34 results about "Prediction residual" patented technology

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.

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.

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 application discloses an effective wave height double-stage space-time prediction method and system based on diffusion residual correction, relates to the technical field of sea wave prediction, and inputs historical effective wave height data and corresponding historical wind field data into an effective wave height space-time prediction model for processing to obtain a preliminary effective wave height prediction result, and then inputs the preliminary effective wave height prediction result, the historical effective wave height data and synchronous forecast wind field data into a residual correction module based on a diffusion model to compensate for the preliminary prediction error and obtain a final effective wave height prediction result. The diffusion model based on the EDM architecture is innovatively introduced to finely reconstruct the prediction residual, the basic prediction field, the synchronous wind field dynamic factor and the prediction time limit are taken as multi-dimensional physical strong conditions to inject the diffusion generation process, the application can deterministically restore the high-frequency texture details and nonlinear wave components missed by the first-stage model, to a certain extent, the application can relieve the long-term deviation accumulation, improve the prediction quality in high wave areas and improve the prediction stability.

The application discloses a refrigerated cabinet multi-sensor heterogeneous data fusion method based on artificial intelligence, which comprises the following steps: collecting multi-source heterogeneous data in the operation process of the refrigerated cabinet, and constructing a unified event time axis; pre-processing multi-modal data, combining a heat-air flow-electricity model to generate a physically enhanced multi-modal tensor; constructing a disturbance activation vector and a modal mask graph and embedding them into the 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 the disturbance activation vector and the modal mask graph to generate an abnormal type label and a confidence fusion state representation; and using a federal training framework for sparse gradient upload, parameter aggregation and structure synchronization among multiple devices. The application can improve the multi-modal data fusion precision and abnormal recognition reliability of the refrigerated cabinet, and realize cross-device collaborative intelligent optimization.

The application relates to a kind of implementation methods of FPGA-based JPEG-LS encoder, sequentially including context modeling, edge detection, adaptive prediction correction, prediction residual calculation, Golomb coding parameter K calculation, prediction error mapping, Golomb coding is carried out to prediction error MErrval, context parameter update and code stream splicing.This application can realize JPEG-LS lossless high compression rate encoder function on the basis of hardware without the aid of software.One aspect, discard the run length coding in JPEG-LS standard, reduce the complexity of hardware implementation, reduce the consumption of the hardware resource of the designed encoder, the application also obtains the encoder with compression rate of 18%~25% by optimizing the calculation of golomb coding parameter K and golomb coding mode.

This invention discloses a heterogeneous two-order wind power prediction method and system considering physical constraints, belonging to the field of wind power prediction and microgrid energy management technology. The method includes: acquiring and preprocessing raw data; constructing input features including wind speed, wind direction encoding, temperature, humidity, and historical power; dividing the data into training, validation, and test sets; training an XGBoost multi-step direct master prediction model using Bayesian optimization; training an LSTM model based on the prediction residuals and outputting residual prediction values; superimposing the master prediction values ​​and residual prediction values ​​to obtain a two-order prediction result; applying power boundary and ramp rate constraints, and outputting the final prediction result. This invention, by constructing a heterogeneous two-order prediction architecture of XGBoost and LSTM combined with physical constraints, improves prediction accuracy while adapting to small sample scenarios, thus enhancing the engineering usability of the prediction results.

This invention discloses a low-redundancy compressed data transmission method and apparatus for queuing and calling systems in the field of data transmission technology. The method includes: collecting single-service time series and queue waiting time series from multiple transmission endpoints connected to the cloud; extracting time features and inputting an adaptive timing window module to determine the timing window length; constructing endpoint state vectors and calculating state vector difference data; outputting prediction residual data through a lightweight prediction model, compressing the data, and outputting compressed encoded data; and performing redundancy difference analysis on the cloud side based on a preset redundancy level and updating the timing window length based on multiple redundancy differences. This invention solves the technical problem in existing technologies where a unified compression strategy cannot adapt to the differences in real-time waiting time perception between different transmission endpoints, leading to an inability to dynamically balance data redundancy and freshness. It achieves the technical effect of improving data transmission and compression efficiency while minimizing data redundancy.

