28 results about "Delta encoding" patented technology

The invention relates to the technical field of industrial big data storage and management, and discloses an industrial data full life cycle management method and system.The method comprises the steps that original time sequence data of industrial equipment is obtained, Shannon entropy calculation is carried out, and a time sequence importance score sequence is obtained; performing threshold comparison according to the time sequence importance score sequence, and determining a key transient window and a stable normal window; performing lossless differential coding on the key transient window to obtain a high-fidelity data segment; performing feature extraction and sparse processing on the stationary normal window, and performing time domain splicing and index construction to obtain a primary hybrid storage structure; and continuously monitoring the query access path to generate a popularity distribution graph to obtain a hierarchical archiving data set. According to the method, value-aware hierarchical storage and dynamic full-life-cycle management of massive industrial data can be realized.

A method of database operations includes receiving a user query, generating a query vector embedding representative of the user query, querying a vector database using the query vector embedding, retrieving a first database vector of the plurality of database vectors based on the query and representative of a first data file corresponding to a first time and belonging to a first time-series data set, receiving a first plurality of delta encodings describing differences between vector representations of temporally-adjacent data files of the first time-series data set, identifying a second data file of the first time-series data set having a second vector representation that differs from the first database vector and corresponds to a second time, and retrieving the second data file from a database.

A decoder receives, from a video bitstream, a first indication of a resolution of a vector difference for a block. Based on the first indication and a position of a magnitude symbol in a suffix portion of a codeword representing the vector difference to be decoded, a probability model is selected from a plurality of probability models for the magnitude symbol. The decoder arithmetically decodes, from the video bitstream and based on the probability model, a second indication indicating whether the magnitude symbol of the vector difference is equal to a corresponding magnitude symbol of a vector difference predictor. A value of the magnitude symbol of the vector difference is determined based on the second indication and a value of the corresponding magnitude symbol of the vector difference predictor. The block is decoded based on the vector difference represented by the codeword having the magnitude symbol with the determined value.

The invention discloses a WebGL-oriented hydrological grid data coding and rendering method and system, and relates to the technical field of water conservancy data visualization, and the method comprises the steps: carrying out the preprocessing of hydrological grid data, and carrying out the blocking according to a preset time length, and obtaining a plurality of time data blocks; performing joint coding of differential coding and run length coding on the water level data in each time data block, and performing polar coordinate quantization coding on the flow rate data in each time data block to obtain coded hydrological grid data; constructing a WebGL rendering pipeline based on pre-calculation and a double-buffering mechanism; and based on a WebGL rendering pipeline fragment shader, carrying out fusion rendering on the encoded hydrological grid data. The technical problems that the real-time interaction requirement cannot be met and the bandwidth is wasted in the prior art are solved.

A method of database operations includes receiving a user query, generating a query vector embedding representative of the user query, querying a vector database using the query vector embedding, retrieving a first database vector of the plurality of database vectors based on the query and representative of a first data file corresponding to a first time and belonging to a first time-series data set, receiving a first plurality of delta encodings describing differences between vector representations of temporally-adjacent data files of the first time-series data set, identifying a second data file of the first time-series data set having a second vector representation that differs from the first database vector and corresponds to a second time, and retrieving the second data file from a database.

A computer-implemented method for optimizing encoding video frames from a video is provided. In an embodiment, the method comprises receiving a video frame to be encoded. The method further comprises using one or more machine learning models to generate an encoding parameter value for encoding the video frame. The method further comprises comparing a first set of delta encoding values, based on the encoding parameter value, representing differences between groups of pixels of the video frame to a second set of delta encoding values, based on an alternative encoding parameter value, representing differences between the groups of pixels of the video frame, and in response to determining that the first set of delta encoding values is less than the second set of delta encoding values, selecting the encoding parameter value. The method further comprises based on the encoding parameter value, encoding the video frame to generate an encoded video frame.

