Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

33results about How to "Increase computational cost" patented technology

CFD-based wind turbine generator load simulation method under high slope

A CFD (computational fluid dynamics)-based wind turbine generator load simulation method under a high slope comprises the following steps: step 1, acquiring topographic data and position information of the high slope: acquiring the topographic data and position information of the high slope through a multi-source data acquisition means; step 2, constructing a high slope simplified physical model according to the high slope topographic data and the machine position information; 3, a CFD tool is used for flow field numerical simulation calculation, and according to the constructed high slope simplified physical model, the CFD tool is used for obtaining the wind speed distribution and turbulence intensity distribution conditions of the whole wind wheel face of the wind turbine generator; step 4, generating a target wind seed file by using a TurbSim tool; 5, establishing a high slope wind file library; step 6, calculating the dynamic response of the whole machine; and obtaining an accurate load result of the wind turbine generator under the high slope terrain. The accuracy of the simulation result is improved, and the safety, the power generation efficiency and the service life of the wind turbine generator are improved.
Owner:THREE GORGES INTELLIGENT CONTROL TECHNOLOGY CO LTD

TransUNet-based thyroid nodule ultrasonic image segmentation method and system

The invention discloses a thyroid nodule ultrasound image segmentation method and system based on TransUNet, and relates to the technical field of image segmentation. The method comprises the steps that thyroid nodule ultrasonic images are received and preprocessed, and image data are divided into a training set, a verification set and a test set; and constructing an improved segmentation model based on a TransUNet model network, wherein the improved segmentation model comprises a CNN-SETrans encoder, a decoder and a jump connection. On the basis of a TransUNet network structure, the SE module is introduced into a Transform encoder and is connected with the MSA module in series to construct a SETrans encoder, channel information and spatial information are comprehensively extracted, the feature representation capability of the network can be improved, and by introducing the Ghost module into jump connection, the performance of the network is improved. According to the method, the loss of image features in the long-distance transmission process from the low resolution of the CNN to the high resolution of the decoder can be compensated, the calculation cost cannot be remarkably increased, and meanwhile, the common convolution at the tail of each layer in the decoder is replaced by the DSConv, so that the effectiveness of feature extraction is improved.
Owner:SUZHOU UNIV

Image or video generation method and system based on diffusion model self-guidance

The invention discloses a diffusion model self-guiding-based image or video generation method and system. The method comprises the steps that the input content is a noise potential variable and a future offset step length of a current time step; the diffusion model predicts the score network twice based on the input content to obtain a conditional score network of the current time step and the future offset time step; carrying out linear combination on the two conditional score networks to generate a combined conditional score network; denoising the noise potential variable of the current time step by using a combined condition score network; calculating and updating a potential variable of the previous time step according to a denoising result; in each time step, the self-guiding strategy is adopted to carry out denoising processing on the noise, and noise prediction is optimized; and after de-noising processing of all time steps, a de-noising potential variable is finally output as a final model generation result, and an image or a video is obtained. According to the invention, the quality of the generated image or video can be obviously improved.
Owner:WESTLAKE UNIV

MRI reconstruction method of subspace diffusion model based on orthogonal decomposition

The invention provides an MRI (Magnetic Resonance Imaging) reconstruction method of a subspace diffusion model based on orthogonal decomposition. The MRI reconstruction method comprises the following steps: forming an open source brain data set; a subspace diffusion model based on orthogonal decomposition is constructed, and k-space data is decomposed into subspaces for diffusion, so that data distribution is better captured, and robustness is kept in a noise environment; designing an iterative solution algorithm for extracting prior information in the model; and applying an iterative solution algorithm to the model to test the test data set, and outputting a final restored image. According to the method, the learning of effective information in complex distribution is promoted, the inherent high-dimensional extrapolation challenge is solved, and the calculation cost related to a diffusion model is reduced. Comprehensive experiments on different data sets clearly show that the method is superior to other methods in the aspects of reconstruction speed and quality.
Owner:NANCHANG UNIV

Transformer three-dimensional electromagnetic simulation method based on NU-UNET and application

