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

41results about How to "Avoid Gradient Explosion" patented technology

T-S fuzzy and TD3-based slow dynamic unknown coal-fired power generation system coordination control method

The invention discloses a T-S fuzzy and TD3-based slow dynamic unknown coal-fired power generation system coordination control method, which comprises the following steps: firstly, decomposing a coal-fired power generation system into a fast subsystem and a slow subsystem by using a singular perturbation theory, and decomposing an original control task into a stabilization task of the fast subsystem and a tracking task of the slow subsystem; for the fast subsystem, a fuzzy set is selected on a definition domain of the fast subsystem by using the slow variation characteristic of steam pressure, and a T-S fuzzy model is constructed; a fast subsystem controller is obtained by solving algebraic Riccati equations corresponding to a plurality of linear systems. For a slow subsystem, a double-Q network is adopted to reduce overestimation of a Q value, a dynamic learning rate and a batch size adjustment mechanism are introduced to accelerate training convergence, and learning is carried out under a TD3 algorithm framework to obtain control input of the slow subsystem. The fast subsystem and the slow subsystem control inputs to be combined and then act on the original system to obtain state information at the next moment, and the intelligent agent interacts with the original system to complete collaborative optimization.
Owner:CHINA UNIV OF MINING & TECH

Universal multi-level fusion attention model for urban coast SAR image ship detection

The invention relates to a universal multi-level fusion attention model for urban coast SAR image ship detection, which comprises the following steps: establishing sub-modules DFEM, CMAFM and SMAFM: firstly designing the DFEM to extract rich features with different resolutions and semantic information for multi-scale ship detection; then designing CMAFM and SMAFM to fuse channel attention blocks and space attention blocks on different levels of feature maps so as to better learn dependency features from channel and space dimensions to enhance feature representation; and finally, inputting the fused feature map into an existing ship detection network to obtain a detection result. According to the method, more important part of information in an input image can be focused, the position information of the ship is enhanced, unimportant feature information is weakened, the model performance is improved through a small number of additional parameters, and the precision of identifying large and small ships under the complex urban coast background is improved.
Owner:ZHEJIANG COLLEGE OF SECURITY TECH +2

Urban road network traffic state prediction method, device, equipment and medium

The invention discloses an urban road network traffic state prediction method and device, equipment and a medium, and relates to the technical field of traffic, and the method comprises the steps: obtaining the basic parameters of a target urban region and the target traffic state data of a historical time period; determining road network density data according to the basic parameters of the target city area by using a road network density calculation formula; inputting the target traffic state data of the historical time period into the trained dynamic space-time residual convolutional neural network model to obtain traffic state prediction data of the next time period of the historical time period; the dynamic space-time residual convolutional neural network model comprises a residual convolution module, a dynamic convolution module based on road network density, a space-time attention mechanism module and a fusion module which are connected in sequence; and the weight corresponding to each dynamic convolution kernel in the dynamic convolution module based on the road network density is adjusted according to the road network density data. The prediction efficiency and accuracy are improved.
Owner:XIAN TECH UNIV

Face and tongue coat identification and tea drink recommendation method and system based on artificial intelligence

The invention relates to the field of image recognition or understanding, in particular to a face and tongue coat recognition and tea drink recommendation method and system based on artificial intelligence. The method comprises the following steps: identifying a plurality of image annotation labels, training a feature extraction model and a classification model by using the annotated images, obtaining a new image, performing feature extraction by using the trained feature extraction model, and inputting the feature into the trained classification model to obtain a label of the new image. The device is used for applying the method. The existing method and system for identifying the face and the tongue coating of the tea making machine and recommending the tea making have the problem that the identification of a face image and a tongue coating image is not accurate enough. The method for identifying the face and the tongue coating of the tea making machine and recommending the tea making based on the artificial intelligence provided by the invention is accurate in identification of the face image and the tongue coating image.
Owner:XI AN PENGPAIYUEDONG ELECTRONIC TECH CO LTD

