Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
42results about How to "Optimize network parameters" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
End-to-end semantic instant-positioning and graph building method based deep learning
ActiveCN108665496AOptimize network parametersImprove performanceImage enhancementImage analysisPattern recognitionImage sequence
The invention discloses an end-to-end semantic instant-positioning and graph building method based on deep learning. A continuous original image sequence and an original three-dimensional point cloudsequence corresponding to the image sequence are respectively collected and obtained through a color camera and lidar, and pose transformation information, depth information and semantic segmentationinformation of continuous five-image sequences are obtained by processing; and a multi-task deep neural network with branches is constructed, the sequences are input into the multi-task deep neural network, and the multi-task deep neural network is trained for obtaining parameters, and the trained multi-task deep neural network is adopted to process a consecutive to-be-tested five-image sequence to obtain pose transformation information, depth information and semantic segmentation information among images. Compared with a traditional ORB-SLAM algorithm and methods similarly based on deep learning, the method of the invention has better performance.
Owner:ZHEJIANG UNIV
SAR image target classification method based on NSCT double CNN channels and selective attention mechanism
ActiveCN107358258AImprove classification performanceOptimize network parametersCharacter and pattern recognitionMass centerHigh frequency
The invention discloses an SAR image target classification method based on NSCT double CNN channels and a selective attention mechanism. The method comprises steps that training sample sets D1 and D2 for target detection and classification are acquired; the D1 and D2 are expanded to acquire sample sets D3 and D4; models M1 and M2 for target detection and classification are respectively trained; significance detection and morphological processing on test images are carried out, connected domain marking is further carried out, target candidate areas corresponding to a connected domain mass center are extracted, translation in multiple surrounding pixel points is carried out, and the target candidate areas are generated; classification determination of the target candidate areas is carried out through utilizing the M1, and accurate positioning of a target is acquired; a final class of the target is determined through voting decision after M2 classification. The method is advantaged in that a non-down-sampling contour wave layer is added, low frequency and high frequency characteristic images are inputted to a double channel CNN to form the NSCT double channel CNN, the selective attention mechanism is applied to SAR image classification, SAR image target detection classification accuracy is improved, and a problem of low target classification accuracy in the prior art is solved.
Owner:XIDIAN UNIV
Space-time bone characteristic and depth belief network-based human body behavior identification method
InactiveCN108537145AImprove accuracyImprove recognition efficiencyCharacter and pattern recognitionDeep belief networkCharacteristic length
The invention belongs to the field of computer vision and specifically relates to a space-time bone characteristic and depth belief network-based human body behavior identification method. According to the method, a bone sequence is obtained via a depth camera based on each motion; bone point time sequence characteristics are extracted from the bone sequence and comprise displacement characteristics, speed characteristics and acceleration characteristics of each bone point; a bone reference point is chosen, relative distance between each bone point and the bone reference point is calculated, space characteristics of a motion bone sequence is obtained, the bone point time sequence characteristics and the space characteristics of a motion bone sequence are subjected to clustering operation,a characteristic length is fixed, a global motion description operator is established and subjected to dimensionality reducing operation, and motion classification is finished based on cross validation and a classifier.
Owner:NORTHEAST DIANLI UNIVERSITY
Edge detection network optimization method, and pavement disease identification method and system
ActiveCN107871133AOptimize network parametersThe recognition result is accurateImage enhancementImage analysisDiseaseEdge detection
The embodiment of the invention discloses an edge detection network optimization method, a pavement disease identification method and a pavement disease identification system. An edge detection network is used for identifying the specified pavement disease. The optimization method comprises the steps of segmenting each sample pavement picture into a plurality of first grids; obtaining a first matrix corresponding to each sample pavement picture according to the fact whether the first grids possess the specified pavement disease; inputting each sample pavement picture to the edge detection network for identification of the specified pavement disease, and outputting a second matrix corresponding to each sample pavement picture; calculating to obtain a loss function of each sample pavement picture according to vectors of the first matrix and the second matrix corresponding to each sample pavement picture; and optimizing a network parameter of the edge detection network until the second matrix output by the edge detection network with optimized network parameter makes an arithmetic average value of the calculated loss functions of all sample pavement pictures minimum. According to theembodiment, an identification result of the edge detection network is accurate and efficient.