This invention discloses a distribution network state prediction method based on deviation attribution and data augmentation, comprising: acquiring historical operating data of the distribution network; using a pre-constructed distribution network state prediction model, identifying deviations in the historical operating data based on an error index to obtain predicted deviation data; performing deviation attribution analysis based on the standardized prediction residuals to obtain deviation attribution results; wherein, the prediction residuals are the difference between the actual observed values ​​and the predicted deviation data values; based on the deviation attribution results, selecting a target strategy from a preset dimension augmentation strategy library to augment the dimensions of the real-time operating data of the distribution network, and evaluating the performance of the distribution network state prediction model after dimension augmentation; dynamically adjusting the parameters of the distribution network state prediction model based on the deviation attribution results and the performance evaluation results to achieve adaptive optimization of the distribution network state prediction model, and using the optimized prediction model to predict the distribution network state in real time.

This invention discloses a method and system for predicting pressure anomalies in gas pipelines based on time-series data mining, specifically relating to the field of gas transmission and distribution safety monitoring. This invention collects multi-source time-series pressure data from gas pipelines; performs data cleaning, missing value compensation, and multi-scale normalization preprocessing; constructs a multi-scale feature extraction module integrating wavelet transform and sliding window statistics to extract short-term fluctuation features, long-term trend features, and time-frequency domain features; employs an improved attention mechanism time-series prediction model for pressure sequence prediction; dynamically calculates adaptive thresholds based on prediction residuals, and combines this with continuous sliding window detection to determine anomalies; this invention effectively improves the accuracy of anomaly prediction and early warning capabilities, reduces false alarm rates, and adapts to dynamic changes in pipeline operating conditions.

This invention discloses a prediction residual-driven multi-target tracking method, belonging to the field of computer vision technology. First, target detection is performed on the current frame. Based on the updated model transition matrix and model probabilities from the previous frame, multiple motion models are used to predict existing target trajectories in parallel within a unified state space, and the prediction results are weighted and fused to obtain the target prediction state. Motion constraints are constructed using the motion uncertainty index from the previous frame, and together with geometric overlap and motion direction consistency, they form an association cost, realizing data association between the trajectory prediction state and the detection boxes. Kalman filtering is performed on the successfully associated detection boxes to update the data, and the prediction residual is calculated. Based on the prediction residual, the model transition matrix is ​​adaptively adjusted, and an equivalent prediction residual is constructed to quantify the target motion uncertainty index for data association in subsequent frames. This method improves the robustness and stability of multi-target tracking in complex scenes without relying on appearance features.

The application relates to the technical field of robots, and discloses a fault diagnosis method, an electronic device, a storage medium and a program product. In the method, an electronic device obtains actual state information of N joints of a multi-joint robot at a current time; for any target joint of the N joints, the actual state information of N-1 joints except the target joint is input into a first model to obtain first state prediction information and a predicted fault token of the target joint at the current time; a first prediction residual of the first state prediction information and the actual state information of the target joint is determined; in the case that the first prediction residual is greater than or equal to a first preset residual threshold, sequence features of a prediction residual sequence of the target joint within a first preset time length are determined; and based on the sequence features, the actual state information of the target joint and the predicted fault token, a target fault mode and a target severity of the target joint are obtained. The method can improve the accuracy of fault diagnosis.

A determination is made to use an over-complete transform for decoding a block having N prediction residuals. A primary transform associated with a set of orthonormal primary bases is selected. An additional transform basis is selected. At least N+1 quantized transform coefficients are decoded from a compressed bitstream. The block is then obtained using the at least N+1 quantized transform coefficients, the primary transform, and the additional transform basis.

The application discloses a water quality mutation prediction method and system based on explicit period decoupling and a multi-scale hybrid expert network, comprising the following steps: constructing a multi-source water quality time series dataset and performing batch tensor packaging to obtain a batch start time index; constructing an explicit period memory module and performing period decoupling to separate a pure residual sequence; extracting a low-frequency trend component and a high-frequency mutation component based on the pure residual sequence; constructing an environment perception gated multi-scale hybrid expert network, performing multi-step prediction on the low-frequency trend component and the high-frequency mutation component to obtain corresponding prediction values, and weighting and fusing the low-frequency trend expert prediction values and the high-frequency mutation expert prediction values by using the constructed dynamic gating weight to obtain a prediction residual tensor; and performing result reconstruction and mutation weighting optimization. The application can explicitly decouple the period background and the mutation residual from the physical mechanism, and can perform fine modeling on different frequency characteristics, thereby improving the response speed and the prediction accuracy for water quality emergencies.