Multi-mode agricultural pest and disease prediction method

The invention provides a multi-modal agricultural disease and pest prediction method. The method comprises the following steps: acquiring image data and audio data by using a sensor, respectively preprocessing the image data and the audio data, extracting semantic features of the image data by using a CNN model, extracting MFCC features of the audio data by using a Mel system, and splicing and fusing the semantic features and the MFCC features to obtain fused features; the fusion features are input into a BiLSTM model according to time steps, an output result is obtained through training, the output result is input into a full connection layer for classification, probability distribution of all classes is obtained, a pest and disease damage prediction model is obtained, and a final model is obtained after verification set testing and parameter adjustment and used for prediction. Through feature extraction and fusion of the image and audio data, the BiLSTM model is used to carry out classification prediction on the fusion features, automatic prediction of agricultural diseases and insect pests is realized, and the prediction accuracy is higher on the basis of reducing manpower and time costs.
Owner:NANJING UNIV OF INFORMATION SCI & TECH

Personnel ability automatic evaluation method and system based on attention neural network

The invention discloses an attention neural network-based personnel ability automatic evaluation method and system. The method comprises the following steps of: dividing an ability evaluation data set # imgabs0 # into a training set # imgabs1 #, a verification set # imgabs2 # and a test set # imgabs3 #; setting a plurality of different neural network configurations to form a configuration set # imgabs4 #; the obtained hyper-parameter # imgabs5 # is selected, and a neural network model # imgabs6 # is initialized; using a training set # imgabs7 # to train a neural network model # imgabs8 # obtained through initialization; after a # imgabs9 # round is carried out, the neural network model trained in the # imgabs10 # round is recorded as # imgabs11 #, a verification set # imgabs12 # is used for evaluating the trained model, and the neural network model and the evaluation result after each round of training are stored; the model obtaining the highest evaluation index on the verification set in N rounds is selected to be tested and stored on a test set # imgabs13 #; and until all hyper-parameters are used for initializing the neural network models and tested by a test set # imgabs14 #, deploying the neural network model with the highest model evaluation result for personnel ability grade evaluation work. And the requirement on multi-dimensional personnel ability portraits in a personnel evaluation scene is met.
Owner:浙江众合科技股份有限公司 +2

Generative adversarial network algorithm applied to image super-resolution reconstruction

The invention discloses a generative adversarial network algorithm applied to image super-resolution reconstruction. The algorithm comprises the following steps: constructing a residual dense connection module for connecting different layers of feature information of an image by adopting a residual connection module and a dense connection module; establishing a feature weight distribution attention mechanism for distributing different weights according to the importance degrees of different features; combining the content loss function, the perception loss function and the adversarial loss function to obtain a mixed loss function; a residual dense connection-based generative adversarial network is established through a residual dense connection module, a feature weight distribution attention mechanism and a mixed loss function; training the generative adversarial network based on the residual dense module according to the image training set; an image needing resolution reconstruction is input into the trained residual dense connection-based generative adversarial network to obtain a reconstructed super-resolution image, fusion of multi-layer information is improved, background noise is suppressed, and gradient disappearance and gradient explosion are prevented; and the definition of the image is enhanced.
Owner:NANTONG INST OF TECH

Intelligent recognition method of radar target constraint meta-learner

The present invention discloses a radar target constraint meta-learner intelligent recognition method, which belongs to the field of radar signal processing. By combining a normalized exponential loss constraint meta-learner with a multi-class balanced central cosine loss basic learner, a small sample radar target high-resolution distance pixel learning recognition framework is constructed; by designing a regularized parameter update method, while introducing the differential learning experience of different meta-training tasks, the mutual influence of different task losses in the same meta-test task set is fully taken into account, ensuring the effective integration of different task experiences in the parameter update process, making full use of the differential characteristics of different task learning experiences, and realizing the regularized constraint transfer and reference of different task learning experiences, which is used to guide new recognition tasks to perform effective small sample recognition, improve the accuracy of multi-category recognition of high-resolution distance images of radar targets under small sample conditions, and has promotion and application value.
Owner:NAVAL AVIATION UNIV

Intelligent recognition method of radar targets based on mission experience transfer

