25 results about "Source tracking" patented technology

The invention relates to a source tracking and forgery detection method and system based on a face semantic mask, and belongs to the technical field of computer vision. The method comprises the following steps: acquiring a to-be-processed image; dividing a face semantic core region and a face semantic non-core region of the to-be-processed image, and generating masks of the face semantic core region and the face semantic non-core region; respectively embedding a detection watermark and a traceability watermark into the human face semantic core area and the human face semantic non-core area through a watermark encoder to obtain an image embedded with the detection watermark and the traceability watermark; processing the image embedded with the traceability watermark and the detection watermark through a noise layer to obtain a noise image; a traceability watermark and a detection watermark are extracted from the noise image through a traceability decoder and a detection decoder respectively, and whether the image is deeply counterfeited or not is judged; and carrying out loss function constraint training. According to the invention, the accuracy of image forgery detection can be improved.

Systems and methods for decentralized data storage and retrieval within a distributed computing system include partitioning datasets into encrypted data shards that are each associated with a unique identifier. A processing system operating within the network may distribute these shards across storage nodes selected based on performance metrics such as latency, storage utilization, and reliability. Metadata for each shard, including its identifier, unique cryptographic key, and node location, may be recorded in a tamper-evident distributed ledger. Upon receiving a retrieval request, the system may query the ledger to identify shard locations, retrieve the shards in parallel, validate their integrity by comparing ledger metadata with metadata from storage nodes, decrypt the shards using their respective cryptographic keys, and reconstruct the dataset. These systems and methods enhance data security, integrity, and fault tolerance while allowing efficient and verifiable decentralized storage and retrieval in dynamic, distributed environments.

This application concerns software-based improvements to computer systems. It relates to an apparatus, method, or program that allows computer systems to manage content generated with artificial intelligence by providing mechanisms for representing and reasoning about the provenance of said content. The application discloses several embodiments in different practical contexts, including change tracking for legal document generation, version control for source code, and real-time collaborative document editing.

The invention discloses a time-varying multi-interference source direct tracking method. The method comprises the following steps: respectively modeling a new target hypothesis and a survival target hypothesis in a prediction stage; in the updating stage, a direct tracking spatial spectrum function is used for replacing a likelihood function in the filter updating step on each prediction target particle set, target-measurement data association is not needed, and information loss from interference source DOA tracking to position tracking is avoided. According to the invention, direct position tracking in a time-varying scene of the number of interference sources can be realized, and the adaptability of interference source tracking is improved.

The invention discloses an intelligent networked automobile spatio-temporal data safety monitoring system and method, and relates to the technical field of spatio-temporal data safety supervision, and the system is characterized in that a vehicle end and road end platform uploads spatio-temporal data processing flow information to an enterprise platform; the enterprise platform reports the spatio-temporal data processing flow information to the attribution spatio-temporal data safety monitoring platform according to the periodic frequency, and forms rectification measures and reports for feedback according to the risk alarm information; the attribution spatio-temporal data security monitoring platform performs security risk identification on spatio-temporal data processing flow information submitted by the enterprise platform, pushes obtained risk alarm information to the enterprise platform, performs traceability tracking on the risk alarm information to obtain a traceability report, and verifies whether rectification measures and reports fed back by the enterprise platform are valid or not; and if yes, stopping pushing the risk alarm information. According to the method and the device, efficient and accurate safety supervision on the spatio-temporal data is realized.

The invention relates to a fault detection method and system for an Internet of Things module, and the method comprises the following steps: carrying out the real-time collection of the operation data of the Internet of Things module, and carrying out the preprocessing of the operation data, and obtaining the operation parameters of the module; performing fluctuation analysis on the module operation parameters to obtain fluctuation operation parameters, and when the fluctuation operation parameters exceed a fault threshold value, performing fault classification on the Internet of Things module based on the fluctuation operation parameters to obtain a fault type judgment result; and carrying out fault source tracking on the Internet of Things module based on the fault type determination result to obtain fault source data, and carrying out comprehensive analysis on the fault source data to obtain a fault detection report, thereby solving the problem that performance degradation or even sudden fault of the module frequently occurs due to temperature change, voltage fluctuation, communication interference or hardware aging factors, and improving the reliability of the module. If the information cannot be found and processed in time, chain reaction is easily caused, and data loss, service interruption and even safety accidents are caused.

