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

19results about How to "Improve classification ability" patented technology

Standard section automatic identification method based on 3D ultrasonic data

The invention discloses a standard section automatic identification method based on 3D ultrasonic data, and belongs to the technical field of medical imaging, and the method comprises the following steps: obtaining a 2D ultrasonic standard section image and 3D ultrasonic data; the obtained 2D standard section image is preprocessed, and corresponding parameters are set according to classification of different ultrasonic acquisition devices for batch automatic cutting; and meanwhile, automatic cutting is realized by adopting a maximum connected domain method, so that interference of a pure color background and an equipment interface is removed, and pure standard section data is obtained. According to the method, anatomical information is provided by utilizing 3D ultrasonic data, the neural network is guided to realize efficient standard section automatic identification, interference of a background and an equipment interface is removed by designing an automatic cutting method and a 2D data expansion method, a large number of 2D slice images are generated, and the richness of training data is increased. And a VGG16 and OC-CNN design network is combined, and the preprocessed and expanded 2D image is trained to obtain a model for automatically identifying the standard section.
Owner:WUHAN CHILDRENS HOSPITAL

Improved vehicle and pedestrian tracking method based on Deep-sort

The invention discloses a vehicle and pedestrian tracking algorithm recommendation method based on Deep-sort improvement. Relates to the field of recommendation methods, an improved Deep-Sort serves as a tracker, due to the fact that a tracking object of a Deep-Sort algorithm is a person, a research object of the invention is a vehicle and an appearance model is not applicable, the improved Deep-Sort improves a feature extraction network of the Deep-Sort in three aspects: the feature extraction network is rebuilt based on ResNet-34, so that the feature extraction network is more suitable for extraction of vehicle features; a center loss function and a triple loss function are introduced and are combined with a Softmax loss function for use, so that intra-class and inter-class distances of the network are better optimized; and a cosine annealing learning rate attenuation method is used, so that the final convergence effect of the network training model is better. The improved feature extraction network is trained, and experimental results show that the vehicle appearance model obtained through training has good classification capability, and the recognition capability of vehicle features is improved compared with an original model. In a multi-moving vehicle tracking experiment, the D-MMVT algorithm is tested by using an OTB-100 data set and a daily monitoring road video, the accuracy rate MOTA, the precision rate MOTP and the ID switching frequency are used as evaluation standards, a comparison experiment is performed with other tracking algorithms, and an experiment result shows that the D-MMVT algorithm has a better tracking effect on a severely shielded target and a small target; and the tracking accuracy and precision are improved.
Owner:NANJING TECH UNIV

RecycleNet-based sea surface target detection and classification method and system

The invention discloses a RecycNet-based sea surface target detection and classification method and system, and belongs to the technical field of radar target classification. The RecycNet-based sea surface target detection and classification method comprises the following steps: acquiring a sea surface echo amplitude sequence by using a radar, and processing the sea surface echo amplitude sequence by using a recurrence plot construction technology to obtain recurrence plot data; the method comprises the following steps: constructing a RecycNet network; training a RecycNet network by using the recurrence plot data, and adjusting network parameters; performing classification by using the trained RecycNet network to obtain a classification result; compared with the traditional convolutional neural network, the RecycNet can more efficiently capture and recover the detail characteristics of the input data, so that the network shows stronger detection and classification capabilities under the complex background of sea clutters, and the detection precision and robustness are remarkably improved.
Owner:NANJING UNIV OF POSTS & TELECOMM

DC fault arc state monitoring method and device, and computer program product

The invention provides a direct-current fault arc state monitoring method and device and a computer program product, and the method comprises the steps: sampling an electrical signal of a monitored load, and obtaining a current and voltage signal sequence; an electrical characteristic signal sequence is extracted by adopting a wavelet packet decomposition and reconstruction technology, when the load type is unknown, similarity measurement is carried out through electrical characteristic sample sequences of various preset types of loads, and the load type is determined by utilizing a deep belief network model. And when the type of the load is known, fault analysis is carried out through the similarity measurement value, the second deep belief network model is utilized to output a DC fault arc state monitoring result, and whether the load is in normal operation or has a DC fault arc is judged, so that the monitoring accuracy and efficiency are improved.
Owner:SHENZHEN POWER SUPPLY BUREAU

Method for realizing experiment teaching behavior analysis based on deep convolutional neural network