The invention discloses a traffic simulation data adaptive quantitative compression method based on vehicle dynamics constraints, and aims to solve the technical problems of huge vehicle trajectory data volume, high storage resource occupation and low transmission efficiency in a traffic simulation scene. According to the method, efficient compression is carried out on vehicle state data by innovatively combining time sequence differential coding, self-adaptive quantization and a data grading technology. Specifically, according to the method, every 10 frames serve as a compression unit, the first frame records complete vehicle state data to serve as a key frame, and other frames record differential data with the first frame; and the data precision is optimized through a self-adaptive quantization strategy based on a vehicle speed limit standard and dynamic constraints. Experimental verification shows that the compression efficiency can be remarkably improved on the premise that the data reduction precision is guaranteed, and the method is particularly suitable for large-scale traffic simulation and real-time data processing scenes.

Methods and apparatus prepare medical or therapy data, such as respiratory therapy data for electronic communication to a server system, such as a monitoring system (100). The data may be accessed. Low-pass data may be generated by applying an anti-aliasing filter to the data. Low-rate data may be generated by applying a down-sampling filter to the low-pass data. Encoded data may be generated by performing one or more of intermediate floating-point interpolation, simple linear predictive delta encoding, and / or Rice-Golomb encoding on the low-rate data. The encoded data may be packaged as compressed data with a compression header.

Data compaction using codebooks, that applies delta encoding methods to entropy encoding methods to improve data compaction of entropy encoding methods under certain conditions and when compacting data having certain characteristics. Delta encoding may be applied to entropy encoding methods to further compact data sets by reducing the number of sourceblocks included in a codebook to those most commonly encountered in data to be encoded and, where mismatches occur during encoding, using delta encoding of bit differences with existing sourceblocks in the codebook rather than adding new sourceblocks to the codebook.

The application discloses a kind of calling chain data efficient compression and query method, equipment, medium and product, the calling chain data efficient compression method includes: the span list of target calling chain data is obtained, traverses the span list, the calling path in it is matched with the high-frequency path dictionary of pre-set, the path of matching success is replaced as short identifier, combined with the path information of unmatching, generate compressed topological structure data stream;Quantization and difference coding are carried out to numerical data in the span list, to generate numerical data stream;For label type data, label key is mapped to global dictionary identifier, and direct or local dictionary coding is carried out according to label value type, to generate label data stream;The three-way data stream is respectively adaptively entropy coded, and assembled into hierarchical compression block with header index, the technical scheme of the embodiment of the application can significantly improve compression ratio, reduce storage and computing overhead.

A method of recreating vector database data from delta encoding data includes receiving first time-series vector data from a vector database and comprising a first plurality of vector embeddings representative of a first plurality of data files corresponding to a first time point, receiving a first plurality of delta encodings representative of differences between corresponding vector embeddings of the first plurality of vector embeddings and a second plurality of vector embeddings, and recreating the second plurality of vector embeddings based on the first plurality of vector embeddings and the first plurality of delta encodings. The second plurality of vector embeddings is representative of a second plurality of data files corresponding to a second time point each vector embedding of the second plurality of vector embeddings corresponds to one vector embedding of the first plurality of vector embeddings.

The invention provides a lightweight vector map compression and encryption method and system, and the method comprises the steps: extracting a polyline or polygon object from a vector map layer, recognizing geometric key points through employing an SDP algorithm, dividing coordinates into key point and non-key point types, processing the coordinates of the key point and non-key point types through employing a differential processing strategy, and obtaining a compression and encryption result. Wherein the non-key points are subjected to differential processing by adopting a delta coding algorithm so as to eliminate data redundancy, and the key points are processed by adopting a grid mapping method so as to realize coordinate conversion and enhanced protection. The coordinate sequence is divided into a decimal sequence and an integer sequence, the integer sequence is compressed and XOR encryption and scrambling are performed by adopting a DRLE algorithm, and positive integer conversion, scrambling and scrambling are performed on the decimal sequence. And finally, the processed coordinate sequence is stored by adopting an LCS and is compressed by adopting an LZMA2 algorithm, so that an encrypted map after lossless compression is obtained.

A method of vector database compression includes receiving first time-series vector data comprising a first plurality of vectors representative of data corresponding to a first time point, receiving second time-series vector data comprising a second plurality of vectors representative of data corresponding to a second time point, generating a plurality of delta encodings, discarding the second time-series vector data after generating the plurality of delta encodings, and modifying database data to store the plurality of delta encodings. Each vector of the second plurality of vectors corresponds to one vector of the first plurality of vectors. The plurality of delta encodings is generated by, for all corresponding vectors of the first plurality of vectors and the second plurality of vectors, generating one delta encoding that describes differences between values of corresponding elements of the vector of the second plurality of vectors and the corresponding vector of the first plurality of vectors.