The invention belongs to the technical field of carbon emission monitoring, and particularly relates to an industrial park carbon source tracking detection method and system.A data acquisition layer is used for acquiring multi-source data related to carbon emission in an industrial park through various heterogeneous devices; the edge calculation layer is in communication connection with the data acquisition layer and is used for carrying out preprocessing and feature extraction on the multi-source data; the block chain layer is in communication connection with the edge computing layer and is used for carrying out credible storage and verification on the preprocessed data to form credible data which cannot be tampered; the central analysis layer is in communication connection with the block chain layer and is used for carrying out fusion, inversion and trend prediction analysis on the credible data to generate a carbon tracking result; and the application layer is in communication connection with the central analysis layer and is used for visually displaying the carbon tracking result and providing decision support. Through a multi-technology fusion mode, comprehensive, real-time and credible monitoring of carbon emission of the industrial park is realized, and a powerful technical support is provided for carbon management of the park.

The application discloses a radioactive solid waste discharge statistical analysis method and system based on multi-source tracking single data fusion. The method comprises the following steps: S1, providing a tracking single standard template and collecting radioactive solid waste discharge data; S2, settling and summarizing the radioactive solid waste discharge data, and converting different types of radioactive solid waste into a unified estimated prepared volume according to a settlement coefficient based on a rule engine; S3, generating a curve graph and trend analysis according to the settlement and summary result; S4, recording all operation processes; and S5, abnormal early warning and threshold management. The application is not dependent on semantic analysis, adopts a template+rule engine mode, is more stable and simple to maintain, provides a trend graph and prediction function, assists nuclear power plant resource planning, and has strong scalability and supports waste type addition and refinement.

A method includes retrieving a plurality of data from one or more data sources, generating, a contextual NFT based on at least a portion of the plurality of data, executing one or more ephemeral computing operations. The method also includes, as a result of deletion of an ephemeral record created during the ephemeral computing operations, generating a unique reference pointing to an original timestamped source of at least a portion of the plurality of data associated with an original recipient, and performing source tracking to track the unique reference associated with the original recipient. The method also includes determining a proof of achievement score based on results of the one or more hybrid transactions, performing analysis and adjudication using a transaction repository and the at least one smart contract, and determining and output a proof of acknowledgement result based on the analysis and adjudication.

This invention discloses an automated processing system and method for multi-source tracking information in supply chain logistics, relating to the field of supply chain logistics information technology. The system includes: an initial weight assignment module, a data fusion and correction module, a credibility profile update module, and a closed-loop iteration module. The initial weight assignment module is used to acquire all multi-source logistics tracking data sources associated with the target waybill, and based on the inherent attributes and historical accuracy of each data source, generates an initial credibility benchmark value and an initial fusion weight for each data source. The automated processing system and method for multi-source tracking information in supply chain logistics proposed in this invention, through a dynamic credibility assessment and fusion correction linkage mechanism, achieves automated integration and real-time correction of multi-source logistics tracking data, significantly improving information processing efficiency. It can continuously track and evaluate the credibility changes of each data source, effectively solve the problems of conflict and inconsistency between multi-source data, and enhance data collaborative processing capabilities.

The application discloses a false video source tracking method and system based on hash data search and a medium, and relates to the technical field of data processing. The method comprises the following steps: uniformly preprocessing a video dataset to be tracked, and constructing a video sample set with consistent structures; constructing a triple sample set comprising original video samples, corresponding fake video samples and different source comparison video samples based on the video sample set; inputting the triple sample set into a hash feature extraction network to generate video hash codes, and dynamically training a plurality of hash centers based on a hash triple loss function; generating a temporary hash center through a voting mechanism and iteratively optimizing the distribution distance between the hash centers during the training process; extracting a video hash code of a video to be detected, performing hash data search in the hash centers, and determining a target hash center; outputting a source tracking result of the video to be detected based on a real video associated with the target hash center, and realizing stable positioning of a unique source video.

Systems, apparatuses, and methods are described for controlling source tracking and delaying beamforming in a microphone array system. A source tracker may continuously determine a direction of an audio source. A source tracker controller may pause the source tracking of the source tracker if a user may continue to speak to the system. The source tracker controller may resume the source tracking of the source tracker if the user may cease to speak to the system, or when one or more pause durations have been reached.

The invention provides a model fingerprint injection and verification method and equipment based on a bypass network, and belongs to the field of artificial intelligence security, and the method comprises four core stages of main model and bypass network design, trigger sample construction, joint training optimization and fingerprint verification. By introducing the lightweight bypass network, the output of the intermediate layer of the main model is dynamically coupled with the input of the bypass network, and the shared feature representation is realized; constructing a fingerprint trigger sample or a non-fingerprint trigger sample, and outputting a fingerprint text response or a normal text response through joint reasoning in the main model and the bypass network; according to the method, a fingerprint data set and a main model set are constructed, a loss function is established to train a joint reasoning process of a main model and a bypass network, a fingerprint identifier with high specificity and high robustness is injected on the premise of not interfering the performance of the main model, and it is ensured that fingerprints are only effectively activated on a target model; and an innovative solution is provided for model copyright protection, watermark verification and source tracking.

