46 results about "Resampling" patented technology

The invention relates to the technical field of numerical control machining measurement and correction, and discloses a machining surface correction method for a numerical control milling center. The method comprises the steps of establishing an original point cloud of a processing surface and extracting a feature structure; analyzing each feature structure to form a structured description list; inputting the list into a geometric growth model, and generating a virtual feature structure set according to rule iterative evolution; extracting and rasterizing the contour and the height field, carrying out registration fusion with a reference plane equation, and carrying out variation calculation to obtain a smooth middle geometric surface; performing adaptive resampling on the intermediate surface to generate an intermediate point cloud; and finally, according to the spatial frequency distribution of a topological residual field formed by the difference between the original point cloud and the intermediate point cloud, adaptively determining a weight, and carrying out weighted mixing on the original point cloud and the intermediate point cloud to generate an optimized surface point cloud. According to the method, intelligent evolution of the machined surface features and frequency domain self-adaptive data fusion are achieved, and the correction machining precision and the surface forming quality are improved.

Some embodiments are directed to a resource allocation analysis system implemented via a back-end application computer server. A resource data store may contain electronic records associated with a set of resource types, each electronic record including an electronic record identifier and resource parameter. The back-end application computer server may receive, from the resource data store, information about a set of resource types to be analyzed, including the associated resource parameters. The computer server may then execute a differential evolutionary algorithm to optimize the set of resource types based on at least one non-linear constraint and generate resource analysis results. The back-end application computer server may, according to some embodiments, perform a resampling process that uses non-parameterized historical data, regression on at least one resource type, and moment matching.

The application discloses a kind of power electronic system fault diagnosis methods combined with resampling and integrated learning, belong to power electronic equipment fault diagnosis technical field, the method includes real-time sampling electric power electronic system in the current data of one-half fundamental period, current data is normalized, and normalized data is obtained;According to the normalized data, based on fast fourier transform algorithm, the frequency domain characteristics of normalized data are obtained;According to the frequency domain characteristics, the feature vector of normalized data is obtained by feature extraction selector;According to the feature vector of normalized data, the fault class label is obtained using integrated classification model, and the electric power electronic system fault diagnosis is completed.The application solves the technical problems that the original data sample is unbalanced in the existing fault diagnosis method, which leads to inaccurate diagnosis and even wrong judgment of the existing machine learning model, effectively realizes the diagnosis of various fault types of sensors and power devices of electric power electronic system.

The invention discloses an injection molding quality soft measurement method for staged resampling and quality enhancement, which belongs to the technical field of injection molding and comprises the following steps: acquiring a process data set and a quality data set; normalizing the process data set and the quality data set to obtain a normalized process data set and a normalized quality data set; performing clustering, staged resampling and merging on the normalized process data set to obtain a low-resolution process data set; and inputting the low-resolution process data set and the normalized quality data set into a quality enhancement prediction network to obtain a prediction result of the injection molding quality. The accuracy of injection molding quality soft measurement is improved.

The chain system fixed-point resampling method belongs to the technical field of rotating machinery signal processing, and aims to solve the problems that the time domain signal regularity of a single chain link is poor, the dynamics characteristics of the whole system are difficult to directly reflect, and the frequency domain feature extraction cannot be performed. The method comprises the following steps: extracting original data of discrete motion units from multi-body dynamics simulation results; at the preset data monitoring point position of the system, the nearest discrete motion unit is searched as a monitoring unit in real time by calculation; according to the data type, one of a direct method, a synthetic method and an average method is adopted to process the original data of the monitoring unit, and feature data representing the dynamics state of the fixed position are generated; and finally, the feature data are written into a file in time sequence. The time sequence data of the fixed position are collected through the fixed-point resampling method, and the problems that the data is strong in contingency, poor in regularity, and difficult to perform in-depth frequency domain and statistical analysis due to tracking a single motion unit are solved.

The invention discloses a fixed-point resampling method for a chain system, belongs to the technical field of signal processing of rotating machinery, and aims to solve the problems that the regularity of a time domain signal of a single chain link is poor, the dynamic characteristics of the whole system are difficult to directly reflect, and frequency domain feature extraction cannot be carried out. The method comprises the following steps: extracting original data of a discrete motion unit from a multi-body dynamics simulation result; searching the nearest discrete motion unit as a monitoring unit in real time by calculating the distance at a preset data monitoring point position of the system; according to the data type, processing the original data of the monitoring unit by adopting one of a direct method, a synthesis method and an average method, and generating characteristic data representing the dynamic state of the fixed position; and finally, writing the feature data into a file according to a time sequence. According to the method, time sequence data of a fixed position is collected through a fixed-point resampling method, and the problems that due to tracking of a single motion unit, data contingency is high, regularity is poor, and frequency domain deepening and statistical analysis are difficult to carry out are solved.