The application discloses a new energy power distribution network multi-resolution power prediction method considering source-load space correlation, relates to the power system power prediction technical field, and through adaptive variational modal decomposition based on source-load correlation guidance, decomposes the power signal into high-frequency, medium-frequency and low-frequency components, then constructs a multi-scale dynamic space correlation tensor, adopts a differentiated window to extract correlation characteristics for different frequency components, inputs the extracted correlation characteristics into a layered differentiated prediction model, trains by using a source-load collaborative constraint loss function, adjusts a fusion weight based on an adaptive integration mechanism of a prediction residual feedback, so as to realize power prediction, effectively solves the technical problems that signal decomposition is disconnected with a prediction task, correlation modeling is static and single, an integrated strategy lacks adaptive feedback, and a prediction model lacks physical constraints, and significantly improves the precision, adaptive capacity and robustness of new energy power distribution network power prediction.

This invention belongs to the field of point cloud encoding and decoding technology, and particularly relates to a point cloud hybrid motion prediction method based on point cloud prediction trees. The specific steps include: obtaining geometric prediction parameters and sequence initialization, generating candidate points for reference frames, selecting prediction points, and reconstructing the point cloud. This invention obtains the final predicted position by superimposing the local motion of a reference node relative to its parent node onto coordinates after global motion compensation. This invention scientifically expands the candidate node list, introducing a wider range of predictors to cover various motion types. This expansion breaks the limitations of traditional single predictors, enabling the encoder to have richer matching choices when facing complex local motions, thereby effectively reducing prediction residuals. The method of this invention provides consistent and significant coding gains in various dynamic scenarios, especially in driving scenarios with densely packed independent moving objects.

The encoding device includes: a predictor configured to generate, for each component, a prediction block corresponding to an encoding-target block; a residual generator configured to generate, for each component, a prediction residual representing a difference between the encoding-target block and the prediction block; a mode selector configured to select one mode either an individual encoding mode performing a transform process and a quantization process on a prediction residual of the first component and a prediction residual of the second component for each single component, or a joint encoding mode performing a transform process and a quantization process on a joint prediction residual generated from the prediction residual of the first component and the prediction residual of the second component; a quantization controller configured to determine a quantization matrix to be applied in the quantization process based on the mode selected by the mode selector.

This invention belongs to the field of video compression, and particularly relates to a method and system for compressing virtual human motion streams under low bandwidth. The method includes: acquiring real-time motion stream data of a virtual human guide; real-time monitoring of network transmission bandwidth and terminal decoding performance, generating corresponding state parameters; configuring an initial compression strategy based on the state parameters; extracting motion features to identify first interactive motion segments and regular motion segments, and generating motion parameter prediction residual sequences using a pre-trained temporal semantic prediction model; compressing the two types of segments using corresponding encoding modes according to the initial compression strategy to obtain compressed motion bitstreams; transmitting the bitstream to the terminal and calling a lightweight reconstruction model for decoding, calculating the decoding frame rate and lip-sync error; if a preset threshold is not met, adjusting the compression parameters and returning to reprocessing; otherwise, outputting a stable compression configuration. This invention achieves efficient compression and smooth reconstruction of virtual human motion streams under low bandwidth, ensuring a smooth interactive experience for virtual human guides.

The application relates to a stacked carton fatigue crushing online detection method, device, system and medium, the method comprising: obtaining vibration data of each target object corresponding to an acceleration time sequence; preprocessing the acceleration time sequence to generate a first acceleration time sequence corresponding to each target object, performing frequency response feature extraction and difference processing on the first acceleration time sequence to generate real-time difference frequency response data corresponding to each carton; inputting the real-time difference frequency response data into an energy modal prediction model to obtain corresponding expected frequency response data, determining a total prediction residual based on the similarity between the real-time difference frequency response data and the corresponding expected frequency response data; judging whether an EWMA value corresponding to the total prediction residual is greater than a preset anti-shake control feature energy threshold value, and determining a stacked carton crushing result according to the judgment result. Through the application, the problem that a scheme for detecting a stacked carton crushing in transportation in the related art is prone to frequent false positives or false negatives is solved.

