192 results about "Prediction residual" patented technology

The invention discloses an energy storage power supply residual life prediction method and system based on dynamic weight distribution, and belongs to the technical field of energy storage power supply health management. The method comprises the following steps: acquiring voltage, current and temperature time sequence signals of energy storage power supply operation through a multi-source sensor, and constructing a degradation characteristic sequence by adopting a sliding window method; a degradation inflection point division stage is detected and identified by using a curvature extreme value, a historical degradation mode is matched based on a dynamic time warping algorithm, and an optimal model group is selected; respectively carrying out dynamic weight distribution on the time step length and the feature dimension by adopting a dual-channel attention network, and fusing a learnable coefficient with space-time attention output to generate a life prediction result; residual errors and feature drift are monitored and predicted in combination with an incremental learning mechanism, and model parameters are updated and a degradation knowledge base is expanded by adopting an elastic weight consolidation algorithm. According to the method, the key information capturing capability is enhanced through a space-time attention mechanism, the model adaptability is optimized in combination with incremental learning, and the energy storage power supply life prediction precision and the working condition generalization performance are remarkably improved.

The invention discloses a refrigerated cabinet multi-sensor heterogeneous data fusion method based on artificial intelligence, and the method comprises the following steps: collecting multi-source heterogeneous data in the operation process of a refrigerated cabinet, and constructing a unified event time axis; preprocessing the multi-modal data, and generating a physical enhanced multi-modal tensor in combination with a heat-airflow-electricity model; constructing a disturbance activation vector and a modal mask pattern, and embedding the disturbance activation vector and the modal mask pattern into a multi-modal tensor; inputting the disturbance enhanced input data into an improved Tide model to perform long sequence modeling, and outputting a state prediction sequence; performing prediction residual analysis, constructing a dynamic anomaly scoring function, and fusing a disturbance activation vector and a modal mask graph to generate an anomaly type mark and confidence fusion state representation; sparse gradient uploading, parameter aggregation and structure synchronization are carried out among multiple devices by adopting a federal training framework. According to the method, the multi-modal data fusion precision and the anomaly recognition reliability of the refrigerated cabinet can be improved, and cross-equipment collaborative intelligent optimization is realized.

The invention relates to the technical field of disaster prevention and reduction of a power system, and discloses an icing risk early warning method based on multi-model fusion and residual time sequence characteristic analysis, which comprises the following steps: collecting meteorological data of a line area in real time, removing abnormal values through secondary judgment of a Pauta criterion and a trend, and standardizing; adopting a TEROL algorithm to screen high-weight key features; running SWD-BP, MUL-GRNN and ELM models in parallel, constructing a dynamic weight by combining DSI, an independence weight method and an entropy weight method, and calculating a final meteorological predicted value; generating a prediction residual signal, extracting time domain features such as a mean value and a peak value, and constructing a residual feature matrix through a sliding window; and inputting an LSTM model to process a time sequence dependency relationship, and judging an icing risk level. According to the method, meteorological prediction is optimized through multi-model dynamic fusion, and deviation is analyzed and corrected in combination with residual time sequence characteristics, so that the problem of weak generalization ability of a single model is effectively solved, and the accuracy of icing risk early warning is obviously improved.

The invention relates to the field of biological medicine, and discloses a method and system for analyzing cough sound by using disease characteristics to diagnose respiratory diseases, and the method comprises the steps: deploying a six-microphone annular array to achieve the precise positioning and triggering of a sound source; self-adaptive spectral subtraction and Wiener filtering cascade are adopted to enhance the audio; segmenting a cough segment based on energy envelope; fusing the Mel-cepstrum, the linear prediction residual error, the harmonic energy ratio and the transient zero-crossing rate to construct a pathological feature matrix; extracting local, medium-range and global time sequence features through a three-branch parallel convolutional network; inputting a disease specific classifier to discriminate asthma, pneumonia and laryngitis respectively, and applying a feature decoupling regular term to improve interpretability. The system correspondingly realizes the modularized processing flow. According to the method, the cough sound collection quality and the disease subtype recognition accuracy in a complex environment are improved, meanwhile, the thermodynamic diagram is output to assist clinical decision making, and the diagnosis credibility and practicability are enhanced.