Owner:ROADMAINT CO LTD
Method for identifying continuous stirred tank reaction process based on deep neural network
ActiveCN108920888AOptimize network parametersRegistering/indicating during manufacturing processNeural architecturesNerve networkEngineering
A method for identifying a continuous stirred tank reaction process based on a deep neural network comprises the following steps: step (1) of obtaining process variable data during continuous stirredtank reactor operation; step (2) of performing data pre-processing on the collected process variable data: first needing to standardize the data; selecting time lag, and organizing the process variable into a three-dimensional input form; and finally, dividing the data into a training set, a verification set and a test set; step (3) of establishing, a three-dimensional long-term and short-term memory nerve network, an identification model and training: utilizing a memory unit to establish a three-dimensional long-term and short-term memory nerve network model, and determining a network structure and hyperparameters; utilizing an adaptive moment estimation algorithm to optimize network parameters on the training set, selecting the hyperparameters of the network model on the verification set, and establishing the identification model based on the three-dimensional long-term and short-term memory nerve network and performing training. According to the method, online monitoring on a process state is performed, and the precise identification of the product concentration is achieved.
Owner:ZHEJIANG UNIV OF TECH
Aircraft trajectory prediction method based on long short-term memory network
ActiveCN114048889ASimplify complexityOptimize network parametersSustainable transportationForecastingSimulationTerm memory
The invention relates to the fields of air combat environments, data processing, deep learning and the like, and provides a method for realizing aircraft trajectory prediction by using a long short-term memory (LSTM) network under an uncertain sensing condition. Therefore, the technical scheme adopted by the invention is as follows: the method for predicting the trajectory of the aircraft based on the long-short-term memory network comprises the following steps of: eliminating noise interference carried by a sensor feature vector by using Kalman filtering; data preprocessing including downsampling, invalid value elimination and missing value complementation is carried out on the directly obtained state parameters, in addition, in order to improve the calculation stability, data is subjected to normalization processing, and the value range of input data is included in the interval of [0, 1]; and an LSTM-based trajectory prediction model is created, input and output of the network is defined, and the network is supervised and trained. The invention is mainly applied to the prediction occasion of the flight path of the unmanned aerial vehicle.
Owner:TIANJIN UNIV
Method and system for locating random access failure reason of user equipment
InactiveCN102244886AImprove performanceOptimize network parametersWireless communicationFailure causesUser equipment
The invention provides a method and a system for locating a random access failure reason of User Equipment (UE). The method comprises the following steps that: the UE reports a random access parameter employed by the UE in a failed random access process to the network side; and the network side locates the reason of the random access failure of the UE according to the random access parameter reported by the UE. The system comprises the User Equipment (UE) and a network side apparatus, wherein the UE is used for reporting a random access parameter employed by the UE in a failed random access process to the network side apparatus; and the network side apparatus is used for locating the reason of the random access failure of the UE according to the random access parameter reported by the UE. By employing the method and the system provided by the invention, the reason of the random access failure of the UE can be located conveniently, so that the network performance can be effectively and quickly optimized by the network side.
Owner:TD TECH COMM TECH LTD
Self-supervised learning fine-grained image classification method based on twin network
PendingCN114676777AOptimize network parametersMaintain class consistencyCharacter and pattern recognitionNeural architecturesImaging FeatureSupervised learning
The invention discloses a self-supervised learning fine-grained image classification method based on a twin network. According to the method, firstly, an attention encoder is used for extracting deep convolution features of an image, an attention graph containing semantic information is obtained, and image features are encoded in a bilinear pooling mode; secondly, positioning a salient region where a high response value is located on the attention map from the original image, cutting and erasing the salient region so as to form views of different view angles, and learning view angle invariance features in a self-supervision mode; and finally, combining the center loss function and the consistency loss function to display and constrain different view angle characteristics, and keeping the intra-class consistency of the different view angle characteristics. The method enables the network to obtain significant performance gain, and can significantly improve the classification accuracy on the baseline of fine-grained image classification.
