Patents
Literature
74results about How to "Improve recognition rate" 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
Deep learning-based galaxy classification method and system
PendingCN121962690AImprove recognition rateSolve the problem of scale misalignmentCharacter and pattern recognitionBiological modelsImaging FeatureComputer vision
The invention discloses a galaxy classification method and system based on deep learning, and the method comprises the steps: obtaining a large-scale patrolling galaxy image, and carrying out the preprocessing of the large-scale patrolling galaxy image; the preprocessed image is input into the ResCAFMS-Net network; a self-adaptive multi-scale fusion module is adopted to align and fuse multi-scale features output by different semantic levels of the ResNet-18 network; adaptive pooling is carried out on the fused large-scale sky patrol galaxy image features, a classification layer is input, probability distribution of galaxy form categories is obtained, and a final classification result is output; according to the method, the structure density of a galaxy main body in an input image is remarkably improved through self-adaptive cutting preprocessing, and background noise interference is reduced; through the convolution and attention feature fusion module and the adaptive multi-scale fusion module, the problems of large target scale change and fuzzy structure in a large-scale sky patrol galaxy image are effectively solved.
Owner:NANJING TECH UNIV
Photovoltaic junction box defect detection system and method based on coupling of eddy current excitation and infrared thermal imaging
The application discloses a photovoltaic junction box defect detection system and method based on eddy current excitation and infrared thermal imaging coupling technology, and relates to the field of photovoltaic junction box defect detection.The photovoltaic junction box defect detection system comprises an eddy current excitation module, which is used for applying an alternating magnetic field to a photovoltaic junction box to be detected, so that eddy currents are induced in the internal conductors of the photovoltaic junction box, and differential Joule heat responses are formed at the welds in different contact states; an infrared thermal imaging module, which is used for collecting temperature distribution information of the surface of the photovoltaic junction box and generating a thermal image sequence; and an intelligent algorithm module, which is electrically connected with the infrared thermal imaging module and is used for processing the thermal image sequence, wherein the intelligent algorithm module comprises an image preprocessing unit, a multi-scale wavelet decomposition unit, a variable point detection unit and a defect judgment unit connected in sequence.The application can intensify defect signals through active eddy current excitation, and can detect early and mild false-weld defects by using a targeted variable point detection algorithm and being sensitive to weak temperature rise signals.
Owner:CHANGSHA UNIVERSITY OF SCIENCE AND TECHNOLOGY
Picking robot
ActiveCN224139609UDoes not interfere with robotic arm operationsDo not interfere with workManipulatorPicking devicesLaser rangingRobotic arm
The utility model relates to a picking robot which comprises a collecting frame, a movable chassis and a picking mechanical arm. The picking mechanical arm is fixed to the middle of the collecting frame through a stand column. The movable chassis is arranged at the bottom of the collecting frame. The movable chassis is a four-wheel drive walking mechanism with an independent suspension; a first motor and a first joint are fixedly mounted on the side surface of the top of the upright; one end of the first joint is connected with a first motor output shaft, and the other end is connected with a second motor; an output shaft of the second motor is connected with a second joint; the second joint is fixedly provided with a telescopic joint and is movably provided with a connecting rod through the telescopic joint; the steering engine is fixedly mounted at the tail end of the connecting rod; an output shaft of the steering engine is connected with the picker; a first camera and a laser distance measuring sensor are fixedly mounted on one side, close to the steering engine, of the connecting rod; a second camera and a contact sensor are fixedly mounted on the picker tray; and an AI controller is mounted on the picking mechanical arm in an embedded manner. The fruit picking machine has the advantages of being high in fruit recognition rate, low in omission ratio and high in picking efficiency.
Owner:BEIHANG UNIV
Satellite positioning interference identification method and device, electronic equipment and storage medium
PendingCN121995405Aaccurate captureImprove adaptabilitySatellite radio beaconingFeature setRemote sensing
The invention relates to the technical field of satellite positioning interference recognition, in particular to a satellite positioning interference recognition method and device, electronic equipment and a storage medium, and the method comprises the steps: receiving a satellite signal, and carrying out the preprocessing of the satellite signal, and obtaining a preprocessed baseband signal; jointly extracting a phase feature, a carrier-to-noise ratio feature, a power intensity feature, a ranging residual feature, a Doppler change feature, a correlation peak distortion feature and a time-space consistency feature from the preprocessed baseband signal to form a multi-dimensional feature set; screening the multi-dimensional feature set to obtain key features after screening; according to the screened key features, constructing a feature combination vector in a unified format; and identifying the type of the interference signal according to the feature combination vector. Therefore, the problem of incomplete coverage of single-dimensional feature recognition is solved, the interference signal features can be captured more comprehensively and accurately, misjudgment is reduced, and the recognition rate of deception, suppression and other interference signals is improved.