The invention provides a flight path prediction method and system based on predictive coding. The method comprises the following steps: collecting historical flight path data, and performing flight path prediction on the historical flight path data by using a reference prediction module to obtain a reference prediction point; calculating a high-order motion track residual error between the reference prediction point and the real track point by using a residual error encoder, and carrying out one-hot coding on the high-order motion track residual error to establish a residual error coding index; a high-order probability distribution mapping module is used for learning a deep mapping relation between historical flight path data and residual error coding indexes so as to carry out training; inputting a to-be-predicted historical flight track into the reference prediction module and the trained high-order probability distribution mapping module in parallel to respectively obtain a predicted track point and a predicted residual error coding index; decoding the prediction residual error coding index through a residual error decoder to obtain a residual error at a prediction moment; and adjusting the predicted track point based on the residual error at the prediction moment to obtain a final prediction result.

The invention provides an effective wave height real-time prediction method and device based on dynamic residual correction, and relates to the field of ocean engineering. The method comprises the following specific steps: acquiring wave height data and performing multi-dimensional feature screening; constructing an integrated filter fusing L1 trend filtering and variational mode decomposition, optimizing parameters by using a sea image optimization algorithm, introducing a causal sliding window to extract features so as to construct a time sequence input tensor, and inputting the time sequence input tensor into a stacked bidirectional long-short-term memory network based on an attention mechanism after noise addition standardization so as to obtain a basic predicted value; calculating a manifold coherent structure, PID dynamics and physical statistical characteristics, and cascading with the basic prediction characteristics to construct a comprehensive element characteristic vector; a LightGBM architecture is constructed, and a prediction residual error is fitted after optimization is carried out through a sea image optimization algorithm; and finally, executing linear reconstruction based on the dynamic safety threshold constraint, and outputting a real-time correction result. According to the method, the error evolution rule is deeply mined by using manifold geometric features, and the real-time precision and robustness of significant wave height prediction are remarkably improved.

The invention discloses a force control joint dynamic error compensation method based on Kalman filtering, and the method comprises the steps: collecting and preprocessing multi-source signal data, and forming a force control joint input data set; constructing an energy hierarchical Kalman filtering model, executing energy constraint correction and outputting state estimation; calculating a prediction residual error, carrying out time-frequency analysis, adjusting a compensation gain, and generating residual error information; torque and angular velocity signals are extracted, a semantic state is recognized, and semantic gating parameters are output; fusing the state estimation, the residual information and the semantic parameters to generate a dynamic compensation instruction signal; energy layer and residual information changes are monitored, self-calibration is triggered, and a feedback closed loop is formed. According to the invention, by introducing energy hierarchical Kalman filtering, time-frequency modulation compensation and a semantic gating feedback mechanism, dynamic error self-adaptive accurate compensation of the force control joint in a complex multi-disturbance environment is realized.

The invention provides a voice reconstruction method and system based on entropy coding residual quantization and frequency spectrum restoration, and relates to the technical field of artificial intelligence voice signal processing, and the method comprises the steps: obtaining an original voice waveform, inputting the voice waveform into a neural voice coding and decoding model, firstly entering a coder to map the input voice waveform into acoustic potential representation, and then entering a frequency spectrum restoration model; performing residual quantization on the acoustic potential characterization layer by layer through a residual vector quantization module, introducing a gating-based dynamic layer number selection mechanism and entropy regularization constraint, enabling bits to be adaptively distributed among different voice segments, reconstructing reconstructed acoustic features of the potential characterization, inputting the reconstructed acoustic features into a decoder, restoring the reconstructed acoustic features into a time domain waveform, and outputting the time domain waveform. And mapping to a logarithmic magnitude spectrum domain through a spectrum repairing module, predicting a residual error in the logarithmic magnitude spectrum domain and performing confidence gating fusion to obtain a complex spectrum, and outputting after time domain synthesis to obtain reconstructed speech. According to the invention, high fidelity, intelligibility and transmission reliability of the voice can be considered at an extremely low bit rate.