The invention discloses a method for realizing experiment teaching behavior analysis based on a deep convolutional neural network, and the method comprises the steps: collecting physical, chemical and student experiment teaching data, carrying out the preprocessing, and segmenting a physical, chemical and student experiment teaching process; extracting a plurality of features from the physical, chemical and biological experiment teaching data, and performing multi-modal feature fusion to form comprehensive feature representation; constructing a deep convolutional neural network model by using the fused features; training a deep convolutional neural network model, adjusting model parameters, and identifying teaching behaviors; analyzing the model identification result, and evaluating the frequency and duration of different teaching behaviors and the reaction and participation degree of the students; and constructing a description framework, converting a model identification result and evaluation data into a report, and outputting the report to a teacher. According to the method, the deep learning technology is combined, multi-modal data is comprehensively analyzed, key features are extracted, modal information is integrated through the feature fusion technology, a deep convolutional neural network model is constructed, and therefore teaching improvement and strategy adjustment are supported.
Owner:JIANGSU TECHTRANS IOT CO LTD

Multi-label image classification model training method, application method and related system

The invention discloses a multi-label image classification model training method, an application method and a related system, and relates to the technical field of image classification processing, and the training method comprises the steps: obtaining a to-be-classified sample image; performing feature extraction on the sample image to obtain image features; calculating Euclidean distance-based feature neighborhoods and Hamming distance-based label neighborhoods of the sample images on each category representation so as to respectively determine feature maps and label maps of the sample images on each category representation; respectively determining an aggregation feature represented by each category on the feature graph and an aggregation feature represented by each category on the tag graph; inputting into a classifier of a multi-label image classification model for classification; a knowledge distillation technology is adopted, prediction response of a label graph is extracted from classification loss to serve as supervision knowledge which is used for guiding learning of a feature graph, and a trained model is obtained after training optimization. According to the invention, the accuracy of multi-label image classification can be improved.
Owner:NANJING UNIV OF SCI & TECH

Production identification screening method and system based on machine-made sand dry method

PendingCN121786717AReduce the differential impact of different featuresImprove the efficiency of production quality assessmentData processing applicationsEnsemble learningProcess engineeringIndustrial engineering
The invention relates to the technical field of ensemble learning, in particular to a production identification screening method and system based on a machine-made sand dry method, and the method comprises the steps: obtaining a first vector of each placer particle sample; obtaining a hierarchical clustering tree, and obtaining an integrated weight of a base classifier of each dimension in a dimension set of each class cluster of each hierarchy of the hierarchical clustering tree; obtaining a base classifier of each level of the hierarchical clustering tree in each dimension; obtaining an integrated classifier of each level of the hierarchical clustering tree; obtaining a classification error rate of an integrated classifier of each level of the hierarchical clustering tree; obtaining the comprehensive weight of the integrated classifier of each level; obtaining a total classifier of the training set; and obtaining the production quality of the ore sand particles to be analyzed. According to the method, the difference influence of different characteristics of the ore sand particles is reduced, and the efficiency and accuracy of production quality evaluation of the to-be-analyzed ore sand particles are improved.
Owner:GANSU HUAJIAN NEW MATERIALS CO LTD

A welding machine fault diagnosis method and system

ActiveCN114169414BImprove classification abilityDatabase modelsButt weldingData set
The present invention discloses a welding machine fault diagnosis method and system, the method comprising the following steps: collecting welding machine data under various fault states and constructing a database; preprocessing the welding machine data in the database and generating a training data set; training a hidden Markov model using the training data set to generate a fault state model corresponding to different fault states; and monitoring the welding machine state in real time based on the fault state model. The present invention can determine the welding machine fault location and fault mode by analyzing and calculating the welding machine fault state model and the real-time collected data, so as to facilitate technicians to determine a fault solution.
Owner:JIANGSU AUTOMATION RESEARCH INSTITUTE +1

Respiratory pathogen detection kit and method

The invention relates to the technical field of respiratory pathogen detection, in particular to a respiratory pathogen detection kit and method.According to the respiratory pathogen detection kit and method, pathogen sample data are obtained and subjected to multiple detection through a data collection module, and PCR data, immune data and microarray data are generated; the detection target analysis module fuses PCR data, immune data and microarray data to generate real-time target fusion data, and predicts pathogen types and analyzes feature importance by using a decision tree algorithm; the customized detection module adopts a targeted data fusion strategy according to a prediction result and selects a corresponding detection method, so that the whole-process individuation from data processing to detection method selection is realized, the accuracy, efficiency and pertinence of detection are remarkably improved, and the method is particularly suitable for rapid diagnosis of complex and newly-onset pathogens and has a wide application prospect. And a reliable basis is provided for clinical treatment.
Owner:BEIJING KANGRUN CHENGYE BIOLOGICAL TECH CO LTD

Data-free federated bidirectional knowledge distillation method and system based on feature adversarial