Owner:HUNAN SPACETIME XINAN TECH CO LTD
Unstructured road identification method and system suitable for agricultural unmanned vehicle, medium and terminal
PendingCN121963117AAchieve automatic maintenanceprevent disengagementCharacter and pattern recognitionSpatial perceptionAgricultural engineering
The invention provides an unstructured road identification method and system suitable for an agricultural unmanned vehicle, a medium and a terminal, on one hand, the method does not depend on color threshold segmentation, is not limited by illumination and surface texture change any more, is suitable for scenes with dense trees on two sides of a road, can also adapt to tree-free open scenes, has a very strong generalization ability, and is suitable for popularization and application. The method does not need to depend on the weed coverage condition, color difference and other external characteristics of roads. And on the other hand, the method does not depend on artificial experience like a traditional machine learning technology, can well adapt to various farmland scenes, greatly increases the universality, and greatly improves the recognition rate in a large-range unstructured environment. And finally, large-scale accurate data labeling is not needed, and the generalization performance under different crop, soil texture and weather conditions is excellent. In short, the spatial perception precision can be improved; the influence caused by illumination is avoided; the method is suitable for scenes with many or no trees; the path is prevented from being identified only through trees; and the generalization ability is strong, and no requirement on road characteristics exists.
Owner:HEILONGJIANG HUIDA TECHNOLOGY CO LTD
Live working method and system based on fusion of laser radar and visual recognition
ActiveCN121722123BRealize autonomous obstacle avoidanceAchieve autonomous obstacle avoidanceImage analysisNavigational calculation instrumentsRobotic armEngineering
This invention discloses a method and system for live-line working based on the fusion of lidar and visual recognition. The method includes: transforming lidar point cloud data into coordinates and calculating the weighted height difference using an extrinsic parameter optimization strategy, then performing adaptive weighted fusion with switch images through different feature extraction branches; registering the power equipment point cloud and environmental point cloud in the lidar point cloud data using an IMLS-ICP algorithm based on improved intensity consistency, and using a lidar-visual inertial odometry fusion algorithm for pose estimation and optimization to obtain the globally optimal pose and construct a three-dimensional environment model based on it; using the globally optimal pose as the starting point, the switch position as the ending point, and the three-dimensional environment model as the map, within the safe area defined by electric field simulation on the map, planning an obstacle avoidance path using an improved A* algorithm including electric field strength constraints, and controlling a robotic arm to perform gear shifting operations according to the obstacle avoidance path. This invention improves the safety and efficiency of live-line working.
Owner:HUNAN YONGZHOU ELECTRIC POWER RECONNAISSANCE DESIGN INST +1
A method for identifying a P-wave band burst signal modulation mode
ActiveCN117857269BSolve the problem of difficult to distinguish featuresavoid dependencyFrequency spectrumAlgorithm
The application discloses a P-wave burst signal modulation mode recognition method. The method utilizes the preamble characteristics of the P-wave burst signal, solves the problem that the signal information segment characteristics are difficult to distinguish in a low signal-to-noise ratio environment, and avoids the dependence of the traditional signal preamble autocorrelation recognition algorithm on the signal preamble prior information and the influence of the frequency offset. The method first utilizes the signal preamble spectrum to complete image construction, and utilizes the spectrum characteristics formed by different burst signal preamble rule codes to distinguish; secondly, the multi-semantic feature fusion network fully utilizes the texture features such as signal preamble spectrum edge contour extracted by the low-layer residual network, and fuses the high-level residual network to extract the abstract complex high-level semantic features of the signal spectrum, so as to solve the problem that the residual network only utilizes the high-level abstract semantic features and ignores the low-level features, and improve the burst signal modulation mode recognition performance.