The invention discloses a significant wave height two-stage space-time prediction method and system based on diffusion residual correction, and relates to the technical field of sea wave prediction, historical significant wave height data and corresponding historical wind field data are input into a significant wave height space-time prediction model for processing, and a preliminary significant wave height prediction result is obtained; and inputting the initial prediction error, historical significant wave height data and synchronous forecast wind field data into a residual error correction module based on a diffusion model, and compensating the initial prediction error to obtain a final significant wave height prediction result. A diffusion model based on an EDM framework is innovatively introduced to carry out refined reconstruction on a prediction residual error, and a basic prediction field, a synchronous wind field dynamic factor and a prediction aging code are used as multi-dimensional physical strong conditions to be injected into a diffusion generation process. According to the method, high-frequency texture details and non-linear fluctuation components missed by the first-stage model can be certainly recovered, long-term deviation accumulation is relieved to a certain extent, the high-wave region prediction quality is improved, and the prediction stability is improved.

The invention belongs to the field of real-time monitoring, particularly relates to a chemical process real-time monitoring method and system based on digital twinning, and aims to solve the technical problem that an existing method lacks an online model correction and updating mechanism. The monitoring method comprises the following steps: S1, constructing a state snapshot data set, and synchronously collecting real-time measurement data; s2, identifying a process fluctuation interval, and selecting to-be-selected state snapshots to form an optimal snapshot matrix; s3, calculating a reduced-order truncation error energy ratio, and constructing a reduced-order model based on a reduced-order basis function; s4, in the real-time monitoring stage, if the norm of the prediction residual error is smaller than a monitoring threshold value, a Kalman filtering algorithm is adopted to correct the state coefficient of the reduced-order model; otherwise, calculating the projection error of the prediction residual on each primary function in the standby primary function library; and outputting a full-order state vector reconstructed based on the corrected or updated reduced-order model. The monitoring method provided by the invention ensures the continuous effectiveness and accuracy of monitoring when the chemical process changes.

A 3D data decoding apparatus for decoding encoded data includes a mesh prediction unit that is configured to derive a prediction value of a base mesh vertex position and / or a base mesh attribute from the encoded data and an arithmetic decoder that is configured to arithmetically decode a prediction residual. The arithmetic decoder decodes M first bins of a prefix of a coefficient of the prediction residual by using a context, decodes N first bins of a suffix of a coefficient of the prediction residual by using another context, and adds the prediction value and the prediction residual to derive the base mesh vertex position and / or the base mesh attribute.

The invention provides a time series data prediction deviation correction method and system of a power system. A systematic deviation model is established by extracting a time dimension feature in a prediction residual between reference prediction data of an electric power time sequence and actual electric power time sequence data and combining an exogenous variable feature (time period regularity features such as temperature, irradiation, load, power generation and holiday and festival prompt); the deviation estimate is regarded as a combined effect interpretable by time dimension deviation components (hours and weeks) and exogenous deviation components. And then, adjusting the reference prediction data according to the deviation estimation so as to obtain corrected power time sequence data. According to the technical scheme, through a staged prediction and correction mechanism, while the calculation efficiency is ensured, the problems that systematic deviation always exists in the power time sequence data prediction process, so that the scheduling strategy of a power system is frequently adjusted, and the scheduling efficiency is low are effectively solved. The service life of equipment is shortened; and the power fluctuation of a computing power system is large.

The invention discloses a multivariable time sequence anomaly detection method based on spatial-temporal feature fusion, and the method comprises the steps: outputting time sequence features through employing a series structure of a multi-scale TCN and a lightweight Transform network; constructing an adjacency matrix through a static graph learner and a dynamic graph learner, processing by using a multi-order GCN based on the adjacency matrix, and outputting spatial features; performing weighted fusion on the time sequence features and the spatial features by using a gating mechanism, and outputting a predicted value in combination with multiple residual connection; in a training process, dynamically calculating a sample weight according to a predicted residual error and calculating a weighting loss; and according to a residual error between a predicted value and a real value, generating an abnormal score through a reconstruction error calculation mechanism based on principal component analysis, and comparing a threshold value to complete abnormal detection.