The invention relates to the technical field of three-phase transformer electromagnetic field simulation, in particular to a transformer three-dimensional electromagnetic simulation method based on NU-UNET and application, and the method comprises the following steps: S1, building a three-phase transformer model by using comsol; s2, acquiring and preprocessing three-dimensional electromagnetic field simulation data; s3, an NU-UNET model is built and trained; and S4, carrying out reverse processing and analysis on a model prediction result. According to the method, an NU-UNET model is used, a K-D tree algorithm is used for carrying out regional decomposition, then grids of different sizes are adaptively used for interpolation on different sub-domains, the sub-domain grids subjected to interpolation are used for training a grid-based model of a bottom layer, that is, input grid data features are extracted through a convolutional neural network, and the grid data features are extracted through the convolutional neural network. The method does not depend on specific grid division, and can quickly and efficiently carry out electromagnetic field distribution prediction.
Owner:GUILIN UNIV OF ELECTRONIC TECH

Parameterized level set blade topological optimization method based on conformal geometry theory and isogeometric analysis

PendingCN119129118AAchieve efficient and accurate optimization purposesincrease computational costGeometric CADDesign optimisation/simulationAlgorithmMathematical model
The invention discloses a parameterization level set blade topological optimization method based on a conformal geometry theory and isogeometric analysis, and the method comprises the following steps: building an initial blade model according to an existing blade profile; performing fluid-solid coupling mechanical analysis on the blade structure model, and calculating a blade surface load; the blade structure model is mapped into a two-dimensional rectangular domain through conformal mapping, and a standard orthogonal grid with the same specification as a mapping grid is established; establishing an optimized mathematical model of the blade structure; isogeometric analysis of the blade structure; optimizing a parameterized level set; and inversely mapping the blade structure to complete the optimization process. According to the method, through the conformal mapping theory and the isogeometric level set method, three-dimensional curved surface dimension reduction processing is achieved, the calculation cost is reduced, and the optimization efficiency is improved. And meanwhile, the high-precision calculation advantage of isogeometric analysis is combined, a level set optimized smooth structure is also provided, the checkerboard problem of methods similar to variable density and the like is avoided, and the purpose of efficient and accurate optimization of the curved surface structure is achieved.
Owner:GUANGXI UNIV

Numerical simulation method for estimating equivalent seepage of three-dimensional fracture network based on topological property

The invention discloses a numerical simulation method for estimating equivalent seepage of a three-dimensional fracture network based on topological properties, which comprises the following steps: firstly, constructing a small three-dimensional fracture network sample set on the basis of related data of a fracture network in a research area, then mapping the sample fracture network into an equivalent two-dimensional undirected graph by using a graph theory, solving equivalent permeability, and calculating the equivalent seepage of the fracture network. Forming a training sample set; then, a maximum information index method is adopted to mine fracture network equivalent permeability main control factors, and the control relation between fracture network parameters and an underground flow field is determined; and finally, constructing a substitution model between the main control factors and the equivalent permeability based on a BPNN machine learning model, and realizing conversion from the target discrete fracture network to an equivalent medium. According to the method, a simulation scheme which reflects topological connectivity and is high in prediction precision and low in calculation cost is provided for numerical simulation of the fractured stratum underground flow field under the field scale.
Owner:NANJING UNIV +1

Beidou navigation unmanned aerial vehicle inspection path planning method based on simulated annealing algorithm

The invention discloses a Beidou navigation unmanned aerial vehicle inspection path planning method based on a simulated annealing algorithm. The method comprises the following steps: S1, reading latitude and longitude data: reading latitude and longitude data of an inspection target point and a base from a file or a database; s2, calculating a distance matrix: calculating an actual distance between any two points according to the longitude and latitude data, and constructing the distance matrix; s3, constructing a solution space: constructing all path sets; s4, selecting an initial solution: selecting an initial path from the path set as a current solution; s5, setting an initial temperature: setting the initial temperature of a simulated annealing algorithm; s6, loop iteration: performing an iteration process of a simulated annealing algorithm; and S7, generating a new solution: generating a new path as the new solution by using a transformation method 2 or a transformation method 3 according to the current solution. The method is used for solving the problem that a genetic algorithm and a particle swarm optimization algorithm are relatively weak in theoretical convergence guarantee, and particularly are possibly difficult to meet theoretical conditions in practical application.
Owner:CHINA THREE GORGES UNIV

Checkpoint deployment method, system, product and device for heterogeneous computing system