Owner:HANGZHOU DIANZI UNIV
An infrared target detection method and device, electronic equipment and storage medium
ActiveCN115496971Bimprove accuracyIntegration is reasonable and effectiveCharacter and pattern recognitionNeural learning methodsFeature extractionEngineering
This invention discloses an infrared target detection method, apparatus, electronic device, and storage medium, relating to the field of infrared target detection. The method includes: inputting a currently acquired infrared target image into a trained infrared target detection model for category detection; wherein the trained infrared target detection model is determined based on an attention mechanism neural network and a training dataset; the attention mechanism neural network includes a feature extraction network, an attention module connected to the output of the feature extraction network, a feature enhancement module connected to the output of the attention module, and a target classification and detection network connected to the three outputs of the feature enhancement module; the attention module is used to calculate the fusion coefficients of the original feature map output by the feature extraction network to determine a multi-scale feature map. This invention enables low-cost, high-efficiency, and accurate detection of infrared target images.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS +1
A method for identifying a carpet region, a robot and a chip
The application provides a carpet area identification method, a robot and a chip. A first line laser sensor is used to detect the height and distance of an obstacle. When a first preset condition is met, an edge-following mode is started. A second line laser sensor is used to record a grid that meets a second preset condition during edge following. Finally, whether the grid area is a real carpet area is determined by combining the current of the middle brush and the point cloud data obtained by the first line laser sensor. The carpet area identification method is screened three times to determine whether the obstacle detected by the robot is a carpet, greatly improving the carpet identification rate, and without the aid of other sensors such as ultrasonic waves, the cost is reduced.
Owner:AMICRO SEMICONDUCTOR CO LTD
A document handwriting real-time generation comparison tracing method for judicial expertise
PendingCN122510915ASolve the problem of single-dimensional information extractionOvercoming hard-to-discern flaws
This invention relates to the field of image data processing technology and discloses a real-time generation, comparison, and tracing method for document handwriting for forensic identification. The method involves performing a two-dimensional optical scan of the sample to be identified, calculating the paper-based fiber topology matrix; performing physical medium classification and calculating the microscopic spatial quantization feature sequence; calling a multi-state physical model for inverse mapping to reconstruct instantaneous velocity and acceleration time-series signals; performing frequency domain analysis on the acceleration time-series signals to extract biological tremors and mechanical harmonic energy to calculate the initial screening judgment factor and locate the core physical frequency of vibration; combining the core frequency to extract the actual spatial physical resonance component and generate a theoretical vibration waveform; performing a space-frequency dual-domain closed-loop verification to obtain the maximum phase cross-correlation coefficient; and finally, combining the judgment factor and cross-correlation coefficient to perform logical judgment and output the identification conclusion and parameters of the forged device. This invention can effectively distinguish between genuine handwriting and high-precision mechanical imitation, achieving accurate tracing of handwriting evidence.
Owner:BEIJING HUIZHENG EXCELLENCE TECHNOLOGY CO LTD
One-dimensional code recognition method and device, electronic equipment and storage medium
ActiveCN115983303BImprove recognition rateimprove accuracyNeural learning methodsSensing by electromagnetic radiationAlgorithmTheoretical computer science
The application provides a one-dimensional code recognition method and device, an electronic device and a storage medium. The method comprises the following steps: performing multi-scale recognition on a one-dimensional code image obtained to obtain a first recognition result; recognizing a character region in the one-dimensional code image to obtain a second recognition result; and performing result fusion on the first recognition result and the second recognition result through a fusion algorithm to obtain a recognition result of the one-dimensional code image; the recognition rate of the one-dimensional code when the one-dimensional code is polluted or incomplete can be effectively improved; meanwhile, the one-dimensional code recognition is effectively verified, and the error rate of a scanning result is reduced.
Owner:FUZHOU ROCKCHIP SEMICON
Emergency help signal panoramic projection device, method and terminal
ActiveCN121768121Bimprove intelligibilityIncrease the chance of being rescuedPicture reproducers using projection devicesElectric transmission signalling systemsPanoramaRadiology
The application belongs to the technical field of emergency help equipment, and provides an emergency help signal panoramic projection device, method and terminal, the device can continuously detect the spatial posture and rotating motion state of itself through the environment positioning module, generate corresponding posture data and direction data, the control module can obtain the data and determine the real-time projection pointing angle of the device, and then when the device rotates, the projection control instruction corresponding to the real-time projection pointing angle is generated and sent to the image projection module, so that the image generation unit in the image projection module drives the corresponding rotating transformation of the help image, and the light source in the image projection module irradiates the help image to project the visual help signal which is right and complete. The scheme realizes omnidirectional controllable intelligent projection, can make the help signal cover all directions, at the same time ensures that the projected image can be clearly observed, is more easily identified by the rescuer, reduces the delay caused by signal blur, and improves the probability of being rescued by the help seeker.