This invention discloses a deep learning-based engine condition monitoring method, comprising the following steps: collecting multi-source sensor data and corresponding speed data during engine operation; preprocessing the original multi-source data to obtain preprocessed multi-source data; performing condition-locked resampling to obtain a set of window sequences; constructing a Krylov subspace and performing orthogonalization to obtain an orthogonal basis for the subspace; performing Krylov subspace change point and state transition detection to obtain the current monitoring state; generating a subspace fingerprint and selecting a deep learning monitoring branch based on the subspace fingerprint routing structure to obtain a health index and anomaly score; writing the health index and anomaly score into a candidate alarm queue, switching the current monitoring state, and outputting alarm results based on the candidate alarm queue. This invention improves the accuracy of engine condition monitoring by combining condition-locked resampling with Krylov subspace change point detection.

Apparatuses, methods, and systems for determining winning arms of electronic testing. One method includes obtaining historical data values related to the A / B test of a user, storing the historical data values, determining a historical weight for the historical data values, receiving new data values from the plurality of computing devices collected based on recipient actions during execution of the A / B, constructing a Dirichlet distribution, inferring corresponding central tendencies of samplings of a metric distribution, wherein each central tendency of the corresponding central tendencies is determined by sampling the Dirichlet distribution, constructing an overall utility distribution for each arms of the A / B test by combining the central tendency of each sampling of the metric distribution with a corresponding sampling of a conversion probability distribution, determining a winning arm of the A / B testing by comparing the overall utility distribution of each arm with each other arm of the A / B test.

The invention discloses a minimum region method flatness error evaluation algorithm based on improved teaching and learning optimization, and belongs to the technical field of geometric quantity measurement. In order to solve the problems that a traditional minimum region method is slow in convergence and prone to falling into local optimum, a teaching and learning optimization algorithm is improved through a Sine chaotic mapping population, a self-adaptive learning factor and local resampling search, and the method is applied to minimum region evaluation of flatness errors. The method comprises the specific steps of point cloud data preprocessing, Sine chaos population initialization, teaching stage adaptive updating, local resampling, convergence judgment and result output. Experiments show that compared with traditional TLBO, particle swarm optimization and other algorithms, the algorithm is higher in convergence speed and evaluation precision, higher in robustness and suitable for flatness detection scenes in high-precision manufacturing.

The invention provides a multi-model multi-resolution self-adaptive image recognition and extraction method based on deep learning, and the method comprises the following steps: an unmanned plane collects an aviation picture, and generates a digital orthoimage through a picture splicing algorithm; judging whether natural language analysis needs to be performed or not according to a processing requirement input by a user; obtaining a target category to be recognized and extracted, and adaptively selecting an optimal deep learning algorithm; according to the overall contour pixel size of the target category, selecting a self-adaptive spatial resolution to carry out reprojection resampling processing; obtaining the recognized and extracted original contour; according to the method, a complex natural language can be analyzed, an unknown target category can be recognized and extracted, a high-precision specific category can also be matched, generalization and expandability are enhanced, and the method has the advantages of being high in robustness, high in robustness and the like. And optimal algorithm parameters can be adaptively selected, and the precision requirement of the result is ensured to the greatest extent.

The application discloses a small sample modeling stability evaluation method and system based on Bootstrap resampling, and particularly relates to the technical field of computer and data intelligence, and the method comprises the following steps: completing data preprocessing and stable stage identification under a unified time base, and forming a segmented set; taking the average block length as the decision quantity to optimize SBB to construct a sample space, and determining the optimal block length in combination with the variance consistency and nominal coverage consistency criteria; performing resampling training around three scales of short, medium and long, and collecting layered indexes such as prediction, parameters and features; completing stability calculation and empirical coverage calibration based on scale-in statistics and scale-weighted mixing, and outputting a pseudo-stationary diagnosis score; finally, engineering judgment and rollback optimization are performed according to the coverage, correlation retention and risk threshold. The system records random seeds, block metadata and model version generation evidence pointers, has traceable and retestable characteristics, and is suitable for scenes such as production lines, network traffic and finance.