The invention discloses a check point deployment method, system, product and device for a heterogeneous computing system, and relates to the technical field of computers. According to the demand parameter of each heterogeneous computing power device, calculating a fault recovery time-consuming expected value corresponding to each storage device when the check point is deployed on each storage device, so as to balance the influence of the recovery of the check point corresponding to each demand parameter of each heterogeneous computing power device. And determining a preliminary first check point deployment relationship according to the demand parameter of each heterogeneous computing power device and the fault recovery time-consuming expected value, so as to realize the predictability of local optimal selection from the corresponding deployment check point to each storage device for the characteristic parameter of each heterogeneous computing power device in the deployment process. According to the preliminary first check point deployment relation, the second check point deployment relation is determined through optimization of the deployment positions of the heterogeneous computing power devices, the possibility of finding a global optimal solution by effectively jumping out of a local optimal solution is increased, and the execution efficiency of a training task is improved.
Owner:LANGCHAO ELECTRONIC INFORMATION IND CO LTD

Student classroom behavior detection method based on Transform and task dynamic alignment

The invention provides a student classroom behavior detection method based on Transform and task dynamic alignment. On one hand, the GLTB is added into the backbone network to enhance the feature extraction capability. Wherein the CGLU is used for enhancing the non-linear feature expression capability, and the global feature extraction capability of the MHSA is combined, so that the method has higher feature extraction capability. And on the other hand, a neck structure network improved by GLSA is adopted, global features can be effectively extracted, fine-grained local features can be captured, and interaction between global information and local information can be promoted. According to the method, when a complex task is processed, key information in the image can be identified and understood more accurately. In addition, a novel detection head ETADH is adopted, interactive features of classification and positioning tasks are learned through a feature extractor, and task decomposition is performed through a task decomposition module. In this way, the interactivity between the two tasks is enhanced, and the detection precision is improved by effectively utilizing the associated information between the two tasks.
Owner:CHONGQING UNIV OF POSTS & TELECOMM

Unfavorable geology detection method, system and device based on multi-source aerial remote sensing data fusion

The invention relates to the technical field of remote sensing, and discloses an unfavorable geology detection method, system and device based on multi-source aerial remote sensing data fusion, and the method comprises the steps: obtaining laser radar data, hyperspectral image data, CCD image data, infrared image data and integrated navigation data of a target area, and carrying out the preprocessing; carrying out space matching processing on the preprocessed laser radar data, the preprocessed hyperspectral image data, the preprocessed CCD image data and the preprocessed infrared image data; fusing the laser radar data, the hyperspectral image data, the CCD image data and the infrared image data after application space matching processing for geological detection to obtain different geological detection data; and obtaining an unfavorable geological detection result of the target area in combination with different geological detection data. According to the method, the multi-source data fusion processing precision can be improved, multi-source heterogeneous information is effectively integrated, comprehensive and efficient detection of various unfavorable geologies is realized, the misjudgment and missed judgment risks are reduced, and the credibility of unfavorable geology detection results is improved.
Owner:CHINA RAILWAY DESIGN GRP CO LTD

Self-planning method, device and equipment for mathematical geometry problem of intelligent agent and medium

The embodiment of the invention provides a self-planning method for a mathematical geometry problem of an intelligent agent, which can be applied to the technical field of artificial intelligence. The self-planning method for the mathematical geometry problem of the intelligent agent comprises the following steps: generating a seed image according to scene demand information of the mathematical geometry problem; performing complex feature planning on the seed image based on a preset guide search rule until a target mathematical problem set is output; and performing problem solving path matching on at least one target mathematical problem in the target mathematical problem set through a preset language model to generate a problem solving planning path set. The embodiment of the invention further provides a self-planning method and device for the mathematical geometry problem of the intelligent agent, equipment, a storage medium and a program product.
Owner:BEIJING INSTITUTE FOR GENERAL ARTIFICIAL INTELLIGENCE

Multi-channel sampling circuit hybrid error modeling and cyclic calibration method and system

The invention discloses a multi-channel sampling circuit hybrid error modeling and cyclic calibration method, and belongs to the technical field of signal sampling. The method comprises the following steps: constructing a system error model, and introducing polynomial series to realize model simplification; generating a reference signal participating in error cycle calibration according to the simplified hybrid error model; initializing a frequency response function of each sampling channel, and calculating an initial tap coefficient of an error correction module; a cascade multi-order error correction structure is used for completing reference signal error iteration correction, so that the reference signal error continuously approaches to an input signal; correcting the signal to calculate a loss function, and completing estimation and updating of a tap coefficient; and judging whether the loss function reaches a convergence condition or not until hybrid error cyclic calibration is realized. On the basis of the reference signal constructed by the hybrid error model, low-complexity online blind cycle calibration is realized, the precision and completeness of the system error model are improved, the quantity of parameters to be estimated and the number of sample points for calibration are remarkably reduced, and the calibration efficiency is improved.
Owner:BEIJING INST OF TECH