Provided are a blockchain-based block compression method and apparatus, a device, and a readable storage medium. The method comprises: acquiring M transaction data packets, each of the transaction data packets comprising transaction attribute values of N transaction attribute types; on the basis of the transaction attribute values and type priorities respectively corresponding to the N transaction attribute types, determining a positional order of the M transaction data packets, so as to obtain a transaction data packet sequence; for the N transaction attribute types, on the basis of the transaction attribute values corresponding to the same transaction attribute types in the transaction data packet sequence, respectively generating attribute value sets according to the positional order, so as to obtain N attribute value sets; respectively performing delta encoding processing on the N attribute value sets to obtain N attribute value delta encodings; and performing block encapsulation processing on the N attribute value delta encodings to obtain a block pending for blockchain uploading.

Provided is a compression method executed by a compression device 1, wherein the following are carried out in the compression method: a step S1 in which a data group observed at a plurality of positions is divided by area, and a plurality of data items included in each area are scanned in a zigzag manner and aligned such that data items at nearby positions are continuous; a step S3 in which a plurality of areas having the same or similar area attributes are grouped together, and a plurality of aligned data strings included in the plurality of grouped areas are integrated; and a step S4 in which the integrated data strings are compressed by XOR delta encoding.

Provided in the present invention is a fractional-order difference analog-to-digital converter for collecting a neural spike potential. Firstly, a fractional-order difference operator is designed to calculate arbitrary fractional-order delta encoding, such that the sparsity of a neural spike potential can be more effectively utilized, thereby significantly improving the compression ratio of the neural spike potential; moreover, the fractional-order difference analog-to-digital converter has higher flexibility by means of tunable differential orders, and can adapt to a wider range of neural spike potential collection scenarios. Secondly, a fractional-order buffer module is designed, such that an arbitrary fractional-order delta encoding value can be generated in a simple and efficient manner. Finally, a serialized bitstream generator is designed, such that serialized packaging can be performed on fractional-order delta data collected by a plurality of parallel channels, which not only exhibits high scalability and is applicable to multi-channel collection, but also enables reliable interfacing with a subsequent synchronous clock wireless transmission system, thus avoiding overheads and latency of synchronous-asynchronous interface coordination, and further improving the efficiency of neural spike potential collection.

The invention relates to the technical field of data processing, and more specifically, the invention relates to an agricultural product supply chain data storage method based on a block chain, and the method comprises the steps: obtaining supply chain logistics distribution data of agricultural products, events between transfer points, and a compartment temperature data sequence; dividing the supply chain logistics distribution data into a plurality of data blocks, and determining adjustment factors of the data blocks according to differences between actual logistics distribution data of the data blocks and corresponding theoretical logistics distribution data; determining the abnormal degree of the data block according to the influence factor difference between the data block and other data blocks; the adjustment degree of the data block is obtained by combining the adjustment factor and the abnormal degree of the data block to obtain the reference value of the data block, and the reference value of the data block is used for compression in differential coding to realize storage of the agricultural product supply chain data, so that the storage efficiency of the agricultural product supply chain data is effectively improved.

The application relates to the technical field of industrial big data storage and management, and discloses an industrial data full-life-cycle management method and system, which comprises the following steps: acquiring original time sequence data of industrial equipment, and performing Shannon entropy calculation to obtain a time sequence importance score sequence; performing threshold comparison according to the time sequence importance score sequence to determine a key transient window and a stable normal window; performing lossless difference coding on the key transient window to obtain a high-fidelity data segment; performing feature extraction and sparsification processing on the stable normal window, and performing time domain splicing and index construction to obtain a primary mixed storage structure; continuously monitoring a query access path to generate a heat distribution map, and obtaining a hierarchical archiving data set. The method can realize value perception layered storage and dynamic full-life-cycle management of massive industrial data.