The present invention discloses a radar target intelligent recognition method based on task experience transfer, which belongs to the field of radar signal processing. A meta-learning recognition framework combining a multi-class balanced central cosine loss basic learner with a task loss inverse weighted update meta-learner is constructed, and a multi-class balanced cosine center loss function of a convolutional neural network in the basic learner is designed. By introducing Focal loss, the problem of decreased recognition rate of targets of a category with a small number of samples caused by the imbalance of the number of samples of different categories is avoided. Based on the smoothed mean absolute error loss function, a task loss inverse weighted update method for the parameters of the meta-learner is designed, which improves the degree of reference to high-correlation experience, and avoids the problem of non-convergence of the parameter update process caused by the small meta-training task loss value by taking a small operation. The method of the present invention improves the accuracy of multi-category recognition of high-resolution range images of radar targets under small sample conditions, and has popularization and application value.
Owner:NAVAL AVIATION UNIV

Deep learning method and system for electromagnetic signal classification and identification

The invention discloses a deep learning method and system for classification and identification of electromagnetic signals, and the method comprises the steps: constructing electromagnetic signal data sets, each data set comprising multi-dimensional electromagnetic signal features and corresponding label information; selecting and configuring a corresponding deep learning model; label pollution is carried out on the obtained electromagnetic signal data set, pollution of Y individual levels and pollution of M matrix levels are included, the pollution of the Y individual levels is carried out by randomly selecting data points of a specific proportion and modifying labels of the data points, and the pollution of the M matrix levels is carried out by applying a predefined pollution matrix to carry out systematic modification on the labels of all or part of the data set; training an obtained deep learning model by using the obtained electromagnetic signal data set with the pollution label; and inputting an unpolluted electromagnetic signal data set into the obtained deep learning model, and outputting an identified electromagnetic signal. Data pollution of different levels can be effectively handled, and it is ensured that the model can achieve the optimal classification effect in various environments.
Owner:XIDIAN UNIV

Gastrointestinal tract image classification method based on improved Adam optimization algorithm

The invention discloses a gastrointestinal tract image classification method based on an improved Adam optimization algorithm, and belongs to the technical field of medical image processing. The method comprises the following steps: a data preprocessing module for acquiring gastrointestinal tract image data for preprocessing, constructing a data set, and randomly dividing the preprocessed gastrointestinal tract image into a training set and a test set; the feature extraction module is used for extracting key features from the preprocessed image; the image classification module is used for classifying the images subjected to feature extraction by adopting an improved Adam algorithm (AdaWAGC algorithm); the result output module is used for displaying an image classification result; and the evaluation module is used for evaluating the classification result. By adopting the method to classify the gastrointestinal tract image, the accuracy and efficiency of gastrointestinal tract image recognition can be improved.
Owner:SHENYANG UNIV

Consultation question optimization method and system based on big language model prompt project

The invention discloses a consultation question optimization method and system based on a big language model prompt project. The consultation question optimization method comprises the steps that a prompt template used for model input reconstruction is constructed; obtaining a legal consultation question data set, and reconstructing original questions in the legal consultation question data set by using a prompt template to obtain corresponding input data; for each piece of input data, performing optimization problem generation by using an existing large model, and performing expert evaluation and re-optimization to obtain a final optimization problem data set; taking the generated optimization problem data set as fine tuning data, and performing fine tuning on the small base model to obtain a problem optimization model; when the user consults the legal questions, the user questions are optimized based on the question optimization model, data can be generated and the question optimization model can be constructed by utilizing the prompt project in the scene of insufficient data, and the professionality and the accuracy of expression of the legal consulting questions of the user are improved.
Owner:CENT SOUTH UNIV

CNN-LSTM-based high-voltage user power consumption data prediction method

The invention discloses a CNN-LSTM-based high-voltage user power consumption data prediction method, which takes high-voltage user power consumption data as a research object, cleans historical power consumption data of high-voltage users by using an LOF method, removes unreasonable outlier data, and reduces the influence of the outlier data on the performance of a prediction model. Through data classification and feature engineering, original data are converted into a more representative feature set, which is convenient for model learning. And the volatility of the data is reduced through data normalization, so that the stability and convergence speed of model training are improved. And then constructing a CNN-LSTM hybrid model, that is, adding a convolutional neural network CNN on the basis of a long short-term memory network LSTM. The hybrid model is composed of a CNN part and an LSTM part, the CNN part is used for extracting local features of data, the LSTM part is used for capturing a dependency relationship of a time sequence, and then more accurate power consumption prediction is carried out. After the construction and training of the model are completed, the prediction result of the model is effectively and intuitively evaluated through a mean square error and a graphical method.
Owner:南京鼎研电力科技有限公司 +1