Owner:INST OF SOFTWARE - CHINESE ACAD OF SCI
Method for optimizing grid-related parameters for power generator unit based on double-stranded quantum genetic algorithm
ActiveCN106777521AImprove transient stabilityHigh speedDesign optimisation/simulationSpecial data processing applicationsQuantum gateGlobal optimal
The invention discloses a method for optimizing grid-related parameters for a power generator unit based on a double-stranded quantum genetic algorithm. The method comprises the following steps: selecting the grid-related parameters required for the optimization of the power generator unit in a power grid system; calculating constraint conditions of the grid-related parameters required for the optimization of the power generator unit and total objective functions with optimal transient stability; determining fitness functions of the grid-related parameters for the power generator unit; initializing a double-stranded quantum population; judging whether a current chromosome is mutant or not by utilizing the mutation probability and performing quantum bit non-gate mutation in case of mutation; converting the probability amplitude of each quantum bit of the current chromosome into a solution space, determining a value-substituted transient stability calculation program obtained by conversion into a total objective function value with optimal transient stability and performing fitness evaluation to determine individual fitness and store a global optimal solution; calculating a quantum rotation angle advance step length and updating a quantum gate to obtain a next generation of chromosome; by optimizing the grid-related parameters for the power generator unit, the transient stability of the system is improved, and the machine-grid coordination is realized.
Owner:JIANGSU ELECTRIC POWER CO +2
Rain image rain stripe removing method and system based on image filtering and CNN
ActiveCN111815526AEasy to trainFast executionImage enhancementNeural architecturesData setAtmospheric sciences
The invention provides a rain image rain stripe removing method and system based on image filtering and CNN. According to the method, a rain image data set is expanded and trained by using a rain image data set enhancement method, multiple groups of rain images and label images are acquired, the rain images are subjected to image filtering to obtain rain image high-frequency parts, the rain imagehigh-frequency parts are input into a rain removal network to obtain rain removal images, and an SSIM loss function is added to optimize the rain removal network; the rain removal image is input intothe H-G discrimination network to obtain a discrimination result, the discrimination result is fed back to the rain removal network to further improve the rain removal quality, and the problems that an existing rain removal method is difficult to accurately describe a rain model, rain stripes cannot be completely removed, and details are easy to lose are solved.
Owner:CHINA UNIV OF GEOSCIENCES (WUHAN)
Remote sensing image thin cloud removal method and system based on full-wave band feature fusion
ActiveCN114066755AEnhance thin cloud removal abilityThin cloud removal with high precisionImage enhancementImage analysisWave bandImage resolution
The invention discloses a remote sensing image thin cloud removal method and system based on full-band feature fusion. The method comprises the following steps: carrying out multispectral influence thin cloud removal on a multispectral remote sensing image to be processed by using a trained thin cloud removal network; wherein the trained thin cloud removal network is obtained through the following steps: obtaining multispectral remote sensing images in the same region under the condition of clouds and no clouds, and obtaining a training set and a test set; sampling to obtain spatial features and spectral features of spectral bands with different resolutions of the image; carrying out feature fusion to respectively obtain feature maps of the images under the cloud condition and the cloud-free condition; calculating multi-path supervision loss, and optimizing network parameters of a preset thin cloud removal network; and training and testing the optimized thin cloud removal network by using the training set and the test set to obtain a trained thin cloud removal network. The method is high in thin cloud removal precision and small in error, compared with the prior art, the removal training is greatly improved, and the method has a wide application space in multispectral remote sensing images.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
Multi-focus image fusion method based on multi-scale feature interaction network
ActiveCN113705675AAvoid lossImprove accuracyCharacter and pattern recognitionNeural architecturesConditional random fieldInformation transmission
The invention discloses a multi-focus image fusion method based on a multi-scale feature interaction network. The method comprises the following steps: 1, performing channel connection on a pair of color multi-focus source images, and inputting the color multi-focus source images into the designed multi-scale feature interaction network (MSFIN) to generate a focus image representing pixel focus attributes; 2, binarizing the focused image (namely setting a pixel value higher than a threshold value as 1, and otherwise, marking the pixel value as 0) to obtain an initial decision diagram, and refining the initial decision diagram by adopting a full-connection conditional random field (FC-CRF) to remove some misclassified small regions to obtain a final decision diagram; and 3, based on the refined final decision diagram, adopting a pixel weighted average strategy to obtain a fused image. The images with different focus areas can be effectively fused, a new solution is provided for fusion of the multi-focus images, the fused image contains enhanced spatial information, and information transmission and image processing are facilitated.