ActiveCN118627644BImprove classification abilityimprove performanceMachine learningEngineeringData mining
The present invention provides a data-free federated bidirectional knowledge distillation method and system based on feature adversarial, which relates to the field of federated learning technology in machine learning. The method includes constructing a federated learning framework, combining multiple clients and servers, using a global generator to generate pseudo data for enhanced training, and combining adversarial training between local feature discriminators and global generators to promote bidirectional knowledge distillation between global models and local models. During the iteration process, the global model continuously integrates local knowledge, and the local model also obtains the classification ability from a global perspective, ultimately achieving high performance and stability of the global model. The present invention is suitable for scenarios with uneven data distribution.
Owner:JIANGNAN UNIV +1

Rock damage identification methods and systems, electronic equipment

PendingCN122087546AStrong characteristicImprove classification abilityOriginal dataEngineering
This invention relates to the field of machine learning, providing a method, system, and electronic device for rock damage identification. The method includes: acquiring multi-stage damage data of a rock under test, wherein the multi-stage damage data is acquired through a piezoelectric sensing element disposed on or inside the surface of the rock under test; constructing an original dataset based on the multi-stage damage data; and inputting the original dataset into a pre-constructed convolutional neural network model to obtain the damage identification result of the rock under test. This invention addresses the shortcomings of related technologies, such as the easy loss of key damage details like local spectral deformation and peak shift, enabling real-time, non-destructive, and continuous monitoring of rock damage. It automatically extracts damage features and accurately classifies damage at each cycle stage from initial loading to final failure, providing reliable technical support for early warning of geotechnical engineering disasters.
Owner:NANCHANG UNIV

Deep convolutional neural network learning-based hearth water-cooled wall anomaly detection and identification method

The invention discloses a hearth water wall anomaly detection and identification method based on deep convolutional neural network learning, and relates to the field of hearth water wall anomaly detection. In order to solve the technical problems of inaccurate identification and inaccurate classification of detection, identification and classification of the water cooling wall of the cooling pipe in the prior art, the technical scheme provided by the invention is as follows: the method comprises the following steps of: obtaining original pipeline data shot at each pipeline position in a hearth water cooling wall system of a thermal power plant; preprocessing the original pipeline data, and dividing the original pipeline data into a training data sample and a test data sample; respectively carrying out hybrid enhancement processing on the training data sample and the test data sample; and inputting the training data sample subjected to hybrid enhancement processing into a deep convolutional neural network for model training to obtain a trained model. The method can be applied to abnormal detection and identification of the hearth water wall.
Owner:DATANG DONGBEI ELECTRIC POWER TESTING & RES INST

Channel pruning method for visual understanding model of classification task and application

The invention belongs to the technical field of visual understanding, and discloses a channel pruning method and application of a visual understanding model for classification tasks, and the method comprises the steps: employing a divergence-based channel importance function to calculate the importance value of each channel in a trained visual understanding model; wherein the divergence-based channel importance function is a product of a standard deviation of sample distribution corresponding to each channel output feature map in the visual understanding model and a variable coefficient, and the variable coefficient is equal to a ratio of the standard deviation of the sample distribution to a mathematical expectation of the sample distribution; the standard deviation of the sample distribution is used for reflecting the dispersion degree between the samples; and sorting the importance values of the channels, and pruning the corresponding channels from the channels with low importance values until the parameters and operand required by the visual understanding model are met, thereby obtaining the pruned lightweight visual understanding model. According to the method, the classification performance of the model can be improved under a relatively high pruning rate.
Owner:HUAZHONG UNIV OF SCI & TECH

A deep learning classification model and its application in phenotypic detection of maize with multiple growth periods

The present invention discloses a deep learning classification model and its application in phenotypic detection of corn during multiple growth periods. The deep learning classification model constructs a global hierarchical data structure with the synergy of a multi-layer feature fusion mechanism and an attention mechanism as the core; the steps of performing corn phenotypic detection based on the model include: constructing a high-throughput image database of multiple growth periods based on multi-environment data sources of corn diversity populations; image preprocessing; image classification data set construction; deep learning classification model construction; and deep learning classification model is used for corn phenotypic classification or detection. The present invention can realize real-time, rapid, and non-destructive low nitrogen stress detection in multiple growth periods of corn.
Owner:HUAZHONG AGRI UNIV

A method for verifying laboratory data based on ensemble learning and changing decision threshold