The application discloses a kind of synthetic and real data's adaptive resampling method and device, electronic equipment, comprising: obtaining training dataset from multiple data sources;Uniformly pre-process multiple-source training data, and construct data source identification and sample basic feature representation;Based on basic feature representation, calculate the distribution deviation between different data sources and sample bias feature;According to distribution deviation and sample bias feature, the weight of each sample is calculated or resampling probability is calculated;Based on the weight of removing bias or resampling probability, adaptive resampling is executed to multiple-source training data, and the training sample set of removing bias is constructed;Training or updating is carried out to target model using the training sample set of removing bias, and resampling strategy is dynamically adjusted based on model training feedback, to realize multiple rounds of adaptive debiasing and resampling optimization.

The invention provides a three-dimensional medical image segmentation method without resampling, and relates to the field of computer vision, in particular to the three-dimensional medical image segmentation method without resampling. The three-dimensional medical image segmentation framework does not need to be resampled in training and reasoning stages, information of original medical images can be reserved to a greater extent, detail loss caused by resampling is avoided, and therefore segmentation precision is improved. Convolutional low-rank decomposition is introduced into the attention module, so that global attention calculation which originally needs to be subjected to high-dimensional matrix product can be approximated to low-rank matrix operation, and the calculation complexity is effectively reduced. Meanwhile, the corresponding self-adaptive parameterized European matrix is also compressed to a small scale, video memory occupation and computing resource consumption are greatly reduced under the condition that global dependence modeling capability is guaranteed, and efficient operation on large-scale three-dimensional medical image data can be achieved.

This invention discloses an AUV-UFastSLAM algorithm optimized based on the whale algorithm, comprising a series of processes including initialization, prediction, sampling, map updating, and resampling. This method optimizes the particle sampling process using the whale algorithm, enabling the particle swarm to move towards a high-likelihood region, making the AUV pose estimation closer to the true value. An inertia weight factor is introduced to improve the whale algorithm's position update formula, increasing the population convergence speed and accuracy. Simultaneously, Cauchy mutation is used to randomly perturb the optimal neighborhood to increase population diversity and improve the algorithm's global search capability. An improved resampling method is used for particle filtering to ensure particle diversity. Through these adjustments, the accuracy of AUV simultaneous localization and map creation is improved.

The application discloses a kind of MU-MIMO system wireless resource allocation method and system based on condition diffusion strategy enhancement, for MU-MIMO system downlink transmission, method includes: with the total capacity of the MU-MIMO wireless communication system as target, jointly optimize resource block allocation and transmission power, construct optimization problem;Condition diffusion model is trained based on expert dataset;Multiple candidate solutions are generated using the trained model, and the optimal solution of the optimization problem is obtained by further optimizing the multiple candidate solutions through policy enhancement;By constructing a condition diffusion model, imitation learning is performed on the expert dataset;And introduce resampling mechanism, select the optimal solution from multiple candidate solutions, while ensuring to meet system constraints, significantly improve system throughput and reduce computational complexity.

This invention relates to the field of autonomous driving technology, specifically to a method, system, device, and medium for fast trajectory resampling based on autonomous driving motion constraints. This method introduces vehicle kinematics to control a virtual vehicle to move along a desired trajectory, thereby generating a trajectory that satisfies the kinematics, overcoming the shortcomings of filtering methods that cannot take kinematics into account. Simultaneously, to address the problem that control-based methods cannot converge at the final target point of the trajectory, a combination of forward and backward filtering is used, solving the lag problem caused by simple filtering methods while inheriting the advantage of fast computation speed of filtering methods for rapid trajectory resampling.

The invention belongs to the technical field of artificial intelligence, and relates to an artificial intelligence-based image processing method, which comprises the steps of receiving an input original image; in a processing process of decoding an original image by using a visual language model, performing perception graph construction based on a self-attention mechanism to obtain a corresponding initial perception graph; carrying out iteration refining processing based on the initial perception graph to obtain a first perception graph; performing optimization processing on the first perception graph based on a preset post-processing strategy to obtain a second perception graph; performing nonlinear resampling processing on the original image based on the second perception image to generate a target image; reasoning the target image based on a visual language model to generate text response data; and outputting the text response data. The invention further provides an image processing device based on artificial intelligence, computer equipment and a storage medium. The method can be applied to image processing scenes in the financial science and technology field and the digital medical field, and the accuracy of the generated text response data is effectively improved.