This invention belongs to the field of data mining and recognition technology, and discloses a method, system, and device for constructing a dual-center knowledge graph of domain entities and events. The knowledge graph construction combines entity-centric and event-centric approaches, depicting static information such as entities, entity attributes, and entity relationships in real-world events, and can also express dynamic information such as event attributes and event relationships. A novel four-tuple data structure is designed to enable source tracing of data in the knowledge graph, supporting practical applications such as data access control, privacy protection, and license management. A novel derivative graph computation module is designed to support operations such as aggregation, statistics, association, and transformation of knowledge graph data, while also supporting intelligent computations such as graph embedding and machine learning models. The processed data is stored in a graph storage engine, accelerating knowledge graph querying and retrieval. This invention improves the fine-grained access control, privacy protection, and data management capabilities of knowledge graphs.

The application provides a model fingerprint injection and verification method and device based on a side branch network, and belongs to the field of artificial intelligence security.The method comprises four core stages of main model and side branch network design, trigger sample construction, joint training optimization and fingerprint verification;by introducing a light side branch network, the main model intermediate layer output is dynamically coupled with the side branch network input, and feature representation is shared;fingerprint trigger samples or non-fingerprint trigger samples are constructed, fingerprint text responses or normal text responses are output through joint inference of the main model and the side branch network;fingerprint data sets and main model sets are constructed, and a loss function is established to train the joint inference process of the main model and the side branch network, so that high specificity and strong robustness of the fingerprint identifier are injected without interfering with the performance of the main model, the fingerprint is ensured to be effectively activated only on the target model, and an innovative solution is provided for model copyright protection, watermark verification and source tracking.

This invention discloses an active deepfake detection and robust source tracing method based on region-aware watermarking, belonging to the field of deepfake detection technology. The method includes the following steps: constructing an active deepfake detection framework based on dynamic encrypted watermark generation and region-aware watermarking; training the active deepfake detection framework by generating an encrypted watermark related to image identity information using a dynamic encrypted watermark generator, combining the encrypted watermark with a facial semantic segmentation mask, and inputting it into an encoder to obtain a watermark image; processing the watermark image through a simulated deepfake attack module, and then decoding it through a decoder to recover the global and region-level watermarks; the active deepfake detection framework learns watermark extraction and region-level awareness under different perturbations during the training phase; and inferring the active deepfake detection framework by performing dual verification of identity decryption consistency and global and region-level watermark consistency on the image to be detected. This invention can accurately identify forgery types and provide interpretability.

Techniques are described for performing automated operations related to identifying and using repair and maintenance status information for devices and / or other items or entities, such as summarizing and encoding incoming data, identifying specific encoded information in response to natural language queries, and using the identified repair and maintenance status information in further automated manners in some situations (e.g., to automatically initiate repair or maintenance actions on a particular computing device), along with determining, tracking and providing information about source(s) of data elements stored in the database, and determining, storing and providing information about confidence levels in current accuracy of associated stored data elements. For example, source information tracking may include generating and using a source tracking graph with nodes and edges representing the processing and data flow of incoming data as part of its categorization and encoding into groups of related data items of multiple types.

The invention discloses a maritime target continuous tracking and track association method based on multi-source heterogeneous data, and relates to the technical field of maritime monitoring and intelligent perception, and the method comprises the steps: constructing a multi-source observation model, obtaining multi-source observation data, carrying out the space-time registration of the multi-source observation data through linear interpolation and geographic coordinate transformation, obtaining observation track data, and carrying out the tracking of a target. Feature extraction and fusion are performed on the observation trajectory data to obtain fusion features, a nonlinear regression function is adopted to predict and update the track of the marine target based on the fusion features to obtain a multi-source tracking track, and a multi-detection-source space-time joint association module is constructed to obtain a tracking association track. And introducing an information comparison loss function based on the tracking association track to obtain a target joint optimization track. According to the method, the time evolution and space topology characteristics of the track are fully mined, the multi-target tracking robustness is improved, high-precision track association and continuous tracking capability can still be kept in a complex sea area environment, and the method has good engineering adaptability and expansibility.