Provided are a decoder unit, a decompressor unit, and method for a decoder of multiple stages of adders to decode delta encoded data. A decoder unit implemented in a cache memory having a cache memory cell array comprises a first stage of adder circuits to add, in parallel, pairs of encoded items transformed using a delta encoding, wherein the encoded items include a plurality of deltas of neighbors of sequential source items. The decoder unit further comprises at least one successive stage of adder circuits to add, in parallel in each stage, pairs of outputs from a previous stage of adder circuits, wherein each successive stage has fewer adder circuits than the previous stage, and wherein output from a last of the at least one successive stage comprises the sequential source items.

The application discloses a kind of based on light weight deep model's monitoring video compression method, equipment and medium, it is related to video compression technical field, including to input monitoring video sequence, through interframe difference and time sequence difference degree calculation, front and rear frame comparison, determine original key frame set;In set, with local dynamic target detection and enhancement model, generate enhanced key frame data.Based on enhanced key frame data, do difference coding to redundant frame set, according to the difference degree result of enhanced key frame allocation difference coding, obtain different coding frame set.The set is input light weight deep compression model, after parameter quantization, pruning and distillation operation, output compressed video data.Finally, compressed video data and original key frame set mapping comparison, correct target area distortion in compression, generate and store compression correction video.The method greatly reduces model calculation complexity and storage overhead, while maintaining high-precision compression performance, meets the real-time operation demand of edge device.

The invention discloses a monitored video compression method and device based on a lightweight depth model and a medium, and relates to the technical field of video compression, and the method comprises the steps: determining an original key frame set for an input monitored video sequence through inter-frame difference and time sequence difference degree calculation and front and back frame comparison; and generating enhanced key frame data by using a local dynamic target detection and enhancement model in the set. And performing differential coding on the redundant frame set based on the enhanced key frame data, and distributing differential coding according to a difference degree result with the enhanced key frame to obtain a differential coding frame set. And inputting the set into a lightweight deep compression model, and outputting compressed video data through parameter quantization, pruning and distillation operations. And finally, mapping and comparing the compressed video data with the original key frame set, correcting distortion of a target area in compression, and generating and storing a compressed corrected video. According to the method, the model calculation complexity and the storage overhead are greatly reduced, meanwhile, the high-precision compression performance is kept, and the real-time operation requirement of edge equipment is met.

Provided in the present invention is a total variation analog-to-digital converter for collecting an electroencephalogram. Firstly, a total variation operator is used to calculate second-order delta encoding, such that the sparsity of an electroencephalogram signal can be more effectively utilized, thereby significantly improving the compression ratio of the electroencephalogram signal. Secondly, a non-switching double-buffer module is designed, such that a second-order delta encoding value can be generated in a simple and efficient manner. Finally, a serialized bitstream generator is designed, such that serialized packaging can be performed on second-order delta data collected by a plurality of parallel channels, which not only exhibits high scalability and is applicable to multi-channel collection, but also enables reliable interfacing with a subsequent synchronous clock wireless transmission system, thus avoiding overheads and latency of synchronous-asynchronous interface coordination, and further improving the signal collection efficiency.

The invention belongs to the field of artificial intelligence quality tracing, and mainly relates to a traditional Chinese medicine decoction piece production tracing method based on the Internet of Things, which comprises the following steps: deploying different sensors at nodes for different machines, connecting interfaces of the sensors with an MUC interface, integrating acquisition frames, uploading through DMA, preprocessing and analyzing data, performing differential coding operation on a mean value, a variance and a peak value, and obtaining a traditional Chinese medicine decoction piece production tracing result. The method comprises the following steps of: inputting a sliding window, merging the sliding window with fluctuation of the sliding window, outputting a feature vector, merging the feature vector into a sequence for model training, deploying a deep learning framework, training by adopting a prototype calculation function and a Siamese distance measurement model, outputting a model after the training is finished, issuing an edge node, and training and recording abnormal data after the edge node receives the model. And training the abnormal data, automatically adjusting parameters, enhancing monitoring, calculating an abnormal risk and marking potential abnormal data when an abnormal probability is output, solving an absolute value of a feature vector of the potential abnormal data, normalizing gradient weight, generating a fixed-point report, and issuing the fixed-point report to a quality tracing database.