A generative artificial intelligence detection and evaluation method and system based on blockchain

The present application discloses a blockchain-based generative artificial intelligence detection and evaluation method and system, which involves blockchain technology, including: obtaining a text classification data set; extracting features from the text classification data set; using the extracted text feature vector as training data, inputting the text detection model based on the generative adversarial network for training; using the Paillier homomorphic encryption algorithm to encrypt the trained text detection model; using the secure multi-party computing protocol to distribute the encrypted model parameters to multiple participants; obtaining the text to be classified, and using the Paillier homomorphic encryption algorithm to encrypt the text to be classified; through the secret sharing protocol and the obfuscated circuit protocol in the secure multi-party computing, each participant performs joint reasoning calculations in the encrypted space; uploading the encrypted classification results and metadata including the timestamp and the caller ID to the blockchain network. In view of the privacy leakage of generative artificial intelligence data in the prior art, the present application improves privacy protection.
Owner:DATA INTELLIGENCE HUINENG (DALIAN) TECH DEV CO LTD

Neural network-based fundus microaneurysm image automatic segmentation method, device and system

PendingCN119027431AEnhancing Non-Linear Expression Capabilitiesguarantee the sameImage enhancementImage analysisAutomatic segmentationAlgorithm
The invention provides a fundus microaneurysm image automatic segmentation method, device and system based on a neural network, and provides a new neural network structure, Tuned-PSP-UNet, which is used for microaneurysm segmentation. Specifically, it is suggested that a modified PSP module is used on a jump connection in a Tuned-UNet architecture, and an NAM attention mechanism is added. The combined loss function is used to train the network. The result shows that the provided model can realize microaneurysm segmentation. The performance of the Tuned-PSP-UNet is superior to that of other models, including a baseline U-Net. In addition, through a contrast experiment, the influence of a loss function and preprocessing on the network segmentation performance is known. And the performance influence of adjusting the PSP module in the proposed architecture is comprehensively evaluated. And finally, comparing the proposed model with the highlighted architecture in the literature, and verifying the segmentation performance of the model.
Owner:HUNAN UNIV OF CHINESE MEDICINE

High-voltage line operation abnormal data detection method

The invention discloses a method for detecting abnormal operation data of a high-voltage line, and belongs to the technical field of data anomaly detection. Comprising the steps of collecting operation data traffic in real time, and performing data preprocessing; sending the processed data to a time sequence prediction model; generating a comparison flow change baseline through the time sequence prediction model; judging whether the data traffic is normal; an RIME-CNN-LSTM-Attention model is trained, and the RIME-CNN-LSTM- Detecting line operation abnormal data traffic; judging whether flow early warning is associated or not; abnormal data association early warning information is generated; generating early warning information of abnormal data of unknown reasons; unknown abnormal data analysis and diagnosis; and triggering a line abnormal data emergency measure. According to the invention, manual intervention and operation are significantly reduced, the safety and stability of high-voltage line operation are improved, and the operation cost and maintenance difficulty of the high-voltage line are reduced at the same time.
Owner:BEIJING SHUNYI LIYUAN POWER SUPPLY ENG INSTALLATION CO +1

An aero-magnetic data deep learning forward and inverse method and system under physical constraint

PendingCN122595831AHave learning abilityget rid of shackles
The application discloses a kind of physical constraint under aero-magnetic data deep learning forward and inversion method and system. Including: constructing the forward network based on deep operator network DeepONet, the network adopts logarithm-linear dual space mechanism, to log resistivity profile and flight height as input, predicts log quadratic field;Inversion network is constructed, with the input of observed transient electromagnetic response and flight height, and the output of resistivity profile;When training, the forward network is pre-trained to convergence under data-driven loss and Maxwell equation physical constraint loss;Inversion network is trained under the joint supervision of its model space loss and physical data space loss, wherein the physical data space loss is obtained by comparing the re-synthesized response with the original observation after inputting the resistivity profile predicted by the inversion network into the frozen forward network.The application deeply integrates physical laws into data-driven models, and realizes the integrated efficient solution of aero-magnetic forward and inversion.
Owner:CHANGZHOU UNIV