Owner:HEFEI UNIV OF TECH
Attention-based model training method and device and electronic equipment
PendingCN114494814AOptimize network structureOptimize network parametersNeural architecturesNeural learning methodsAlgorithmImage detection
The invention provides an attention-based model training method and device and electronic equipment, relates to the technical field of artificial intelligence, in particular to the technical field of deep learning and computer vision, and can be applied to scenes such as image processing and image detection. The specific implementation scheme is as follows: obtaining an attention output matrix of an attention module in a neural network model, performing dimension reduction calculation on a sample dimension and a data block dimension of the attention output matrix based on a pooling layer of the neural network model, and determining a pooled first output matrix; carrying out convolution operation on the first output matrix based on a convolution layer of a neural network model, determining a second output matrix after convolution, and carrying out normalization processing and weighting processing on an output value of each header in the second output matrix to obtain an updated second output matrix; and obtaining an updated attention output matrix based on the updated second output matrix, and training the neural network model based on the updated attention output matrix.
Owner:BEIJING BAIDU NETCOM SCI & TECH CO LTD
Neural network training method, image processing method and device, equipment and medium
PendingCN112016576AOptimize network parametersCharacter and pattern recognitionNeural architecturesSample graphData set
The embodiment of the invention discloses a neural network training method, an image processing method and device, equipment and a medium. The method comprises steps of carrying out the splicing of sample images, and obtaining a spliced sample image; extracting feature data of the spliced sample images through a first neural network; splitting the extracted feature data according to a rule of splicing the sample images to obtain feature data of each spliced sample image; determining a prediction result of each image processing task through a first neural network according to the feature data of each sample image; according to the determined prediction result of each image processing task and the corresponding annotation result of each sample image, obtaining an annotation result. Accordingto the method, the network parameters of the first neural network are adjusted, training is carried out through the mixed data set, the network parameters of the first neural network are optimized through different data sets, and the first neural network with the optimized parameters can process a plurality of image processing tasks at the same time.
Owner:ZHEJIANG SENSETIME TECH DEV CO LTD
Multi-source ciphertext image retrieval method based on federated learning and secret sharing
ActiveCN113987255AOptimize network parametersSimple structureDigital data information retrievalRandom number generatorsSecret shareFederated learning
The invention discloses a multi-source ciphertext image retrieval method based on federated learning and secret sharing. The method comprises the following steps that: S1, model training is performed on a convolutional neural network of a double-cloud platform by taking an image owner joining the double-cloud platform as a federated member based on federated learning; and S2, the authorized user completes ciphertext image retrieval by means of the double-cloud platform based on additive secret sharing. According to the scheme, a multi-source ciphertext image retrieval scheme based on federated learning and secret sharing is provided, a neural network model structure used by federated learning simplification retrieval is utilized to obtain better network parameters, better neural network parameters and a simpler network model structure are obtained through overhead of the image owner end, a better convolutional neural network can be used in ciphertext image retrieval, and a user can obtain a better result faster during retrieval.