Methods and apparatuses are provided for encoding and decoding video data based on a supplemental enhancement information (SEI) message. An exemplary method includes: generating a reconstructed frame sequence based on a compressed video; decoding a supplemental enhancement information (SEI) message with respect to the reconstructed frame sequence, according to the compressed video; and performing temporal upsampling to the reconstructed frame sequence based on the SEI message by using a neural network.

The application relates to the technical field of airplane test data analysis, and discloses an airplane task sensor ground test data alignment method. The method firstly performs nearby alignment on data starting time, then realizes unification of sampling intervals based on resampling, performs upsampling operation on data with large sampling intervals, and realizes data filling based on interpolation. The application can effectively improve the overall quality of data, provide more accurate and reliable basis for subsequent data analysis and modeling, and promote airplane performance analysis and optimization.

The invention discloses an adaptive resampling method and device for synthetic and real data, and electronic equipment. The method comprises the following steps: acquiring a training data set from a plurality of data sources; uniformly preprocessing the multi-source training data, and constructing data source identifiers and sample basic feature representations; calculating distribution deviation and sample bias features among different data sources based on the basic feature representation; according to the distribution deviation and sample bias characteristics, calculating a depolarization weight or a resampling probability of each sample; based on the depolarization weight or the resampling probability, adaptive resampling is carried out on the multi-source training data, and a depolarization training sample set is constructed; and training or updating the target model by using the depolarization training sample set, and dynamically adjusting a resampling strategy based on model training feedback, thereby realizing multi-round adaptive depolarization and resampling optimization.

The present application provides an image processing device, an image processing method and a computer program product capable of reducing the amount of calculation and improving the efficiency of post-processing. The image processing device processes a primary segmentation processing result obtained by performing segmentation processing on an image at a segmentation resolution, and comprises: a boundary point searching unit configured to extract a set of specific boundary points on the image according to a probability map as the primary segmentation processing result; a positioning unit configured to map the extracted specific boundary points back to the probability map to obtain coordinates of the specific boundary points; and a resampling unit configured to perform first resampling processing on the specific boundary points on the probability map, perform second resampling processing different from the first resampling processing on other points on the probability map except the specific boundary points, and obtain a segmentation processing result at a target resolution.

The invention provides a large model training method and device and electronic equipment, and relates to the technical field of artificial intelligence, in particular to the technical field of deep learning, large models and the like, and the method comprises the steps that a training sample set is determined, and training samples comprise sample cues, expected output of the sample cues and labels of the training samples; training the large language model based on the training sample, outputting a prediction output of a sample cue word, and determining a first Kalman-Terworth optimization KTO loss of the training sample based on the expected output, the prediction output and the tag; determining a resampling coefficient, a first weight and a second weight of a sample group where the training sample is located, and determining a joint weight of the training sample based on the resampling coefficient, the first weight and the second weight; and according to the joint weight and the first KTO loss, determining a second KTO loss of the large language model, and based on the second KTO loss, performing gradient adjustment on model parameters of the large model to obtain a target large language model.

The invention discloses a self-adaptive arbitrary resampling method and device, and relates to the technical field of digital signal processing, and the method comprises the steps: designing a basic filter based on a windowing function interpolation algorithm, and optimizing the basic filter into an arbitrary group delay filter; the method comprises the following steps: calculating a resampling factor through an input signal sampling rate and an output signal sampling rate, adaptively adjusting a resampling parameter of a filter, calculating an output signal length in combination with an input signal length, judging a resampling periodicity through an error threshold value to determine a cycle period point number and a storage mark, and finally traversing output sampling points. Filter coefficients are obtained in three modes of coefficient real-time calculation, calculation and storage and calling, and target sampling rate signals are output after convolution filtering and gain correction. According to the method, the problems of limited precision and large operand of the existing arbitrary resampling method are solved, and the sampling precision and efficiency are considered at the same time.

A decoding method, an encoding method, a video coding apparatus, and a computer readable medium are provided. The decoding method may be executed by a decoder deployed with an artificial intelligence (AI) model or functionality. The method includes: obtaining a weight information; obtaining a resampling configuration; determining a resampled weight based on the weight information and the resampling configuration; and applying the resampled weight for the AI model or functionality.