Owner:SHENZHEN DOUG HENGTONG TECH CO LTD
Vegetable dehydrating equipment with infrared light and vision fused and foreign matter removing method
PendingCN121847462AImprove quality and safetyIncrease the level of automationSortingImaging processingEngineering
The invention discloses infrared light and vision fused vegetable dehydrating equipment and a foreign matter removing method, and relates to the technical field of food processing, sorting and detection. The device is composed of a feeding and conveying assembly, a vertical plate arranged on the side of a transmission belt, an electric control box arranged on the vertical plate, an infrared light source assembly, a visible light source assembly, corresponding lenses, a limiting baffle, a slidable nozzle assembly, a waste tank and the like. An image processing unit in the electric control box is used for registering and fusing to identify the foreign matters, the time and the transverse position of the foreign matters reaching the nozzle position are calculated by combining the running speed of the transmission belt, and the nozzle assembly is driven to transversely align along the limiting baffle and spray at a fixed point to blow the foreign matters into the waste tank, so that the foreign matters in the dehydrated vegetables are efficiently, accurately and automatically removed; missing detection and mistaken elimination are reduced, and the product quality and the production automation level are improved.
Owner:临沂田源食品有限公司
Dynamic effect identification method and related equipment
PendingCN121879702AImprove recognition ratereduce loadCathode-ray tube indicatorsInput/output processes for data processingView basedOphthalmology
The invention relates to a dynamic effect identification method and related equipment. The method is applied to the electronic equipment. The electronic device displays a first interface in a display screen, and executes a dynamic effect in response to an operation of a user on a first control of the first interface. The electronic device also determines a position change value of at least one view of the dynamic effect based on the transfer matrix of the dynamic effect, and determines a dynamic effect speed of the dynamic effect based on the position change value and a first screen refresh rate of the display screen. According to the method, the position change value of at least one view is determined based on the conversion matrix, and the dynamic effect speed of the dynamic effect is determined based on the position change value and the first screen refresh rate, so that the technical problem that the view dynamic effect corresponding to the user-defined dynamic effect class cannot be accurately identified is solved, and the identification rate of the dynamic effect is improved.
Owner:HONOR DEVICE CO LTD
Method and apparatus for determining a recognition area, and electronic device
ActiveCN116246061BImprove recognition rateAvoid the problem of not being able to recognize the user's body movementsInput/output for user-computer interactionCharacter and pattern recognition
Owner:GUANGZHOU LEAFUN CULTURE SCI & TECH
Iterative updating method, device and system of voice application, vehicle and server
The invention discloses an iterative updating method, device and system for a voice application, a vehicle and a server, and relates to the technical field of voice recognition, the method is applied to a rail transit vehicle, and the method comprises the following steps: after the current voice application is started, obtaining an original voice data stream collected by a vehicle-mounted pickup; performing voice recognition on the original voice data stream by using a voice recognition model of the current voice application to obtain a voice recognition result; generating and storing a model iteration data set according to the voice recognition results and the recognition voice data corresponding to each voice recognition result; sending the model iteration data set to a cloud server; according to the method, the data required for updating the subsequent voice recognition model in the rail transit vehicle can be automatically acquired, so that data support is provided for iterative updating of the voice recognition model, rapid iterative updating of voice application in a rail transit scene is realized, and the engineering landing speed of a rail transit voice intelligent product is increased.
