsistent with the 20th identifier pre-stored in the asset list after OCR recognition. Since the condition that all sub-asset identifiers in the asset list must pass verification was not met, the server performed a transaction rollback and refused to associate the composite information object with all received challenge response data streams as a whole. The composite information object remained in the state of pending verification in the search engine index and could not be publicly retrieved by the querying user, thus ensuring the integrity and consistency of the batch asset index in the information service.

[0046] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for a public service platform for jewelry transactions based on 5G converged applications, characterized in that, The method is executed by the server of a public service platform. The server's database stores information pairs associated with jewelry assets that have been bound through contextual association. These information pairs have an initial state of context verification in the platform's search engine index and are publicly searchable. The initial state is generated in response to the verification of a first dynamic challenge token. The method includes: Step 101: Receive a context recalibration query request initiated by the querying user based on information in the search engine index. Step 102: In response to the context recalibration query request, the server sends a recalibration instruction to the holder user terminal associated with the information pair, and generates and issues a second dynamic challenge token. The second dynamic challenge token has a preset time limit and is different from the first dynamic challenge token corresponding to the information pair when it was generated in the initial state. Step 103: Within a preset time limit, receive a new challenge response data stream containing physical images of the jewelry assets and the second dynamic challenge token uploaded by the holder's user terminal via the 5G network; Step 104: Verify the new challenge response data stream to determine the validity of the second dynamic challenge token; Step 105: Based on the verification results of the new challenge response data stream, dynamically refresh the initial state of the information pair in the search engine index to generate a refreshed context verification state. Step 106: The refreshed context verification status is returned to the querying user as the sole response to the querying user's context recalibration query request.

2. The method for a public service platform for jewelry transactions based on 5G converged applications according to claim 1, characterized in that, The context-verified initial state is generated through the following steps: receiving the main information object associated with the jewelry asset and placing it in a pending verification state; generating and issuing a first dynamic challenge token to the holder's user terminal; within the validity period of the first dynamic challenge token, receiving a challenge response data stream containing a physical image of the jewelry asset and the first dynamic challenge token uploaded by the holder's user terminal; and, only when the first dynamic challenge token is deemed valid, binding the main information object and the challenge response data stream with context to generate the initial state.

3. The method for a public service platform for jewelry transactions based on 5G converged applications according to claim 2, characterized in that, When determining the validity of the first dynamic challenge token, the method further includes: analyzing an accompanying data feature of the challenge response data stream, the accompanying data feature being distinct from the content of the first dynamic challenge token and the content of the physical image of the jewelry asset; identifying whether the challenge response data stream is a real-time collected data stream; wherein the execution of context association binding is further premised on the identification result being yes.

4. The method for a public service platform for jewelry transactions based on 5G converged applications according to claim 3, characterized in that, The accompanying data characteristics include the ambient background noise characteristics in the audio track of the challenge response data stream and the network transmission jitter characteristics of the challenge response data stream.

5. The method for a public service platform for jewelry transactions based on 5G converged applications according to claim 2, characterized in that, The primary information object is a composite information object, which contains an asset list with multiple sub-asset identifiers; the first dynamic challenge token is a session token that remains unchanged within its validity period. The steps of receiving a challenge response data stream uploaded by the holder's user terminal, which includes a physical image of the jewelry asset and a first dynamic challenge token, specifically include: within the time limit, receiving a challenge response data stream containing both a session token and a physical image of the sub-asset for each sub-asset identifier in the asset list; context association binding specifically includes: only when all sub-asset identifiers in the asset list have passed verification, treating the composite information object as a whole and performing context association binding with all received challenge response data streams.

6. The method for a public service platform for jewelry transactions based on 5G converged applications according to claim 2, characterized in that, The steps for determining the validity of the first dynamic challenge token further include: generating a verification confidence level based on the automated recognition quality of the first dynamic challenge token and the automated recognition quality of the physical image; the steps for context association binding to generate an initial state specifically include: when the verification confidence level is higher than a high threshold, updating the information pair to the search engine index with standard weights; and when the verification confidence level is lower than the high threshold, applying a pending review mark to the information pair and updating it to the search engine index with a reduced weight lower than the standard weights.

7. The method for a public service platform for jewelry transactions based on 5G converged applications according to claim 6, characterized in that, Confidence level verification Determined in the following ways: ,in, The confidence level for the automated identification of the first dynamic challenge token. Confidence level for automated recognition of real-object images, and The preset weighting coefficients, and .

8. The method for a public service platform for jewelry transactions based on 5G converged applications according to claim 2, characterized in that, The steps for determining the validity of the first dynamic challenge token also include: extracting the physical feature identifier of the jewelry asset from the challenge response data stream; and determining whether the physical feature identifier is consistent with the feature identifier contained in the main information object.

9. A method for a public service platform for jewelry transactions based on 5G converged applications according to claim 8, characterized in that, The primary information object is the digital certificate, and the physical identifier is the laser engraving on the jewelry asset.

10. A method for a public service platform for jewelry transactions based on 5G converged applications according to claim 2, characterized in that, The steps for generating an initial state by context association binding specifically include: applying a context-verified tag to the information pair and updating it together with the information to the platform's search engine index.

Images