Owner:NANHU LAB
Pulmonary nodule intelligent diagnosis method based on mixed characteristics
InactiveCN110570405AQuickly distinguish between benign and malignantImprove generalization abilityImage enhancementImage analysisPattern recognitionPulmonary nodule
The invention discloses a pulmonary nodule intelligent diagnosis method based on mixed characteristics, and the method comprises the steps: obtaining CT image data which comprises a chest medical image file and a corresponding diagnosis result lesion mark; carrying out resampling, smoothing processing and normalization processing on the acquired CT image data; learning a 3D CT image by using a 3Dresidual-tight connection network to obtain high-dimensional depth features, and obtaining LBP-based texture features and HOG-based shape features for describing characterization features of pulmonarynodules; and classifying benign and malignant pulmonary nodules by adopting a GBM gradient elevator based on the high-dimensional depth features, the LBP-based texture features and the HOG-based shape features. According to the method, benign and malignant pulmonary nodules can be quickly distinguished in CT medical images, the misjudgment rate is reduced, the classification precision is improved, and the generalization ability of a diagnosis network model is enhanced.
Owner:TIANJIN UNIV
Model processing method, related equipment, storage medium and computer program product
ActiveCN113569824BReduce loss valueOptimize network parametersCharacter and pattern recognitionNeural architecturesPattern recognitionVideo processing
The present application discloses a model processing method, related equipment, storage medium and computer program product. The method includes: extracting the first video segment and the second video segment from the sample video, and the distance between the first video segment and the second video segment The spatio-temporal overlap rate is equal to the reference spatio-temporal overlap rate; the video processing model is called to extract the spatio-temporal feature of the first video clip as the first spatio-temporal feature, and extract the spatio-temporal feature of the second video clip as the second spatio-temporal feature; according to the first spatio-temporal feature and the second spatio-temporal feature Two spatio-temporal features, predict the spatio-temporal overlap rate between the first video segment and the second video segment, and obtain the spatio-temporal overlap rate prediction result; calculate the model loss value based on the reference spatio-temporal overlap rate and the spatio-temporal overlap rate prediction result; according to the reduction model The direction of the loss value, optimizing the network parameters of the video processing model. This application can enhance the spatio-temporal representation capability of the model, so that the model can construct more accurate video spatio-temporal fingerprints.
Owner:TENCENT TECH (SHENZHEN) CO LTD
Lithium ion battery health state prediction method based on CRJ network
PendingCN113376541AReal-time online prediction of health statusReduce occupancyElectrical testingArtificial lifeLithium electrodeReal-time computing
The invention relates to a lithium ion battery health state prediction method based on a CRJ network. The method comprises the steps: forming the CRJ network and training the CRJ network; and monitoring a battery constant-current charging time sequence on line, inputting the battery constant-current charging time sequence into a prediction model, and outputting an available discharging capacity sequence to obtain the health state of batteries. According to the method, the constant-current charging time is used as input, the health state is predicted through the CRJ network, and real-time online prediction of the health state of the lithium ion batteries is achieved; and the method has low requirements on hardware conditions and occupies small memory. The prediction model established after the CRJ network is optimized by adopting an optimization algorithm can be used for health state prediction of batteries of the same type; an IPSO algorithm and an AOA algorithm are combined to form an IAPSOA algorithm, and the IAPSOA optimizing algorithm enhances the search capability and stability of the AOA algorithm, so that network parameters can be better optimized; and the accuracy of obtaining a CRJ network model is high.