A method of vector database compression includes receiving first time-series vector data comprising a first plurality of vectors representative of data corresponding to a first time point, receiving second time-series vector data comprising a second plurality of vectors representative of data corresponding to a second time point, generating a plurality of delta encodings, discarding the second time-series vector data after generating the plurality of delta encodings, and modifying database data to store the plurality of delta encodings. Each vector of the second plurality of vectors corresponds to one vector of the first plurality of vectors. The plurality of delta encodings is generated by, for all corresponding vectors of the first plurality of vectors and the second plurality of vectors, generating one delta encoding that describes differences between values of corresponding elements of the vector of the second plurality of vectors and the corresponding vector of the first plurality of vectors.

The application provides a cloud computing-based big data privacy calculation method and system, electronic equipment and storage medium, relates to the technical field of big data privacy calculation, and is based on a lightweight federated learning framework. An edge computing gateway collects intelligent home real-time energy consumption big data, dynamically masks device identification, generates encrypted data streams of type labels and energy consumption data, processes compressed encrypted data streams through Delta encoding, obtains device processing parameters through local processing, and forms global energy scheduling reference information through homomorphic encryption aggregation parameters on a cloud computing platform to optimize strategies through reinforcement learning, combines an alarm mechanism to link devices, adjusts operation through TLS / SSL encrypted transmission instructions, protects data privacy, cooperates with edge and cloud computing through a lightweight federated learning framework, combines encryption and optimization technologies, realizes intelligent home energy consumption data privacy protection and efficient scheduling.

The application discloses a large model content generation content traceability watermark method based on content perception injection, steps S1, embedded area block selection; a detail perception area selector component selects an embedded area block with higher stability and reliability by identifying detail features under different resolutions in an image; step S2, carrier and watermark fusion; an invariant domain difference embedding component selects a specified sector of a near-uniform logarithmic polar Fourier domain of the embedded area block, and fuses the carrier and the watermark in a difference encoding form to obtain a watermark-containing image; step S3, watermark block area selection; a block positioning stripping component fills and blocks a high-resolution image to be detected, selects a watermark block area through a positioner, corrects the watermark block area, and strips a sticking block to obtain a final potential watermark block area; step S4, watermark recovery; a watermark difference decoding component recovers the watermark from a corresponding sector of a near-uniform logarithmic polar Fourier domain of the stripped block.

The invention discloses an image compression method based on singular value decomposition and dynamic region segmentation, and provides a self-adaptive compression scheme aiming at the problems of content distortion caused by traditional fixed block compression and neglect of local feature difference in global singular value decomposition. According to the method, singular value decomposition is carried out on an image matrix, the gradient difference degree of row and column feature vectors is calculated, an optimal segmentation point is dynamically screened in combination with a set compression ratio, and high-density subdivision of a texture complex area is achieved. Further constructing a block density regulation and control model, and balancing the compression efficiency and the visual quality through a three-dimensional optimization mechanism of texture sensitive enhancement, boundary protection and large-area compensation; finally, a coordinate-mean value lightweight coding structure is adopted, the storage space and color information are layered, and the data redundancy is remarkably reduced in combination with differential coding and regional topological optimization technologies. The compression performance is effectively improved while image detail features are reserved, and the method is suitable for low-bandwidth scenes such as remote sensing monitoring and medical image transmission.

Provided in the present invention is an integer-order difference analog-to-digital converter for collecting an intracranial neural field potential. Firstly, an integer-order difference operator is designed to calculate arbitrary integer-order delta encoding, such that the sparsity of an intracranial neural field potential can be more effectively utilized, thereby significantly improving the compression ratio of the intracranial neural field potential; moreover, the integer-order difference analog-to-digital converter has higher flexibility by means of tunable differential orders, and can adapt to a wider range of intracranial neural field potential collection scenarios. Secondly, a multi-stage buffer matrix module is designed, such that an arbitrary integer-order delta encoding value can be generated in a simple and efficient manner. Finally, a serialized bitstream generator is designed, such that serialized packaging can be performed on integer-order delta data collected by a plurality of parallel channels, which not only exhibits high scalability, but also enables reliable interfacing with a subsequent synchronous clock wireless transmission system, thus avoiding overheads and latency of synchronous-asynchronous interface coordination, and further improving the signal collection efficiency.