The invention discloses a video coding method and device and a video decoding method and device based on residual prediction, and is applied to the technical field of video processing. The coding method comprises the following steps: determining a motion vector and a disparity vector of a current coding block; determining a first parallax compensation block corresponding to the current coding block in the reference view according to the parallax vector; determining a prediction residual according to the current coding block, the first parallax compensation block, and the motion vector and the parallax vector of the first parallax compensation block; or determining a first time domain motion compensation block corresponding to the current coding block in the previous frame and / or the next frame of image of the current coding block according to the motion vector; determining a prediction residual according to the current coding block, the first time domain motion compensation block, and the motion vector and the disparity vector of the current coding block; and coding the current coding block according to the prediction residual error. According to the method, the accuracy of the obtained prediction residual error can be improved, so that the coding compression efficiency and the coding effect of the multi-view video are improved.

There is provided an information processing apparatus and method to make it possible to reduce a decrease in encoding efficiency. A predictive residual that is a difference between a reflected light intensity that is attribute data of a point cloud representing an object having a three-dimensional shape as a set of points and a predicted value of the reflected light intensity generated by using a reflection model of light on a surface of the object is generated by decoding encoded data of the predictive residual, a coefficient of the reflection model is derived, a predicted value is derived by performing prediction processing by using the reflection model and the coefficient, and the reflected light intensity is generated by adding the predictive residual obtained by decoding the encoded data and the derived predicted value. The present disclosure can be applied to, for example, an information processing apparatus, an electronic device, an image processing method, a program or the like.

A method and apparatus for processing video data are provided. An exemplary method includes: decompressing compressed frames to generate key frames representing a face; generating a first set of parameters for the key frames, the first set of parameters being associated with a three-dimensional (3D) facial representation of the face; reconstructing a second set of parameters for each of one or more inter frames according to a compressed inter prediction residual of the second set of parameters; and generating a video including the face based on the key frames, the first set of parameters, and the second set of parameters.

Relates to improving accuracy of a primary predictive model based on a residual model. A method includes determining, using a primary predictive model, a first set of training prediction residuals based on a first labeled training set and a first set of testing prediction residuals based on a first labeled testing set. A second dataset includes a second labeled training set, labeled based on the first set of training prediction residuals, and a second labeled testing set, labeled based on the first set of testing prediction residuals. The method further includes training the residual model using the second labeled training set. The primary predictive model is adjusted based on a set of training predictions and a set of testing predictions of the residual model to produce more accurate predictions.

A three-dimensional data encoding method encodes a plurality of three-dimensional points, and includes: selecting one of two or more prediction modes for calculating a predicted value of an attribute information item of the first three-dimensional point, in accordance with attribute information items of one or more second three-dimensional points in the vicinity of a first three-dimensional point; calculating the predicted value by the selected prediction mode; calculating, as a prediction residual, a difference between a value of the attribute information item of the first three-dimensional point and the calculated predicted value; and generating a bit stream that includes the one prediction mode and the prediction residual.

The embodiment of the invention provides a video coding method and device and a video decoding method and device, and relates to the technical field of video coding and decoding. Comprising the following steps: performing feature extraction on a current frame, a reconstructed reference frame and at least one reconstructed auxiliary reference frame to obtain a current feature, a reference feature and at least one auxiliary feature; obtaining a motion offset according to the current feature and the reference feature; obtaining a reconstruction motion offset according to the motion offset; obtaining an initial prediction feature according to the reference feature and the reconstruction motion offset; performing fusion enhancement on the initial prediction feature based on the at least one auxiliary feature to obtain a prediction feature; calculating a residual error between the current feature and the predicted feature, and obtaining a predicted residual error; and generating coded data of the current frame according to the motion offset and the predicted residual error. Some embodiments of the invention are used for solving the problem of insufficient utilization of time domain features caused by inter-frame prediction based on a single reference frame.