Owner:LIAONING TECHNICAL UNIVERSITY
Rodent identification and analysis method based on deep learning and transfer learning
ActiveCN113283306AOptimize network parametersQuick checkGeometric image transformationCharacter and pattern recognitionPattern recognitionImaging processing
The invention discloses a rodent identification and analysis method based on deep learning and transfer learning, and belongs to the technical field of animal classification and image processing. The method comprises the following steps: S1, carrying out rodent video acquisition and preprocessing; S2,making a rodent data set; S3, constructing and training a rodent model; S4,carrying out rodent detection; S5, carrying out rodent image cutting and super-resolution reconstruction; S6, segmenting each part of the rodent image; S7, carrying out data cleaning and calculation of all body index data; and S8, classifying rodents. The invention has the beneficial effects that detection, classification and segmentation of each part of the rodent are realized based on a deep learning method, and each body index data of the rodent is calculated, so that the rodent detection accuracy and efficiency are effectively improved, the traditional rodent classification mode is changed. And each part of the rodent can be automatically segmented, so that animal protection and pest prevention and control work can be favorably carried out.
Owner:青岛云智环境数据管理有限公司
Model processing method, related equipment, storage medium and computer program product
ActiveCN113569824AImprove feature learning abilityEnhanced spatio-temporal representation capabilitiesCharacter and pattern recognitionNeural architecturesEngineeringAlgorithm
The invention discloses a model processing method, related equipment, a storage medium and a computer program product, and the method comprises the steps: intercepting a first video clip and a second video clip from a sample video, and the space-time overlapping rate between the first video clip and the second video clip is equal to a reference space-time overlapping rate; calling a video processing model, extracting a spatial-temporal feature of the first video clip as a first spatial-temporal feature, and extracting a spatial-temporal feature of the second video clip as a second spatial-temporal feature; according to the first spatiotemporal feature and the second spatiotemporal feature, predicting a spatiotemporal overlapping rate between the first video clip and the second video clip to obtain a spatiotemporal overlapping rate prediction result; calculating a model loss value based on the reference space-time overlapping rate and the space-time overlapping rate prediction result; and optimizing network parameters of the video processing model according to the direction of reducing the loss value of the model. According to the method, the space-time representation capability of the model can be enhanced, so that the model can construct more accurate video space-time fingerprints.
Owner:TENCENT TECH (SHENZHEN) CO LTD
A performance optimization method and system for cognitive network
ActiveCN112333804BIncreased probability of interruptionReduced probability of outagePower managementTransmission monitoringStatistical analysisEnergy estimation
The embodiment of the present invention proposes a cognitive network-oriented performance optimization method and system. Through the joint optimization scheme of spectrum sensing, power allocation, and energy loss, the cooperation technology is adopted to effectively improve the cognitive network interruption performance and reduce the cognitive network performance. Outage probability, with the aim of improving the energy efficiency of cognitive networks. Wherein, the system implementing the method includes: a spectrum sensing module, an interruption performance optimization module, a transmission performance optimization module, and an energy consumption optimization module. In the energy consumption optimization module, through the statistical analysis of the energy in the cognitive network, as well as the distribution probability and energy estimation of the nodes, the optimal number of hops and the appropriate number of secondary users are selected to achieve the goal of optimizing network energy consumption.
Owner:江苏移动信息系统集成有限公司 +1
Three-dimensional point cloud classification method fusing persistent coherence
PendingCN114581718AImprove accuracyPerfect structured descriptionCharacter and pattern recognitionNeural architecturesNetwork modelConvolution
The invention belongs to the technical field of computer vision, particularly relates to a three-dimensional point cloud classification method fusing persistent coherence, and aims to solve the problem that point cloud topological features are not represented in the current three-dimensional point cloud classification technology. And further introducing a persistent coherence method in algebraic topology to carry out point cloud classification. The method comprises the following steps: firstly, constructing a point cloud witness simple complex topological structure, and quantifying the persistent coherent topological features of the point cloud from two aspects of Betti number and persistent graph; secondly, defining a loss function based on persistent coherence, and performing training learning on the three-dimensional point cloud classification network model according to the loss function to obtain various parameters of the neural network model; and finally, performing a classification task of the three-dimensional point cloud by using the trained convolutional neural network. Test results show that the point cloud classification accuracy is remarkably improved.