A multitask eeg automatic detection and prediction system for epilepsy

ActiveCN116304575BResolve Healing Effectsproblem solvingForecastingSensorsFeature extractionMedicine
This design is a multi-task automatic EEG detection and prediction system for epilepsy. The system first receives the EEG signals from epilepsy patients to be identified or the collected patient EEG signals used for training the module through a preprocessing module. The preprocessing module performs noise reduction, filtering, segmentation, and normalization on the signals, and outputs the processed patient EEG signals to the training module. In the training module, the collected patient EEG signals are used for training and testing. First, feature extraction is performed using a multi-scale convolutional network. Then, the detection and prediction branches incorporating a dual attention mechanism are trained, and the trained model parameters are saved for use in the detection and prediction module. Finally, the detection and prediction module provides detection and prediction results for the EEG signals from epilepsy patients to be identified. This invention employs a modular design, achieving real-time detection and prediction of noisy, unbalanced EEG signals, providing a reliable system for practical applications.
Owner:HARBIN UNIV OF SCI & TECH

A smart contract vulnerability detection method based on artificial intelligence

The present invention relates to the field of smart contract security, and discloses a smart contract vulnerability detection method based on artificial intelligence, which aims to reduce the false alarm rate of smart contract vulnerability detection. The method comprises: obtaining an unlabeled smart contract data set; obtaining a labeled smart contract data set according to the smart contract vulnerability type; converting the smart contract bytecode in the smart contract data set into an operation code sequence, and using the operation code sequence as a training sample, including an unlabeled sample and a labeled sample; a smart contract vulnerability detection model comprises a pre-trained model and a vulnerability classification model; training the pre-trained model with the unlabeled training samples and training the smart contract vulnerability detection model with the labeled training samples to obtain a trained smart contract vulnerability detection model; converting the smart contract to be detected into an operation code sequence, and inputting it into the trained smart contract vulnerability detection model to obtain a vulnerability detection result corresponding to the smart contract.
Owner:UNIV OF ELECTRONICS SCI & TECH OF CHINA

Adaptive antenna trajectory planning method for secure communication of unmanned aerial vehicle

PendingCN122111066AIncrease confidentiality capacityLow eavesdropping signal-to-noise ratio suppressionVehicle position/course/altitude controlPosition/direction controlSecure communicationAntenna gain
The application discloses a secure communication trajectory planning method for unmanned aerial vehicles with integrated control of an adaptive antenna, constructs an air-ground secure communication model containing rigid body dynamics constraints, and determines that the attitude of the unmanned aerial vehicle body is uniquely determined by the translational acceleration vector and the gravity acceleration; for horizontal-horizontal (HH) and vertical-vertical (VV) two kinds of airborne antenna configurations, an explicit coupling model of the antenna gain with respect to the acceleration is established; a non-convex trajectory optimization problem is constructed to maximize the average secrecy capacity; a sequential convex approximation combined with a convex difference programming strategy is adopted for solving, wherein the implicit function differentiation method is adopted to solve the gradient singularity problem of the VV antenna null point; and finally, a flight strategy containing an optimal acceleration sequence is output. The application utilizes the active change of the flight attitude of the unmanned aerial vehicle to realize mechanical beamforming, significantly improves the secrecy capacity through "active drift" or "null alignment" maneuvering, and the generated trajectory strictly satisfies the physical feasibility.
Owner:NANTONG UNIV