Corn microbial marker screening and phenotype prediction method

The invention relates to a corn microbial marker screening and phenotype prediction method, and belongs to the technical field of corn breeding. The method comprises the following steps: preprocessing a microorganism sample: acquiring corn microorganism sample data, and performing data cleaning and standardization operation to generate a standardized microorganism abundance table; differential microorganisms are analyzed and screened through LEfSe, and finally a significant differential microorganism feature set is obtained; performing feature importance evaluation based on LightGBM: dividing the significant difference microorganism feature set screened by LEfSe into a training set and a test set; a LightGBM (LightGBM) model is constructed; the LightGBM model is used for calculating and evaluating the feature importance of the microorganisms; screening important characteristics to obtain a microbial marker for deeply analyzing key microbial community characteristics of corn phenotypic changes; and model verification and performance evaluation: verifying the prediction effectiveness of the screened microbial characteristics through model training, model verification and model performance evaluation. According to the method, the prediction effect of the model can be improved while important biological characteristics are reserved.
Owner:UNIV OF JINAN

Method of data transfer between fluid domain and solid domain in fluid-solid coupling simulation

The present invention discloses a method for data transmission between a fluid domain and a solid domain during a fluid-solid coupling simulation process, and belongs to the technical field of fluid-solid coupling simulation of impeller machinery. The method comprises: obtaining data information on the known data side of the fluid-solid coupling interface to form a known data set; obtaining data information on the unknown data side of the fluid-solid coupling interface to form an unknown data set; selecting a target node in the unknown data set; finding the three nodes 1, 2, and 3 closest to the target node in the known data set; determining the distance d from the target node to the plane where nodes 1, 2, and 3 are located, and if d is less than or equal to a preset threshold, calculating the target node data; selecting the next target node in the unknown data set, and repeating the above steps until all nodes in the unknown data set obtain corresponding data information. The present invention can realize accurate data transmission at any grid scale, and does not significantly increase the time of data mapping.
Owner:BEIHANG UNIV

Large-scale travel order connection plan automatic generation method and device considering time-varying network, and readable storage medium

PendingCN119624022Asimplify complexityincrease computational costForecastingGenetic algorithmsPathPingSimulation
The invention provides a time-varying network-considered large-scale travel order connection plan automatic generation method and device and a readable storage medium, and the method comprises the steps: carrying out the parameter configuration through reading order and vehicle information; carrying out space-time clustering and local path planning on the orders, and merging on-the-way orders; and assigning the orders by using a genetic algorithm, and generating a connection plan of each vehicle / driver. According to the method, the calculation amount is effectively reduced, the solving efficiency is improved, the workload of workers is reduced, and the utilization rate of the connection vehicles is improved.
Owner:HANGZHOU SHUZHIMENG TECH CO LTD

Method for calculating edge drop control efficiency coefficient of 20-roller cold rolling mill using opening degree

The invention discloses a method for calculating the edge drop control efficiency coefficient of a 20-roller cold rolling mill by using the opening degree, and relates to the technical field of cold rolling. A more applicable edge drop control efficiency coefficient calculation model is proposed, and the edge drop control efficiency coefficient is obtained by the opening degree of the first intermediate roll and the strip. The method does not depend on a specific production line, and the edge drop control efficiency coefficient curve obtained has stronger applicability, which can improve the adjustment accuracy of the first intermediate roll of the 20-roller cold rolling mill, thereby quickly and effectively controlling the edge drop of silicon steel and improving the silicon steel finished product rate.
Owner:NORTHEASTERN UNIV CHINA

Transient interaction characteristic analysis method for optical storage direct-flexible system in high-temperature and high-humidity environment