Owner:CRRC ZHUZHOU ELECTRIC LOCOMOTIVE RESEARCH INSTITUTE CO LTD
Electric power engineering supervision and shooting method and system based on artificial intelligence
PendingCN121960934Aensure unique bindingEnsure the performance of duties is in placeData processing applicationsDigital data authenticationFeasibility studyElectric power equipment
The invention belongs to the technical field of power equipment monitoring, and particularly relates to an artificial intelligence-based power engineering supervision and monitoring method and system, and the method comprises the steps: obtaining an authorized technical improvement and overhaul project list from a power grid resource business platform in response to successful login, building the correlation between a supervision and monitoring device and a target project, and carrying out the supervision and monitoring of the target project; downloading feasibility research report structured data corresponding to the project, an associated power equipment asset code and power industry mandatory regulation text data of a specialty to which the project belongs; according to the invention, through double authentication of face living body detection and passwords in the step S1, unique binding of'human-machine-account 'is ensured, and false use of the identity is completely eradicated; and S3, Beidou differential positioning and anti-metal electronic tag scanning are creatively fused, the positioning precision is improved to a meter level / sub-meter level in a 500kV strong electromagnetic interference environment, whether a supervisor accurately arrives at an equipment operation point or not is directly checked, the recognition rate of virtual post behaviors such as card punching and regional wandering is improved, and supervision is guaranteed to perform in place from the source.
Owner:SHANDONG LUNENG SOFTWARE TECH
Speech recognition method, apparatus and electronic device
PendingCN122266364AImprove recognition rateGuaranteed accuracySpeech recognitionSpeech synthesisPersonalizationAcoustic model
The application discloses a speech recognition method, device and electronic equipment, the method sets a hot word database containing multiple hot words in advance, and hot word database is searched and biased decoding is carried out based on the speech feature representation of speech data in the recognition stage, and accurate recognition of personalized and low-frequency hot words is realized. By so setting, the defects that the traditional shallow fusion mode is insufficient to correct the internal bias of the acoustic model and is easy to affect the general recognition performance are overcome, the hot word recognition rate is effectively improved, and the accuracy and stability of the overall recognition system are ensured.
Owner:BEIJING YUNSHANG TECH CO LTD
Tire surface barcode correction method and device based on density segmentation
ActiveCN121962178AImprove recognition rateGood segmentation effectImage enhancementImage analysisImaging processingData set
The invention relates to the technical field of image processing, in particular to a tire surface barcode correction method and device based on density segmentation. The method comprises the steps of obtaining an initial barcode image; acquiring a bar code feature data set based on the initial bar code image, determining a bar code segmentation boundary position based on the bar code feature data set, and segmenting the initial bar code image based on the bar code segmentation boundary position to obtain a plurality of bar code area images; and performing correction processing based on the bar code area image to obtain a corrected bar code image. In this way, the uniformity of the bar code can be effectively recovered, and then the bar code recognition rate is remarkably improved.
Owner:ZHEJIANG UNIV OF TECH
Gas pipeline leakage judgment method and storage medium
InactiveCN121953258AImprove recognition rateThe principle is simpleMaterial analysis by optical meansPipeline systemsBackground concentrationsPhysical chemistry
The invention discloses a gas pipeline leakage judgment method and a storage medium. The method comprises the steps that S1, adaptive background concentration extraction is carried out; for the data points one by one, taking the median in the sub-window as an initial background concentration value, and further correcting the background concentration by adopting the change condition of the initial background concentration in the main window to obtain a background concentration value adapted to the data points one by one; s2, judging a methane source; the methane source is judged by comprehensively considering the ratio of methane to ethane enhancement value, methane and ethane change synchronism and the ethane change condition at the methane peak value. The storage medium is realized based on the method. The method has the advantages of simple principle, wide application range, easiness in implementation, high detection precision and the like.
Owner:CHANGSHA GAS IND CO LTD +1
Model bag laying quality evaluation method and device based on image recognition, equipment and storage medium
PendingCN121767276Acost savingReduce the need for manual interventionImage analysisBiological modelsImage segmentationMechanical engineering
The invention discloses a mold bag laying quality evaluation method, device and equipment based on image recognition and a storage medium, and relates to the technical field of mold bag laying quality evaluation, and the method comprises the steps: obtaining an orthoimage of a target area mold bag laying region shot by shooting equipment; carrying out automatic identification and image segmentation on the orthoimage, and determining boundary information and overlap edge information of a mold bag unit in the orthoimage; positioning a defect part according to the boundary information and the overlap edge information; and according to the boundary information and the lap joint edge information, measuring the lap joint width between the adjacent mold bags, and according to the defect part and the lap joint width, completing mold bag laying quality evaluation. Through the image recognition technology, automatic quality evaluation of the mold bag laying area is realized, and the evaluation efficiency and accuracy are improved. Compared with a traditional manual evaluation mode, the method has the advantages that the defect part can be positioned more quickly, and the overlapping width between the adjacent mold bags can be accurately measured, so that powerful support is provided for controlling the laying quality of the mold bags.