The present application relates to the technical field of industrial intelligent monitoring and fault diagnosis, and particularly relates to an unbalanced fault diagnosis method based on attention perception and decision coupling. The technical scheme comprises the following steps: obtaining a one-dimensional vibration time series observation signal of a target industrial equipment in a running state, and pre-processing the signal; and inputting the pre-processed time series signal into a Gram angle field sine transformation, a Gram angle field difference transformation and a Markov transition field for parallel coding. The present application can autonomously capture weak fault fingerprints submerged in strong background noise without relying on manual resampling or prior threshold setting, can guarantee the reliability of the diagnosis of majority class samples, can significantly improve the upper limit of the recognition of a small number of key fault patterns, and can provide an intelligent diagnosis solution with high sensitivity, strong robustness and without manual intervention for complex industrial scenes such as wind power, aviation and rail transportation which have high reliability requirements and difficulty in obtaining fault data.

The invention discloses a rolling bearing fault diagnosis method based on physically enhanced double-branch ResNet, and belongs to the technical field of intelligent fault diagnosis of rotating machinery. According to the method, for the technical problems of fault feature frequency drift, difficulty in impact modulation information extraction and poor generalization ability of a data driving model under a variable speed working condition, two types of complementary features including an angular domain resampling envelope feature and a cepstrum mode enhancement feature are constructed, and a double-branch residual network is input for fusion and diagnosis; in training, spectrum energy constraint loss fusing fault order priori and TVCS-PINN dynamic residual loss including non-negative and smooth constraints are introduced to enhance the physical consistency and generalization ability of the model. According to the method, the stability and characterization capability of fault features under the variable speed condition are effectively improved, and the diagnosis accuracy and the cross-working-condition adaptability are improved through physical constraint driving.

The application provides a three-dimensional medical image segmentation method without resampling, relates to the field of computer vision, and particularly relates to a three-dimensional medical image segmentation method without resampling. The three-dimensional medical image segmentation framework of the application does not need resampling in the training and reasoning stages, can retain the information of the original medical image to a greater extent, avoids the loss of details caused by resampling, and thus improves the segmentation accuracy. By introducing convolutional low-rank decomposition into the attention module, the application can approximate the global attention calculation of high-dimensional matrix multiplication to low-rank matrix operation, thereby effectively reducing the calculation complexity. At the same time, the corresponding adaptive parameterized Euclidean matrix is also compressed to a small scale, under the condition of ensuring the global dependence modeling capability, the memory occupation and the calculation resource consumption are greatly reduced, so that the application can be efficiently run on large-scale three-dimensional medical image data.

The invention provides a diffusion model privacy data protection method based on resampling and anomaly detection. The method comprises the following steps: performing fine adjustment on a pre-trained text graph diffusion model; inputting a text prompt required to be forgotten by a user into the fine-tuned text graph diffusion model, and calculating a distribution condition of neurons concerned by the user in a corresponding neuron module; screening neurons of which the activation values are greater than a preset threshold value in the distribution; carrying out re-sampling operation on the screened neuron activation values according to original parameter distribution; according to the method, the screened neuron parameters are re-sampled by adopting a self-service method, and the original parameters are replaced by the parameter values obtained by sampling, so that efficient privacy data removal is realized on the premise of not damaging the generation quality and diversity of the model; the forgetting depth and speed of the model are optimized, and the privacy of the user is effectively protected; and the efficiency of the model is greatly improved, and the required calculation consumption is reduced.

The application discloses a kind of sensor cooperative processing method and system based on industrial internet of things, it is related to industrial internet of things technical field, method includes: obtaining original sampling data and executing dynamic resampling according to highest instantaneous frequency;Resampling data is converted into frequency sequence and equivalent energy fluctuation sequence based on dimension energy equivalent mapping, so as to calculate difference sequence;According to physical coordinates, construct space-time delay compensation matrix to carry out phase alignment;Combining sampling density and background fluctuation standard deviation filters data and generates dynamic active coefficient;Using physical space span calculates space weight factor, and active coefficient is combined to generate correction coefficient carrying topological attenuation state;The coefficient is input into semantic fusion network to obtain correlation coefficient, and normalized sequence is weighted and fused to generate cooperative representation vector;Finally, its local normalized physical deviation degree is calculated, and compared with safety boundary to output early warning instruction.The application has the advantages of anti-aliasing, space-time accurate alignment and semantic fusion.