The present invention discloses a method for verifying laboratory data based on ensemble learning and changing decision thresholds, and belongs to the field of big data. The present invention combines clustering algorithms, the idea of ​​ensemble learning and several methods of changing decision thresholds together to obtain a new strong classifier, and learns based on existing laboratory data to obtain a classification model, and pre-classifies new laboratory data to increase its accuracy, thereby solving the technical problem of effectively verifying the classification accuracy of laboratory data. The present invention has stronger classification ability by combining four different single classifiers, and for the problem of data imbalance in disease data sets, in the final decision stage, by changing the decision threshold, the accuracy of classification is improved, and the classification results are evaluated from multiple angles through multiple indicators, so as to better understand the accuracy of the classification results.
Owner:THE FIRST AFFILIATED HOSPITAL OF ZHENGZHOU UNIV

Array type iris face multi-mode acquisition and recognition system

The invention discloses an array type iris face multi-mode acquisition and recognition system, which belongs to the technical field of biological feature recognition and comprises a main control board and a multi-core processor adopting an ARM framework. The main control board is connected with a binocular face enhancement camera, a binocular distance measurement camera, 16 groups of iris acquisition modules, 16 groups of near-infrared light supplement lamps, a three-color LED indicating lamp, a touch display screen and a real-time living body detection and anti-fraud module based on deep learning. According to the method, the iris features and the face features are deeply fused by adopting a multi-dimensional feature fusion and matching algorithm, so that comprehensive and accurate identity feature vectors are formed. The iris has extremely high uniqueness and stability, the face feature provides an additional identity information dimension, and the fusion of the iris and the face feature greatly improves the accuracy of identity recognition. Meanwhile, the real-time living body detection and anti-fraud module based on deep learning is trained by a large number of real samples and fraud samples, and has strong feature extraction and classification capabilities.
Owner:SHANGHAI IRISIAN OPTRONICS TECH CO LTD

Cell classification method, system, cell graph Transformer model training method and electronic device

The present invention relates to a cell classification method, system, cell graph transformer model training method and electronic equipment, belonging to the field of cell classification technology, including: acquiring a cell image; inputting the cell image into a cell segmentation model to obtain a cell segmentation map or inputting the cell detection model to obtain a cell detection map; inputting the cell image into the feature extractor to obtain a feature map; using the cell graph construction module to generate a cell map according to the feature map and the cell segmentation map or generating a cell map according to the feature map and the cell detection map; using the cell graph symbolization module to generate a symbol vector according to the cell map; using the transformer encoder to obtain a cell classification according to the symbol vector. The present invention makes full use of the powerful classification performance of the transformer encoder and the global topological relationship information of the cell nucleus of the cell image to improve the accuracy of cell classification.
Owner:SHENZHEN RES INST OF BIG DATA

Method and system for detecting and positioning faults of narrow-frame TFT-LCD (Thin Film Transistor-Liquid Crystal Display) display unit

The invention provides a method and a system for detecting and positioning faults of a narrow-frame TFT-LCD (Thin Film Transistor-Liquid Crystal Display) display unit. Belongs to the technical field of display. The method comprises the following steps: carrying out image acquisition on a narrow-frame TFT-LCD display unit through a high-resolution camera; carrying out the preprocessing of the collected image; based on an image processing technology, fault features existing in a display unit are extracted, pixels in an image are subjected to point-by-point analysis by setting a specific algorithm, abnormal pixel points or areas are identified, and feature information is extracted. Through a high-resolution camera and a fine image processing technology, detail information of the display unit can be captured, and accurate extraction of a fault area is ensured.
Owner:SHENZHEN GILANTE TECH CO LTD

Mechanical fault diagnosis model migration training method, mechanical fault diagnosis method, device, equipment and medium

ActiveCN119474865AImprove classification abilitystrong discriminationBiological modelsMachine learningModel migration
The invention relates to a mechanical fault diagnosis model migration training method and device, a mechanical fault diagnosis method and device, equipment and a medium. The training method comprises the steps of obtaining a mechanical fault diagnosis model, a source domain sample set and a target domain sample set; outputting a sample feature and a fault prediction result of the source fault sample and a sample feature and a fault prediction result of the target fault sample by using the mechanical fault diagnosis model; determining sample classification loss according to the fault prediction results of the plurality of source fault samples and the sample labels; determining feature difference loss according to the sample features of the plurality of source fault samples and the sample features of the plurality of target fault samples; determining category confusion loss according to the fault prediction results of the multiple target fault samples; and determining target loss according to the sample classification loss, the feature difference loss and the category confusion loss, and optimizing the mechanical fault diagnosis model by using the target loss. Therefore, the accuracy of mechanical fault type diagnosis of the trained target mechanical fault diagnosis model can be improved.
Owner:ELECTRIC POWER RESEARCH INSTITUTE OF STATE GRID JIBEI ELECTRIC POWER CO LTD +1