Policy model of instant policy game and training method thereof

The invention discloses a strategy model of a real-time strategy game. The strategy model comprises the following steps: describing an RTS game as a Markov game form; the environment size # imgabs0 #, the environment size # imgabs1 # and the environment size # imgabs2 # respectively represent the height and the width of the grid; the state space is # imgabs3 #; a global observation space is # imgabs4 #, # imgabs5 # represents a real number space, and # imgabs6 # represents a feature channel dimension; the joint action space is # imgabs7 # and # imgabs8 # and represents the action dimension of the control unit; the state transfer function is # imgabs9 #; the reward function is # imgabs10 #; the strategy is # imgabs11 #; the strategy model is # imgabs12 #. According to the method, the problems existing in the prior art are solved, training can be carried out with lower computing resource expenditure through innovation of a strategy model architecture and a training method, and meanwhile the method can adapt to setting of various RTS games.
Owner:SHENZHEN THUNDER DIGITAL ENTERTAINMENT CO LTD

A method for constructing a road cavity detection model based on three-dimensional ground penetrating radar

This invention discloses a method for constructing a road cavity detection model based on 3D ground-penetrating radar (GPR). The method includes constructing a hybrid spatiotemporal detection model, which comprises a feature preprocessing module, a soft clipping module, a spatiotemporal feature extraction module, and an output module. The feature preprocessing module performs feature pre-extraction on preprocessed and labeled data, outputting a pre-extracted feature map. The soft clipping module performs spatiotemporal perception on the pre-extracted feature map, outputting a clipped feature map. The spatiotemporal feature extraction module analyzes the clipped feature map, outputting feature information with global spatiotemporal information. The output module integrates the feature information and predicts the existence of cavities and the range of their occurrence channels. By combining the correlations between data in the spatiotemporal dimension, it accurately detects underground cavities, solving the problems of high computational load and difficulty in adjusting model parameters in existing 3D GPR high-resolution data interpretation. This provides an efficient and reliable technical solution for intelligent road cavity detection and infrastructure safety maintenance.
Owner:JIANGSU CHENGAN PIPE NETWORK TECHNOLOGY CO LTD

Multi-beam synthesis method and device based on collaborative null generation strategy

The invention provides a multi-beam synthesis method and device based on a collaborative null generation strategy, and belongs to the technical field of antenna design. The method provided by the invention comprises the following steps: randomly generating a design index in a beam direction range, and constructing a mask matrix based on the design index; constructing a multi-beam comprehensive model, inputting the mask matrix into the multi-beam comprehensive model, and training the multi-beam comprehensive model to output amplitude excitation and phase excitation of each beam; constructing a target mask matrix according to beam performance indexes in an actual application scene; and inputting the target mask matrix into a trained multi-beam comprehensive model, outputting amplitude excitation and phase excitation of each array unit for different beams, loading the amplitude excitation and the phase excitation to a multi-beam phased array, and generating a multi-beam far-field directional diagram. The invention provides a multi-beam synthesis method and device based on a collaborative null generation strategy. The method and the device are used for reducing the loss of equivalent omnidirectional radiation power on the premise of effectively suppressing multi-beam interference.
Owner:ZHEJIANG UNIV

Intelligent network attack detection method based on flow analysis and electronic equipment

The invention discloses an intelligent network attack detection method based on traffic analysis and electronic equipment. The method comprises the following steps: acquiring traffic data in a network; a pre-trained automatic encoder AE is constructed, the automatic encoder AE comprises an encoder module, a potential space and a decoder module, the encoder module adopts two layers of long short-term memory networks, and a Dropout layer is inserted between the two layers of long short-term memory networks and is used for receiving input traffic data and performing data compression and feature extraction to obtain key traffic features; and the potential space adopts a bidirectional gating circulation unit to carry out key traffic feature classification, and outputs a detection result indicating whether the traffic data has an abnormal traffic type or not through a decoder module. According to the method, the important traffic characteristics in the traffic can be fully extracted, and the problems of low detection rate and high false alarm rate caused by solving the problem of data imbalance by using a sampling technology are avoided while data dimension reduction is completed.
Owner:NANKAI UNIV

Image deblurring method based on generative adversarial network and storage medium

The invention discloses a construction method of a generative adversarial network, which comprises the following steps: adopting a basic architecture based on the generative adversarial network GAN, adopting an FPN feature pyramid structure for a generator, and adopting a CNN convolutional neural network structure for a discriminator; adding a self-attention module in each layer of structure of the FPN to obtain an improved generative adversarial network; wherein the self-attention module receives a feature map from the FPN as input, converts the feature map into a query matrix Q, a key value matrix K and a value matrix V in a matrix space, determines weight distribution of the V by calculating similarity between the Q and the K, and obtains output fused with a self-attention feature vector; the deblurred image quality is high, the training efficiency is high, and the training process is stable.
Owner:WUHAN POLYTECHNIC UNIVERSITY +1