The application relates to the technical field of cloud computing software, and particularly discloses a cloud data fusion method and system, which comprises the following steps: acquiring multi-source cloud data records and data source state information, generating a data source context image according to a field list, field completeness, an update time, an interface state and historical conflict records, generating a field trust chain according to the data source context image, constructing a candidate record cluster according to a primary key value, a time window, a field combination and the field trust chain, determining a fusion field value or writing an abnormal queue according to a trust condition, scheduling according to an update frequency, a conflict density and a cloud resource state slice, and updating the data source context image and the field trust chain according to downstream calling exceptions, verification failures or correction records. Through the cooperation of the data source state, the field trust chain, the conflict condition and the feedback record, the application forms a closed relationship among field source tracking, conflict shunting and fusion according to the update.

The invention provides an image super-resolution model watermarking method based on feature space hidden label mapping, and aims to realize copyright identification and source tracking of an image super-resolution model and maintain model performance and data concealment at the same time. The method comprises the following steps: generating a high-resolution watermark image with specific semantic features, and enabling the watermark image to be approximate to an original high-resolution image in a perception level through a feature space alignment technology; meanwhile, lightweight disturbance is injected into the low-resolution image to serve as a triggering mechanism, and the model is guided to output a high-resolution image result with preset semantics in the reasoning stage. Furthermore, through gradient optimization and a sample selection strategy, image samples which have the most influence on model feature migration are screened, so that the robustness and effectiveness of watermark embedding are enhanced. The method is suitable for the field of information security, effectively realizes watermark embedding and subsequent watermark detection or affiliation verification of the model, and has good concealment, robustness and practical value.

Watermarking media content, in combination with blockchain and distributed storage networks, prevents the proliferation of Deepfake content. Digital watermarks are embedded in the audio and video tracks of video clips of trusted content producers at the time the videos are captured or before they are distributed. The watermarks are detected at the social media network's portals, nodes, and back ends. The embedded watermark imparts a unique identifier to the video, that links it to a blockchain. The watermarks also allow video source tracking, integrity verification, and alteration localization. The watermark detectors can be standalone software applications, or they can be integrated with other applications. They are used to perform three main tasks: (1) they alert the Internet user when he watches an inauthentic news video, so that he may discard it, (2) they prevent a Deepfake content from propagating through the network (3) they perform forensic analysis to help track and remove Deepfake content postings.

Watermarking media content, in combination with blockchain and distributed storage networks, prevents the proliferation of Deepfake content. Digital watermarks are embedded in the audio and video tracks of video clips of trusted content producers at the time the videos are captured or before they are distributed. The watermarks are detected at the social media network's portals, nodes, and back ends. The embedded watermark imparts a unique identifier to the video, that links it to a blockchain. The watermarks also allow video source tracking, integrity verification, and alteration localization. The watermark detectors can be standalone software applications, or they can be integrated with other applications. They are used to perform three main tasks: (1) they alert the Internet user when he watches an inauthentic news video, so that he may discard it, (2) they prevent a Deepfake content from propagating through the network (3) they perform forensic analysis to help track and remove Deepfake content postings.

Systems, methods, and computer-readable storage media to track provenance of assets utilizing non-fungible tokens (NFTs). One method includes receiving a provenance request, generating a first non-fungible token (NFT) encapsulated with a first control structure, generating a second NFT encapsulated within a second control structure, and tracking provenance of assets utilizing the first NFT and second NFT.

The application relates to the technical field of intelligent manufacturing equipment industry, and particularly discloses a quality safety intelligent management and control data dynamic processing method and system, which comprises the following steps: collecting quality abnormal event data, constructing a dynamic mapping atlas, identifying the risk diffusion track of the quality abnormal event, and extracting a high-risk correlation node group; identifying whether the high-risk correlation node group accelerates the track diffusion, performing quality scheduling imbalance analysis based on interaction data, obtaining an intelligent quality scheduling bias value, and judging whether an algorithm bias early warning needs to be triggered; embedding a dynamic attenuation function to adjust the intelligent quality scheduling weight of a quality safety intelligent management and control platform, and correcting and intervening in the quality abnormal event in the quality safety intelligent management and control platform; improving the efficiency of quality abnormal event source tracking and high-risk node identification, and timely discovering the risk diffusion imbalance problem caused by excessive correlation of a high-sensitivity quality risk label in an intelligent quality scheduling strategy.

Systems, apparatuses, and methods are described for controlling source tracking and delaying beamforming in a microphone array system. A source tracker may continuously determine a direction of an audio source. A source tracker controller may pause the source tracking of the source tracker if a user may continue to speak to the system. The source tracker controller may resume the source tracking of the source tracker if the user may cease to speak to the system, or when one or more pause durations have been reached.