Owner:ZHONGBEI UNIV
Method and system for removing rain streaks in rainy images based on image filtering and CNN
ActiveCN111815526BEasy to trainFast executionImage enhancementNeural architecturesAtmospheric sciencesRain removal
The present invention provides a method and system for removing rain stripes from rainy images based on image filtering and CNN. The present invention uses a rain image data set enhancement method to expand the training rain image data set to obtain multiple sets of rainy images and label images. The rainy images pass through The high-frequency part of the rainy image is obtained by image filtering, and the high-frequency part of the rainy image is input into the rain-removing network to obtain the rain-removing image, and the SSIM loss function is added to optimize the rain-removing network; the rain-removing image is input into the H-G discriminant network to obtain the discrimination result. The discrimination results are fed back to the deraining network to further improve the quality of deraining, and solve the problems that the existing deraining methods are difficult to accurately describe the rain model, the rain streaks cannot be removed cleanly, and the details are easily lost.
Owner:CHINA UNIV OF GEOSCIENCES (WUHAN)
Image change detection method based on CNN-CDCN
PendingCN112418282AMinimize Contrast ErrorOptimize network parametersCharacter and pattern recognitionNeural architecturesFeature vectorFeature Dimension
The invention discloses an image change detection method based on CNN-CDCN, and relates to the field of machine learning, and the method comprises the following steps: constructing a convolutional neural network structure, and defining a change detection problem; preprocessing data, removing noise of corresponding images, reducing feature dimensions through a convolution layer and a pooling layer,removing redundant information, performing multi-layer convolution and pooling to finally obtain feature vectors capable of representing input images, and converting the two input images into the same feature space through a coupling layer to enable feature representation of the two input images to be more consistent; and finally, learning network parameters by optimizing a target function. According to the invention, the convolutional neural network is utilized to extract the local features of the image, feature conversion is carried out, an unsupervised method is adopted, additional prior information is not needed, and the invention is an autonomous learning intelligent algorithm.
Owner:无锡禹空间智能科技有限公司
Self-inspection method for camera mechanism of robot, electronic equipment and storage medium
PendingCN112836711ASave resourcesSimplified Computational ComplexityCharacter and pattern recognitionNeural architecturesComputation complexitySimulation
The invention relates to a self-inspection method for a camera mechanism of a robot, electronic equipment and a storage medium, and the method comprises the steps of collecting an image when the camera mechanism is shielded in a preset scene, and carrying out the preprocessing of the image, thereby forming a training set; constructing a self-checking network model of the camera mechanism, wherein the self-checking network model of the camera mechanism comprises multiple layers of sensing units; and training a camera mechanism self-inspection network model by adopting images of the training set, optimizing network parameters of the camera mechanism self-inspection network model, thereby adopting a robot camera shielding self-inspection strategy based on a neural network algorithm, and meanwhile, a mode of combining a convolution module and a multi-layer perceptron are used for simplifying the calculation complexity of a shielding detection module, and in combination with business scene characteristics, the shielding self-checking module is dynamically called, so that robot resources are saved.
Owner:SHANGHAI YOGO ROBOTICS CO LTD
Multi-source ciphertext image retrieval method based on federated learning and secret sharing
ActiveCN113987255BOptimize network parametersSimple structureDigital data information retrievalRandom number generatorsCiphertextImage retrieval
This scheme discloses a multi-source ciphertext image retrieval method based on federated learning and secret sharing, which includes the following steps: S1. Based on federated learning, the image owner who joins the dual cloud platform is a federation member to perform convolutional neural network analysis on the dual cloud platform. Network for model training; S2. Authorized users complete ciphertext image retrieval based on additive secret sharing with the help of dual cloud platforms. This solution provides a multi-source ciphertext image retrieval solution based on federated learning and secret sharing, uses federated learning to simplify the neural network model structure used in retrieval, obtains better network parameters, and exchanges the cost of the image owner for better neural network parameters With a more streamlined network model structure, better convolutional neural networks can be used in ciphertext image retrieval, allowing users to get better and faster results when searching.