A general multi-level fusion attention model for ship detection in urban coastal SAR images

The present invention relates to a universal multi-level fusion attention model for ship detection in SAR images of urban coasts, comprising establishing submodules DFEM, CMAFM and SMAFM. Firstly, DFEM is designed to extract rich features with different resolutions and semantic information for multi-scale ship detection; then CMAFM and SMAFM are designed to fuse channel attention blocks and spatial attention blocks on feature maps of different levels to better learn dependent features from channel and spatial dimensions to enhance feature representation; finally, the fused feature maps are input into an existing ship detection network to obtain detection results; the model can focus on more important part of the input image, enhance the position information of the ship, weaken unimportant feature information, improve model performance through a small number of additional parameters, and improve the accuracy of identifying large and small ships under the complex urban coast background.
Owner:ZHEJIANG COLLEGE OF SECURITY TECH +2

Intelligent recognition method of radar targets based on negative exponential updating

The present invention discloses a radar target intelligent recognition method based on negative exponential update, which belongs to the field of radar signal processing. By combining a multi-class balanced central cosine loss basic learner and a task loss negative exponential correction meta-learner, a small sample radar target high-resolution distance image meta-learning recognition framework is constructed; a task loss negative exponential function weighted update method for the meta-learner parameters is designed, and the negative correlation between the meta-training task loss value and the task training effect is utilized. Based on the gentleness of the change of the negative exponential function, the reference of the task learning experience is closely associated with the task relevance, which reduces the unstable effect of the parameter update process caused by the drastic change of the loss value of different meta-training tasks, avoids the divergence problem of the parameter update process caused by the meta-training task loss value being too small or too large, and the calculation process is simple and convenient. The method of the present invention improves the accuracy of multi-category recognition of high-resolution distance images of radar targets under small sample conditions, and has the value of popularization and application.
Owner:NAVAL AVIATION UNIV

A multi-level land resource segmentation and migration fine-tuning method with input alignment

This invention discloses a multi-level land resource segmentation migration fine-tuning method with input alignment, comprising the following steps: acquisition and processing of remote sensing image data and label data related to the land resource segmentation task; selection and slicing of sampling areas based on prior knowledge of the study area; data augmentation to expand the dataset and pixel value normalization; partitioning the model dataset using stratified sampling techniques; determining the optimal initial learning rate through hyperparameter optimization, entering input feature alignment, and establishing a quick connection between the input alignment module and the model output; employing three Transformer architectures and their pre-trained models for land resource segmentation, and optimizing the model training process using gradient-value-constrained gradient clipping and two-stage training strategies; step six: segmentation prediction and visualization. This invention improves the model's training efficiency and segmentation performance, achieving high-precision multi-level land resource segmentation and recognition.
Owner:INST OF GEOGRAPHICAL SCI & NATURAL RESOURCE RES CAS

Intelligent recognition method of radar target based on adaptive reverse truncation

The present invention discloses a radar target adaptive reverse truncation intelligent recognition method, which belongs to the field of radar signal processing. Aiming at the problem of poor recognition accuracy of one-dimensional range image multi-category under small sample conditions, a recognition framework of fuzzy truncated cosine loss basic learner combined with normalized exponential loss constraint meta-learner is constructed. Based on the potential difference between source domain data of meta-training task and target domain data of meta-test task, the difference design of convolutional neural network loss function is carried out in stages, and the universal features of source domain data are extracted while fully exploring the individual features of small samples in target domain; in addition, by taking into account the mutual influence of different task losses in the same meta-test task set, the regularized constraint stable update of meta-learner parameters is realized, the calculation process has strong regularity, and can be applied to the effective migration and fusion of meta-training task learning experience from different sources, which improves the overall reference efficiency of meta-training task experience and has promotion and application value.
Owner:NAVAL AVIATION UNIV