Owner:CCCC SHEC FOURTH ENG
Long-tail Federated Learning Method and Apparatus Based on Gradient Cleanup and Residual Learning
This invention proposes a long-tail federated learning method and apparatus based on gradient cleanup and residual learning. The method includes: the server selecting clients and distributing global parameters; each client performing local training based on the global parameters to obtain updated local shared parameters, which are then uploaded to the server. During training, the task gradient and the balanced alignment gradient are calculated, and the task gradient is cleaned by projection to remove components opposite to the balanced alignment gradient, resulting in a cleaned gradient. The server aggregates the local shared parameters uploaded by each client to obtain global shared parameters and determines whether the current number of communication rounds has reached a preset limit. If so, the global shared parameters are distributed to all clients for personalized training to obtain client-specific models. During personalized training on each client, global branches and personalized branches are used to improve the long-tail category recognition rate.
Owner:XIAMEN UNIV
A method and system for diagnosing faults in a control valve
This invention relates to the field of industrial automation control technology, specifically disclosing a fault diagnosis method and system for control valves. The method includes: acquiring operating data of the control valve, including pressure data, temperature data, and vibration acoustic data; extracting features from the vibration acoustic data, obtaining features including root mean square value, peak value, peak factor, spectral centroid, energy ratio of a specific frequency band, and main direction of the sound source; wherein, based on a dynamic adaptive beamforming algorithm, spatial filtering and anti-interference mechanisms are introduced to calculate the main direction of the sound source; inputting a fault diagnosis model to calculate the fault type and fault probability; when a fault occurs, adjusting the valve according to the operating data. This invention can accurately identify various control valve faults, and the classification accuracy of cavitation faults reaches over 92%, making it particularly suitable for cavitation fault identification of electro-hydraulic linkage control valves under harsh operating conditions such as high pressure, large diameter, large pressure difference, and high precision regulation.
Owner:TIANJIN AOMEI AUTOMATION SYST
A complex building classification extraction method based on a U-Net model
ActiveCN115017968Baccurate segmentationImprove Segmentation AccuracyInternal combustion piston enginesCharacter and pattern recognitionPattern recognitionData set
The application discloses a kind of based on U-Net model's complex building classification extraction method, including with steps:Step1.Getting unmanned aerial vehicle image data;Step2.Make building classification dataset.Step3.Using classification building dataset respectively trains U-Net model, obtains building extraction model;Step4.Enhance dataset;Step5.Using enhanced dataset, U-Net model is fine-tuned training, obtains fine-tuned building extraction model;Step6.Model test: the building image of prediction dataset is input to step Step5 obtains fine-tuned building extraction model, and classification is identified to building.The application can be under the condition that building dataset is limited, to complex building type is effectively identified extraction, the building contour after segmentation is clear, building is complete, can exist in the form of single house, effectively improve the extraction accuracy of complex building.
Owner:GUIZHOU NORMAL UNIVERSITY
A two-dimensional code recognition method, device, storage medium, terminal and vehicle
PendingCN122347159AImprove recognition rateEnhance texture detailsComputer graphics (images)Image resolution
Owner:BYD CO LTD
Method for identifying motion state of artificial limb wearer
PendingCN121845561Acancel noiseImprove signal processing capabilitiesSensorsDiagnostic recording/measuringNeighbor algorithmProsthetic limb
The invention discloses a method for identifying the motion state of an artificial limb wearer. The method comprises the following steps: 1, acquiring a signal; 2, preprocessing the initial signal based on wavelet transform; 3, extracting characteristic values of the signals based on wavelet packets; and 4, classifying and identifying the motion state based on a K-nearest neighbor algorithm. According to the invention, automatic identification of the motion state of the artificial limb wearer is realized, and the identification rate and real-time performance of motion state identification of the artificial limb wearer are effectively improved.