This invention discloses a deep learning-based method and system for predicting oil product sales trends, comprising: collecting and preprocessing multi-source data to obtain a time-series feature sequence; constructing training inputs and prediction targets by slicing using a sliding window; inputting the training inputs into an improved TSMixer model to obtain sales predictions and the penultimate latent space representation; constructing change point observations based on the differences in latent space representations and prediction residuals between adjacent time steps; performing Bayesian online change point detection calculations on the change point observations to obtain the posterior probability of the change point and the run length distribution; generating dynamic gating signals based on the posterior probability of the change point and scaling and updating the weight matrix of the time mixing layer; setting an asymmetric penalty coefficient based on the run length distribution and performing gradient pruning to update the model parameters; and outputting online sales trend prediction results and warning confidence information. This invention achieves change point perception prediction and updating, supporting inventory replenishment and delivery scheduling.

A displacement decoding unit (205) of a mesh decoding device (200) according to the present invention includes: a bypass arithmetic decoding unit (205A) configured to generate a coefficient level value by performing bypass arithmetic decoding on a displacement bit stream; an inverse quantization unit (205B) configured to generate a first transformed coefficient by performing inverse quantization on the coefficient level value; an adder (205D) configured to generate a second transformed coefficient by adding a prediction transformed coefficient and a prediction residual; an inter prediction unit (205F) configured to generate the prediction transformed coefficient by performing inter prediction by using the second transformed coefficient of a reference frame read from the frame buffer; and a second inverse transform unit (205G) configured to generate a decoded displacement by performing second inverse transform on the second transformed coefficient.

The invention discloses a smart pipeline abnormal data sequence prediction method based on a Transform network. The method comprises the following steps: collecting pipeline data, constructing a time sequence, and dividing a monitoring window; a pipe network topology Laplacian matrix is constructed based on the pipe section connection relation, and a normal reference set is extracted; separating a source signal by using topology constraint SOBI and establishing a source space security domain model; constructing a time sequence fusion Transform prediction network to output abnormal sequence prediction; calculating an abnormal weight according to the deviation between the monitoring window and the normal reference set and between the monitoring window and the source space security domain, and updating source signal features; based on the abnormal weight, the source signal feature and the prediction residual error, constructing an improved SOBI optimization target to jointly optimize a blind source separation parameter and a prediction network parameter; and constructing an event triggering criterion according to the source space security domain default index and the prediction residual error to adaptively update the abnormal weight and the joint model, and outputting a pipeline abnormality prediction result. According to the invention, the reliability of pipeline abnormity prediction is improved.

The application relates to the technical field of industrial information and data processing, and discloses a grey model prediction system for building structure deformation analysis, which comprises the following modules: a data preprocessing module for filtering industrial time series data sequences and storing the sequences in a first state register; an adaptive background value calculation module for determining an adaptive background value calibration weight according to a change gradient difference value, limiting the weight within 0.2-0.8 to generate a mean value background value sequence close to the weight; and a residual dynamic correction prediction module for constructing an observation matrix to solve a coefficient to output a trend prediction sequence, and calculating a deviation compensation value according to a prediction residual sequence to output deformation trend prediction data. The application can follow signal changes through adaptive weights, eliminate calculation lag and overshoot deviation in data mutation intervals, avoid divergence risks in parameter identification, and realize fast prediction response and deterministic threshold closed-loop arbitration on a low-power microprocessor.

The invention discloses a multi-source heterogeneous clock group management system and method based on an ALGOS time scale algorithm. The method comprises the steps that an atomic clock group outputs multiple paths of radio frequency analog signals; the signal frequency discrimination comparison module is used for converting the radio frequency analog signal into a digital signal, measuring the frequency and calculating frequency difference data; the atomic time scale calculation module converts the frequency difference data into frequency difference data relative to the atomic time scale, predicts the clock face frequency reading, determines the weight of each atomic clock according to the forecast residual error, and calculates the atomic time correction through weighted average; the frequency scale control module calculates the frequency offset according to the atomic time correction, the crystal oscillator and the atomic clock signal, and generates control voltage; and the local voltage-controlled crystal oscillator receives the control voltage and reproduces a high-precision time-frequency reference signal synchronized with the atomic clock group. According to the invention, the operation state of the atomic clock is not adjusted, only discrete sampling is carried out, the advantages of the clock group are fused by utilizing the ALGOS algorithm, the long-term stability of the output signal is improved by controlling the high-reliability voltage-controlled crystal oscillator, and the influence of frequency drift, intermittent start and stop and noise interference is effectively inhibited.