The present application relates to the technical field of electric power engineering, and in particular to a power load prediction method fusing attention enhancement and dynamic fine-tuning mechanism, which firstly constructs an attention-enhanced Legendre memory model and improves the standardization layer thereof, enhances the processing capability of the model for high-order nonlinear power load data, simultaneously introduces a frequency enhancement channel attention mechanism, adaptively weights the features in the frequency domain dimension, effectively retains valuable information in high-frequency components, and suppresses noise interference; secondly, constructs a lightweight dynamic fine-tuning framework, finely corrects the prediction residual of the main model in a gradient boosting iteration manner, and significantly reduces the prediction deviation of the load trend mutation area; finally, comprehensively uses the trained attention-enhanced Legendre memory model and the fine-tuning framework to perform power load prediction. It has been verified that the present application can effectively improve the accuracy of short-period power load prediction and shows superior performance on both public data sets and real power data.

The application discloses a current transformer data prediction method and device, and relates to the electrical technical field.The method comprises the following steps: using a Prophet algorithm to perform additivity decomposition on time series data of a current transformer to obtain a trend item, a periodic item and a residual item; using an adaptive noise complete set empirical mode decomposition algorithm and a gated recurrent unit to sequentially process the residual item to obtain a prediction residual item and a plurality of prediction intrinsic mode functions; and using a Stacking algorithm to process the trend item, the periodic item, the prediction residual item and the plurality of prediction intrinsic mode functions to obtain a prediction residual, and obtaining a global prediction value based on the trend item, the periodic item and the prediction residual.The application solves the problems of low accuracy and poor stability in online monitoring of a current transformer in the related art.

A partial coefficient mode in which only some of the transform coefficients of a block is signaled is provided. When configured to perform encoding operations, a video coder generates a set of prediction residuals for the current block based on a predictor. The video coder transforms the set of prediction residuals into a set of transform coefficients and signals only a predetermined subset of the transform coefficients in the bitstream by omitting remaining transform coefficients outside of the subset. When configured to perform decoding operations, the video coder inverse-transforms a set of transform coefficients into a set of prediction residuals of the current block. Only a predetermined subset of the transform coefficients are parsed from the bitstream and the remaining transform coefficients in the set are set to one or more predefined values. The current block is reconstructed based on the predictor and the prediction residuals from the inverse-transforming.

The application discloses a multivariate time series anomaly detection method based on space-time feature fusion, which comprises the following steps: a series connection structure of a multi-scale TCN and a lightweight Transformer network is used to output time sequence features; an adjacency matrix is constructed through a static graph learner and a dynamic graph learner, and a multi-order GCN based on the adjacency matrix is used for processing to output space features; a gating mechanism is used to weight and fuse the time sequence features and the space features, and a multiple residual connection is combined to output a prediction value; in the training process, sample weights are dynamically calculated according to a prediction residual, and a weighted loss is calculated; according to a residual between the prediction value and an actual value, an anomaly score is generated through a reconstruction error calculation mechanism based on principal component analysis, and anomaly detection is completed by comparing a threshold value.

An encoder includes circuitry and memory. Using the memory, the circuitry: performs a transform process of (i) applying a first transform to a prediction residual signal indicating a difference between a current block to be encoded and a prediction image of the current block and (ii) further applying a second transform to a transform result of the first transform; and in the second transform, selects one transform basis (i) from a first group of candidates when a size of the current block is a first block size and (ii) from a second group of candidates when the size of the current block is a second block size different from the first block size, the first group including one or more candidates for a transform basis, the second group being different from the first group.