Owner:BEIJING XINGJIAN CHANGKONG MEASUREMENT CONTROL TECH
Speech expression multi-modal emotion recognition method based on weighted feature fusion
This invention discloses a multimodal emotion recognition method for speech and facial expressions based on weighted feature fusion, comprising: extracting speech keyframes and facial expression keyframes, and extracting MFCC features from the speech keyframes; aligning the extracted speech and facial expression keyframes, and calculating speech weights and facial expression weights in the weighted feature fusion by utilizing the complementarity of speech and facial expression modalities; extracting higher-order speech emotion features from the speech MFCC features using a trained residual network, and extracting higher-order facial expression emotion features from the facial expression keyframes using a trained 3D-CNN neural network; weighting the higher-order speech emotion features and higher-order facial expression emotion features using the speech weights and facial expression weights respectively, to obtain weighted higher-order speech emotion features and weighted higher-order facial expression emotion features; inputting the weighted higher-order speech emotion features and weighted higher-order facial expression emotion features into a multilayer perceptron (MLP) for weighted feature fusion and emotion classification. This invention can improve the recognition rate of multimodal emotion recognition for speech and facial expressions.
Owner:CENT CHINA BRANCH OF CHINA DATANG CORP SCI & TECH RES INST CO LTD
Encrypted traffic identification method, apparatus, device, and storage medium
ActiveCN115801435BFine recognitionfine predictionSecuring communicationHigh level techniquesFeature extractionEngineering
The application provides an encrypted traffic identification method and device, equipment and a storage medium, wherein the method comprises: identifying traffic to determine encrypted traffic, and sending the encrypted traffic to a corresponding hierarchical identification model for feature extraction and traffic source prediction, which can more finely identify and predict different types of encrypted traffic, effectively improving the identification rate of encrypted traffic and reducing the misidentification rate. Moreover, in the application, a hierarchical identification model is used to extract features from encrypted traffic, and identification is performed according to the features, avoiding the problem that existing identification technologies cannot be applied to all encrypted traffic, and improving the universality of different types of encrypted traffic.
Owner:中孚安全技术有限公司
Method and system for improving license plate recognition rate of raw fuel truck
ActiveCN115690761BImprove recognition rateAutomatically correct recognition errorsRoad vehicles traffic controlCharacter and pattern recognitionTransportation technologyTruck
This invention discloses a method and system for improving the license plate recognition rate of raw material and fuel trucks, belonging to the field of intelligent transportation technology. The system includes: a checkpoint unit and a vehicle management module; each checkpoint unit includes one main capture camera and zero to multiple auxiliary capture cameras, each capture camera including one main recognition algorithm and zero to multiple auxiliary recognition algorithms; the vehicle management module generates a dynamic license plate number database for each type of checkpoint unit; the checkpoint unit acquires the license plate number recognized by the capture cameras and compares the license plate number with the dynamic license plate number database corresponding to the checkpoint unit to complete the recognition process; the vehicle management module also performs self-learning on the recognition results for use in the next recognition process. It employs multiple capture devices, multiple recognition algorithms, and fuzzy judgment by the vehicle management module, maximizing the recognition rate and reducing manual intervention.
Owner:WISDRI ENG & RES INC LTD
A Multimodal Emotion Recognition Method and System for Voice and Facial Expressions of People in Smart Buildings
ActiveCN115273905BIncrease the loss functionincrease differentiationSpeech analysisBiological modelsNetwork modelFacial expression
This invention discloses a method and system for multimodal emotion recognition of speech and facial expressions in smart buildings. The method includes: preprocessing collected speech information and facial expression videos, including: extracting MFCC features from the speech information; segmenting the facial expression video into frames and extracting facial expression regions from the facial expression images; inputting the MFCC features into a residual network model to extract high-order emotion features from the speech information; inputting the segmented facial expression images and the extracted facial expression regions into a 3D-CNN model to extract high-order emotion features from the facial expression images; and using an MLP model to perform multimodal fusion of the high-order emotion features of the speech information and facial expression images to identify the emotional state of the person. During the fusion process, a pre-designed loss function is used to complete end-to-end training of the multimodal feature fusion of speech and facial expression images, and to increase the difference between the high-order emotion features of the speech information and facial expression images. This invention can improve the recognition rate of multimodal emotion recognition of speech and facial expression images.
Owner:CENT CHINA BRANCH OF CHINA DATANG CORP SCI & TECH RES INST CO LTD