Reduce circuit size and improve coding efficiency. [Solution] The encoding device (100) comprises a circuit (160) and a memory (162). The circuit (160) derives a prediction residual that shows the difference between the target block and a prediction image generated using matrix operation type intra prediction, which generates a prediction image by performing matrix operations on the pixel sequences obtained from the left and upper pixel values ​​of the target block. A linear transformation is performed on the prediction residual, a quadratic transformation is performed on the result of the linear transformation, quantization is performed on the result of the quadratic transformation, the result of the quantization is encoded, and a common transformation set is used as the transformation set for the quadratic transformation for multiple prediction modes.

The invention belongs to the technical field of electric digital data processing, and particularly relates to a PDC drill bit torsion impact signal collecting and processing method which comprises the steps that torsion data are segmented into a plurality of torsion data windows; determining the torsion waveform complexity of the torsion data window; determining the rotation speed slippage strength and the formation meshing complexity of the torsion data window; determining an adaptive order of an AR model of the torsion data window; and performing predictive filtering on the torsion data in the torsion data window to obtain a predictive residual sequence of the torsion data window, and extracting a torsion impact signal according to numerical characteristics of the predictive residual sequence. According to the method, the numerical characteristics of the torsion data and the rotating speed data are analyzed, the order of the AR model is self-adapted, the limitation of the AR model with the fixed order is overcome, background noise is filtered out, and the accuracy and robustness of the torsion impact signal are improved.

In various embodiments, an encoder generates a vector of transform coefficients of prediction residues that are associated with a block of source video data. The encoder computes a block step size scaling value based on contextual metadata associated with the transform coefficients. The encoder computes a first quantizer step size based on the block step size scaling value. The encoder computes a second quantizer step size based on the block step size scaling value. The encoder performs trellis coded quantization operations on the vector of transform coefficients using the first quantizer step size and the second quantizer step size to generate a vector of quantization indices. The encoder performs entropy coding operations on the vector of quantization indices to generate an encoded version of the block of source video data.

A method performed by at least one processor in a decoder includes receiving a coded video bitstream that includes a compressed two dimensional (2D) mesh corresponding to a surface of three dimensional (3D) volumetric object. The method includes predicting a sampled 2D coordinate from at least one previously coded sampled 2D coordinate included in the compressed mesh. The method includes deriving a prediction residual associated with the sampled 2D coordinate. The method further includes reconstructing a 2D coordinate corresponding to a boundary vertex based on the predicted sampled 2D coordinate and the derived prediction residual.

A decoder configured to decode a residual level representing a prediction residual from the data stream and sequentially inverse quantize the residual level: inverse quantize the current residual level in accordance with the current transition state to obtain an inverse quantized residual value, the current transition state is updated according to a current residual level characteristic obtained by applying a binary function to the current residual level and according to quantization mode information contained in the data stream. The media signal is reconstructed using the inverse quantized residual value. A current transition state is in accordance with a full-shot mapping of a domain combined from a set of one or more transition states having a current residual level characteristic to a set of one or more transition states in accordance with quantization mode information, wherein a cardinality of the set of one or more transition states differs in accordance with the quantization mode information. A selection of a quantizer is performed by mapping a predetermined number of bits of a current transform state to a default quantizer using a first mapping regardless of quantization mode information.

Methods, system and apparatus for video processing are described. One example method of processing video data includes performing a conversion between a current block of a video and a bitstream of the video according to a rule. The rule specifies that whether a transform skip mode is enabled is determined based on coding information of the current block. The transform skip mode is a coding mode in which a transform is skipped on a prediction residual of a video block.