This invention discloses a method, system, device, and medium for data acquisition of industrial IoT devices, aiming to solve problems such as resource waste or loss of key features caused by fixed sampling frequencies. The method includes: acquiring raw signals and performing standardization processing; calculating the intensity of nonlinear fluctuations and prediction residuals of the signals, and evaluating the sampling completeness coefficient; extracting features to identify the device's operating modes; combining the completeness coefficient and mode identification results, calculating the target sampling frequency through multi-objective optimization and achieving seamless frequency switching; and adaptively compensating the evaluation model based on reconstruction errors. The system includes modules for data perception, sampling evaluation, mode identification, frequency decision-making, acquisition execution, and closed-loop feedback. This application achieves deep adaptation between the sampling frequency and signal characteristics, significantly reducing the communication and storage burden while ensuring reconstruction accuracy, and improving the system's adaptive capability and fault warning accuracy.

This application relates to a video encoding and decoding method, apparatus, chip, computer device, computer-readable storage medium, and computer program product. The video encoding method includes: determining a target sub-block partitioning method and a target transform kernel for the current coding unit, wherein the sub-block partitioning method corresponds to the transform kernel; partitioning sub-blocks from the current coding unit according to the target sub-block partitioning method; performing sub-block transform on the inter-frame prediction residuals corresponding to the sub-blocks using the target transform kernel to obtain transform coefficients; and obtaining the bitstream of the current coding unit based on the transform coefficients. The target sub-block partitioning method is one of multiple sub-block partitioning methods, which include at least two of the following: the width of the sub-block is half the width of the current coding unit, and the height of the sub-block is one-quarter of the height of the current coding unit; the width of the sub-block is one-quarter of the width of the current coding unit, and the height of the sub-block is half the height of the current coding unit. Using this method can improve the encoding effect.

The application relates to the technical field of runoff prediction, and discloses a factor dimension reduction and error correction method and system for runoff prediction, which can significantly improve the runoff prediction accuracy. The method comprises the following steps: calculating the correlation coefficient of each prediction factor in a data set and runoff data by using a Pearson correlation coefficient, and screening a strong correlation factor set; evaluating the prediction importance contribution of each factor in the strong correlation factor set by using an extreme gradient learning tree; constructing a runoff prediction model, and determining the final input dimension and runoff prediction value according to the principle that the larger the Nash coefficient is and the smaller the root mean square error is; decomposing the prediction residual by using ensemble empirical mode decomposition to obtain mode components and residual components; respectively predicting the mode components and the residual components by using an autoregressive model; superimposing the mode components and the residual components of the prediction residual, adding the mode components and the residual components to the original prediction value, and finally obtaining the sample prediction value after error correction.

An encoder (100) includes circuitry and memory. Using memory, the circuitry performs, on a current block to be processed, a transform process to apply a second transform to transform coefficients obtained by applying a first transform to a prediction residual signal. The current block is included in a plurality of blocks that have a plurality of sizes, and the second transform has a common block size for the plurality of blocks. The circuitry selects, in the second transform, a transform base from a group of candidates which include one or more candidates for a transform base and are different according to the size of the current block.

Provided are a point cloud encoding method and decoding method, and an encoder and a decoder. The encoding method includes: processing attribute information of a target point in a point cloud, so as to obtain a predicted residual of the attribute information of the target point; quantizing the predicted residual on the basis of a quantized weight of the target point and a quantized stride of the target point, so as to obtain a quantized residual of the attribute information of the target point, wherein the quantized weight of the target point is the weight used when weighting the quantized stride of the target point; and encoding the quantized residual, so as to obtain a code stream.

The application discloses a GB-RAR deformation monitoring and early warning method based on a hybrid algorithm optimized prediction model, and relates to the technical field of deformation monitoring and early warning. The method comprises the following steps: acquiring GB-RAR deformation time series data of a research object by using a ground-based real aperture radar system; constructing a neural network deformation prediction model with multidimensional environmental factor time series data as input and deformation time series data as prediction output, and training the model by using a constructed historical sample data set; performing difference processing on the deformation prediction time series data and the pretreated GB-RAR deformation time series data to obtain a prediction residual sequence and a standard deviation sequence arranged in time sequence; taking the median of the standard deviation sequence as a stability benchmark and calculating an alarm threshold corresponding to a current sliding window; and if a continuous number of sliding windows in the prediction residual sequence are all marked as structural abnormalities, determining that the continuous number of sliding windows are structural early warning events. The application realizes early warning, accurate alarm and reliable alarm of a complex research object, and improves the operation safety early warning and pre-control capability.