Owner:NANHU LAB
A Classified Intelligent Extraction Method for Hydropower Station Reservoir Scheduling Rules
ActiveCN109359671BEffectively distinguish hidden featuresFully excavatedCharacter and pattern recognitionArtificial lifeLearning machinePower station
The invention discloses a classification intelligent extraction method for hydropower station reservoir dispatching operation rules, comprising: using the power station output as an output variable, using a correlation analysis method to determine the input variable; obtaining the normalized input variable and output variable, using a clustering method The input variables of all samples are divided into K categories; for the input variables and output variables under each category, the corresponding ELM models are respectively constructed for simulation and approximation, and the improved particle swarm optimization algorithm is used to optimize the parameters of the ELM model, thereby obtaining K different ELM model: Determine the category of the newly acquired input variable, and input it into the corresponding model to obtain the corresponding output value. Perform denormalization processing to obtain the output value of the power station for dispatching decisions. The invention adopts the classified evolutionary extreme learning machine model to extract the scheduling rules of the hydropower station reservoirs, which can significantly improve the long-term operation benefits of the hydropower station reservoirs, and is beneficial to the efficient utilization of water energy resources of cascade hydropower station groups in the river basin.
Owner:HUAZHONG UNIV OF SCI & TECH
Intelligent amplitude comparison angle measurement method based on neural network, storage medium and equipment
ActiveCN111929674AExact fit mappingHigh precisionNeural architecturesNeural learning methodsData transformationData set
The invention discloses an intelligent amplitude comparison angle measurement method based on a neural network, a storage medium and equipment. The method comprises the steps of: setting a radiation source and a sampling experiment at intervals, and completing data collection through the frequency point and angle measurement frequency and the measurement amplitudes and preset angles of four antennas as the known angle information of data; carrying out noise data elimination and data conversion on the collected data; designing a multilayer perceptron neural network model; dividing the processeddata into a training data set and a verification data set, training the neural network model by using the training data set, and minimizing a loss function; performing performance verification on thetrained neural network model by using the verification data set; and estimating the angle of arrival of an actual radiation source by applying the verified neural network model. According to the invention, the traditional theory of amplitude comparison angle measurement is reasonably combined with an advanced neural network intelligent calculation method, and more accurate radiation source anglemeasurement can be realized.
Owner:XIDIAN UNIV
Finger vein verification method, electronic equipment and storage medium
PendingCN111832419ABig amount of dataImprove robustnessDigital data authenticationSubcutaneous biometric featuresData expansionRadiology
The invention discloses a finger vein verification method, electronic equipment and a storage medium, which are applied to a U-Net network architecture, and the method comprises the steps: obtaining afinger vein image, and carrying out the ROI extraction of the finger vein image; extracting vein lines through a method for detecting the local maximum curvature of the cross section of the finger vein, and obtaining a segmented image; carrying out differential calculation processing on any two segmented images to obtain a differential image so as to carry out data expansion; carrying out channelcopying on the difference image and then taking the difference image as input of a pre-training stage, and reserving an optimal weight of the pre-training stage as a pre-training weight of a cascadeoptimization stage pair; and performing channel superposition on the differential image and the two original segmented images to serve as input of the cascade optimization stage, optimizing parametersof a pre-trained network, and obtaining an optimized finger vein verification model.
Owner:WUYI UNIV
A Rolling Bearing Fault Diagnosis Method Based on Improved Model Migration Strategy
ActiveCN111721536BAchieve migrationAvoid gradient non-convergenceMachine part testingCharacter and pattern recognitionData setFrequency spectrum
Owner:HARBIN UNIV OF SCI & TECH
Features
- R&D
- Intellectual Property
- Life Sciences
- Materials
- Tech Scout
Why Patsnap Eureka
- Unparalleled Data Quality
- Higher Quality Content
- 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com