The invention discloses a transient interaction characteristic analysis method for an optical storage direct-flexible system in a high-temperature and high-humidity environment, and the method comprises the steps: firstly building a differential algebraic model of a dynamic behavior, and initializing a state and parameters; then, adopting an SAS method to construct an approximate solution, and dynamically adjusting model parameters according to real-time temperature and humidity; establishing a linear equation of a coupling relationship, recursively solving a high-order coefficient, and calculating residual error evaluation precision; solving the high-order nonlinear model by utilizing a parallelization calculation technology, constructing a local RLC topology to carry out transient simulation, recording a time-varying result of a system state, and analyzing a coupling effect of each component; and finally, adjusting the key variable by adopting a self-adaptive control algorithm. According to the invention, based on the SAS calculation method, for the dynamic response of the optical storage direct-flexible system in the high-temperature and high-humidity environment, efficient transient analysis is realized, and the operation reliability and the regulation and control capability of the system are remarkably improved. The method is applied to an on-line monitoring and control platform of a power system, and real-time optimization is realized.
Owner:STATE GRID SHANGHAI MUNICIPAL ELECTRIC POWER CO +1

Deep learning optimization method based on dynamic domain adaptive adjustment

The invention relates to a deep learning optimization method based on dynamic domain adaptive adjustment. The deep learning optimization method comprises the following steps of performing dynamic domain perception initialization; performing parameter exploration based on gradient disturbance; a dual feedback mechanism; carrying out domain adaptive regularization; adjusting the cyclic domain and evaluating the performance of the model, and if the performance meets the requirement, ending; and if not, continuing to adjust the cyclic domain. According to the method, through dynamic domain perception initialization and parameter exploration based on gradient disturbance, the model can better understand and adapt to different data distributions at the initial stage of training, so that good prediction performance can be kept without data; by introducing a domain adaptive regularization and cyclic domain adjustment strategy, the expression difference between different domains of the model can be remarkably reduced, and the model can be ensured to achieve the optimal performance in various domains; and the dual feedback mechanism allows the model to be self-adjusted according to the real-time effect of the parameters in the training process, so that the training process is more intelligent and efficient, and the resource consumption and the training time are reduced while the model performance is improved.
Owner:ABD SMART EYE ELECTRONICS CO LTD

DCS system internal fault positioning method and system

The invention discloses a DCS system internal fault positioning method and system, and relates to the technical field of fault positioning, and the method comprises the steps: carrying out the abnormal data detection in DCS system event data through combining two types of abnormal data detection algorithms of LOF and OCSVM, taking the intersection of the detection result of the LOF algorithm and the detection result of the OCSVM, and removing the intersection from a DCS system event database; clustering the remaining event data by using a K-Means algorithm in combination with the contour coefficients, calculating the contour coefficients under different clustering number K values for multiple times, and selecting the K value with the highest contour coefficient as the optimal clustering number; arranging the event data in the K clusters after clustering according to an event sequence; and according to the K arranged event sets, troubleshooting one by one. The method and the system for positioning the internal fault of the DCS provided by the invention have good robustness, can accurately position the fault, improve the efficiency and the precision of fault positioning, and improve the stability and the reliability of the system.
Owner:NANJING GUODIAN NANZI WEIMEIDE AUTOMATION CO LTD

Cross-view geographic positioning method and device

The invention discloses a cross-view geographic positioning method and device. Inputting the unmanned aerial vehicle image into a first twin sub-network of the twin network model to obtain an embedded vector feature of the unmanned aerial vehicle image; inputting the satellite image into a second twin sub-network of the twin network model to obtain an embedded vector feature of the satellite image; inputting the unmanned aerial vehicle image embedded vector feature and the satellite image embedded vector feature into an SDCL loss function # imgabs0 # # imgabs1 #, calculating the loss between the unmanned aerial vehicle image embedded vector feature and the satellite image embedded vector feature, training through gradient return, and obtaining an optimal twin network model after the training is finished; and inputting the real-time unmanned aerial vehicle image into the optimal twin network model to obtain a matched satellite image. Symmetrical decoupling comparison learning loss is introduced, effective representation is learned by maximizing the similarity between the positive samples and minimizing the similarity between the negative samples, the problem of imbalance between the positive samples and the negative samples is effectively relieved, the learning ability of the network for the positive samples and the negative samples is improved, and the matching effect of cross-view geographic positioning is improved.
Owner:湖北省地理国情监测中心

Traffic thermal environment three-dimensional image rapid prediction and image display device and method

The invention discloses a traffic thermal environment three-dimensional image rapid prediction and image display device and method. The traffic thermal environment three-dimensional image rapid prediction and image display device comprises a 360-degree rotating camera, a temperature sensor, a wind speed sensor, an image acquisition module, a data processing module and a display. The 360-degree rotary camera is responsible for shooting real-time vehicle condition pictures on a road; the image acquisition module is responsible for carrying out vehicle counting on real-time vehicle condition pictures; the temperature sensor is responsible for collecting real-time air temperature around the road; the wind speed sensor is responsible for collecting real-time wind speed around the road; the data processing module is responsible for processing the acquired data, generating a real-time high-resolution traffic thermal environment dynamic image of the road and judging the average temperature; and the display is responsible for displaying real-time high-resolution traffic thermal environment dynamic images or simultaneously prompting'high-temperature early warning '. The method solves the problem that the traditional technology is limited by the number of monitoring equipment and the use cost, and can be suitable for predicting and displaying any road traffic thermal environment three-dimensional image.
Owner:SOUTHEAST UNIV

A method and system for binocular image super-resolution reconstruction based on cross-scale disparity prior.

ActiveCN116862763BCompact aggregation featuresquality improvementComputer graphics (images)Image resolution
This invention discloses a method and system for binocular image super-resolution reconstruction based on cross-scale disparity prior, comprising: S1: acquiring the original left feature map and the original right feature map corresponding to the low-resolution left and right binocular images; S2: using a binocular attention module to perform cross-view interaction on the original left feature map and the original right feature map respectively, to obtain the interacted left feature map and the interacted right feature map; S3: inputting the interacted left feature map and the interacted right feature map into the cross-scale disparity attention module, and fusing them to obtain an aggregated feature map; S4: inputting the aggregated feature map into a cascaded dynamic upsampling reconstruction network to obtain the reconstructed super-resolution binocular image. This invention can fully utilize disparity information and obtain more realistic and higher-quality super-resolution images in a shorter running time.
Owner:YIBIN GREAT TECH CO LTD

Power grid power failure area prediction method and system under typhoon disaster, storage medium and electronic equipment

The invention provides a power grid power failure area prediction method and system under a typhoon disaster, a storage medium and electronic equipment, and the method comprises the following steps: generating a sample set based on the historical typhoon disaster related state data of a to-be-predicted power grid area; respectively processing the sample set by adopting different dimension reduction modes to generate a corresponding training set; training different machine learning algorithm models by adopting different training sets to obtain a plurality of different prediction models; determining an optimal prediction model according to the evaluation index of each prediction model; and inputting the current typhoon disaster related state data of the to-be-predicted power grid region into the optimal prediction model to obtain a prediction result of whether the to-be-predicted power grid region is powered off or not. The method can effectively predict whether power failure occurs in a power grid area under typhoon disasters, thereby effectively guiding operation and maintenance first-aid repair personnel to carry out maintenance and guarantee work.
Owner:ELECTRIC POWER RES INST OF STATE GRID ZHEJIANG ELECTRIC POWER COMAPNY +3

A deep fake image detection method integrating deep and wide learning

The present invention is a deep forgery image detection method that integrates deep and wide learning. First, the image to be detected is converted into a grayscale image and divided into two groups of image blocks. The centralized frequency domain amplitude spectrum of each image block is extracted and an attention mechanism is applied to it. Then, the two groups of image blocks are channel-connected to obtain primary features. Secondly, a channel convolution autoencoder module is constructed, and the encoder of the pre-trained channel convolution autoencoder module is used to perform feature fusion on the primary features to obtain two intermediate features, which are respectively used as inputs of the feature mapping stream and the feature enhancement stream to obtain two mapping features and enhanced features. Finally, three classifiers are constructed according to the principle of the wide learning system, and the output results of the three classifiers are weighted averaged to obtain the final detection result. This method applies an attention mechanism to the image blocks, and can focus on the areas with obvious tampering traces from the global information. The data and time required for model training are small, and both accuracy and efficiency are achieved.
Owner:HEBEI UNIV OF TECH

A rice seed classification method based on attention residual network

The present invention discloses a rice seed classification method based on an attention residual network. Through the image preprocessing method, the interference of image background on model prediction is effectively avoided, and the influence of objective factors on practical application is reduced. Sufficient effective information is obtained during the sampling process, and feature extraction is more sufficient; drawing on the improved model of the Res2Net network, the attention mechanism is introduced, the depth of the network is increased without changing the number of network layers, the multi-scale feature extraction capability is enhanced, the identification problem of similar varieties is effectively solved, and the information extraction capability is increased without causing a large increase in computing costs.
Owner:ANHUI AGRICULTURAL UNIVERSITY

A small target detection feature enhancement method, system and medium based on cross-scale local contrast modulation

PendingCN122657468AInhibit sync amplificationimprove signal-to-noise ratioPattern recognitionRadiology
The application discloses a small target detection feature enhancement method and system based on cross-scale local contrast modulation and a medium, and aims to solve the problems of blindness and easy amplification of noise in small target feature enhancement. The method comprises the following steps: acquiring shallow high-resolution features and deep low-resolution features; generating a small target priori graph by using the deep features, and constructing a local contrast feature on the shallow features; fusing the priori graph and the contrast feature to generate a modulation coefficient graph; and modulating and enhancing the shallow features by using the modulation coefficient graph and the contrast feature, and sending the enhanced shallow features into a detection head for target detection. The application selectively enhances the small target features by fusing the position priori of deep semantic and the local contrast of shallow physical, improves the detection precision, and suppresses the noise.
Owner:MAINTENANCE BRANCH OF STATE GRID HEBEI ELECTRIC POWER +2

Acceleration method of train-ballast track-roadbed system coupling simulation model

PendingCN121787264Aincrease computational costslow convergenceArtificial lifeDesign optimisation/simulation
The invention discloses a method and a system for accelerating a train-ballast track-roadbed system MBD-DEM-FDM coupling simulation model based on a self-adaptive PID (Proportion Integration Differentiation) algorithm. The method comprises the following steps: firstly, establishing a data interaction interface of MBD and DEM-FDM modules, and calculating and transmitting an acting force and a counter-acting force; a self-adaptive PID control algorithm is introduced, loading force is dynamically corrected according to interface errors, and fast approximation balance in a single time step is achieved; the convergence state of the system is monitored through a sliding window error judgment mechanism, a local fuzzy self-learning or global beetle antenna search optimization algorithm is intelligently triggered according to the convergence state, and PID control parameters are dynamically adjusted in real time and globally optimized. According to the method, a traditional time-consuming space iteration balance process is converted into dynamic feedback adjustment of a time domain, the problems that an existing coupling method is low in calculation efficiency, poor in numerical stability and difficult to apply to a large-scale model are effectively solved, and an efficient and reliable simulation tool is provided for design, operation and maintenance of a railway system.
Owner:NAT ENG LAB FOR HIGH SPEED RAILWAY CONSTR +2

Dynamic gesture recognition method based on lightweight multivariate spatio-temporal convolution

The application provides a dynamic gesture recognition method based on light-weight multivariate space-time convolution, which comprises the following steps: constructing a space feature extraction module based on a pseudo 3D gated attention fusion network; extracting multi-scale space features by using the space feature extraction module; injecting a guided heat map by using a gated attention fusion module to enhance key region features and suppress background interference, so as to obtain a space refinement feature sequence; decomposing the space refinement feature sequence into multiple sub-variables; capturing long-range and local time dependence in parallel by using a modern convolution module; and modeling the intra-variable and inter-variable relationships respectively by using a decoupling feature interaction network, so as to obtain a dynamic gesture recognition result. By combining the multivariate feature decomposition strategy with the modern convolution, the double-branch design advantage of the modern convolution is to balance long-range dependence and local details, solve the limitations of traditional convolution in capturing long-range time dependence, and model complex dynamics.
Owner:JIANGXI UNIVERSITY OF FINANCE AND ECONOMICS

Data feature extraction method based on artificial intelligence

The invention provides a data feature extraction method and system based on artificial intelligence. The method comprises the steps of obtaining heterogeneous data of a plurality of clients; performing feature extraction according to the heterogeneous data to generate a standardized feature set; performing model aggregation processing according to the standardized feature set to determine important features; determining target data in the heterogeneous data according to the important features; performing data preprocessing according to the target data to generate standardized data; determining sensitive information in the standardized data according to the standardized data and a preset deep learning model; and determining target features according to the sensitive information and the standardized data. Reasonable data feature extraction can be performed according to the importance degree of the data while privacy data is protected, so that the processing time of data feature extraction is greatly reduced, and the complexity of data processing and the calculation cost are reduced.
Owner:SHENZHEN TUOBAO SOFTWARE CO LTD