35results about How to "Strong discrimination" patented technology

The invention belongs to the technical field of biological medicine and electronic information technology, and discloses an atherosclerosis model based on nano-molecule magnetic resonance imaging and a machine learning algorithm, a nano-targeting probe and application of the atherosclerosis model based on the nano-molecule magnetic resonance imaging and the machine learning algorithm. A high-resolution MRI image capable of specifically displaying foam macrophage distribution in the plaque is obtained; then, image omics features are extracted from the images, a machine learning model which is constructed based on the features and verified by pathological labels is adopted for analysis, and finally an objective and quantitative plaque vulnerability score is output. According to the model disclosed by the invention, an objective and quantitative risk score can be output by fusing nano-targeted molecular imaging and machine learning, and accurate and reliable quantification of plaque vulnerability is realized. According to the invention, through systematic integration of targeted molecular imaging and machine learning, significant technical progress is brought.

This invention relates to the field of multimodal sentiment analysis technology, and in particular to a method and system for multimodal sentiment analysis based on sequential translation under uncertain and missing modalities. The method includes: acquiring multimodal sentiment data, including visual, textual, and auditory modalities; constructing a multimodal sentiment analysis model based on sequential translation, firstly by extracting self-attention encoding features of each modality using a Transformer encoder; and then designing a bidirectional sequential translation architecture centered on the textual modality. This invention improves the utilization of textual modal features in the feature fusion process, thereby enhancing the quality of multimodal fusion features and providing strong support for subsequent complex sentiment analysis.

The present application relates to a kind of well seismic signal-noise superposition wave field high-dimensional multi-component feature intelligent separation method, belong to geophysical oil and gas resource exploration field.Establish time-domain multi-component decomposition result serves different dimension differentiation characteristic property expression;Six-dimensional characteristic attribute space based on multi-component signal-noise component is established, and the mathematical characteristics of complex signal wave and multiple types of noise in DAS record, such as direct wave and reflected wave, are revealed;Dual-stage highly integrated framework with tree model as base learner is set up to complete the feature point category determination separation task in high-dimensional attribute space, so as to accurately and efficiently realize low-damage preservation of signal wave field and maximum suppression of noise wave field.The method is flexible, accurately preserves the amplitude characteristics and frequency energy components of DAS two-dimensional exploration record signal field in well, the mathematical basis is reliable and reliable, and the feature is strongly interpretable, which provides an important basis for thin and fine layer detection, oil and gas reservoir exploitation and other practical applications.

A multi-source domain EEG emotion recognition method based on self-organizing sparse directed graph convolution includes: constructing EEG feature vectors based on EEG signals; constructing a sparse adjacency matrix based on the EEG feature tensor; performing directed graph convolution operations on the forward and backward adjacency matrices of the sparse adjacency matrix respectively to capture the asymmetric information propagation characteristics between EEG channels and generate domain-invariant EEG emotion features; constructing multiple domain feature extraction branches for multiple source domain samples and target domain samples; inputting the domain-invariant EEG emotion features into the multiple domain feature extraction branches respectively; constructing independent feature representation spaces for each source domain and target domain, separating domain-shared features and domain-specific features; classifying samples based on domain-shared features and domain-specific features; fusing the prediction results of each source domain classifier for the target domain samples to output the target domain EEG emotion recognition result; and performing cross-domain optimization of the model using the weighted result of the overall loss function of each source domain.

The invention discloses a multi-modal depression auxiliary detection algorithm, and the method comprises the steps: obtaining multi-modal data which comprises video data, voice data and scale data of a medical site; carrying out identity recognition, round labeling and format conversion processing on the multi-modal data to obtain structured data containing role information, time sequence round and content fields; performing feature extraction on the structured data to obtain a feature vector set containing an image mode, a text mode and a voice mode; carrying out fusion processing on the feature vector set by adopting a preset cross-modal attention mechanism dominated by a text mode to obtain multi-modal fusion feature information; and inputting the multi-modal fusion feature information into a pre-configured depression risk assessment model, performing feature transformation and compression through a full connection layer and a pooling layer, and outputting a depression disease probability for representing a target individual. According to the method, complementation and dynamic semantic modeling of multi-source data can be realized, and the accuracy and reliability of a detection result are improved.

The invention discloses a DNA bar code system and method for identifying glycyrrhiza medicinal plants and hybrid complexes thereof, and belongs to the technical field of molecular biology and traditional Chinese medicine identification. The system is formed by combining three bar codes of a nuclear gene segment ITS, a chloroplast gene segment ndhA and a chloroplast gene spacer trnH-psbA. Wherein ndhA is a core bar code which is screened and verified for the first time, and the specific variation site of the ndhA can be used for accurately distinguishing three kinds of medicinal liquorice, namely, Glycyrrhiza uralensis Fischhex DC., Glycyrrhiza glabra. And Glycyrrhiza inflata Batalin. The system fully utilizes the characteristics of chloroplast paternal heredity of glycyrrhiza, not only can identify homozygous species, but also can efficiently analyze complex hybridization and introgression types, and traces the source of a male parent. Compared with a traditional multi-bar code combination, the method has the advantages of being high in identification success rate (up to 98.08%), easy and convenient to operate and low in cost, and is suitable for liquorice germplasm resource identification, medicinal material market rapid screening and improved variety breeding.

The invention relates to the technical field of machine vision detection, in particular to an inner package detection and recognition method based on vision detection, which comprises the following steps: synchronously acquiring a multi-view image sequence of a to-be-detected piece through an annularly arranged acquisition device; independently analyzing each image, and generating an initial defect area mask and a local feature descriptor; a three-dimensional defect space distribution model is constructed by fusing multi-view masks, and geometric topology and depth information of the three-dimensional defect space distribution model are extracted to form a comprehensive feature vector. And performing cross-modal correlation analysis on the local features and the three-dimensional comprehensive features to generate an enhanced defect feature spectrum. And calling a pre-training classification model to analyze the atlas, outputting an inner package type, a defect type and a severity level, and generating a quality evaluation report and a processing instruction according to the inner package type, the defect type and the severity level. According to the method, three-dimensional space accurate positioning and multi-feature deep fusion of the defects are realized, and the recognition accuracy and evaluation comprehensiveness of the complex defects are improved.

The application provides a pig behavior recognition method and device based on skeleton topology constraint and an electronic device. The method comprises the following steps: S1, obtaining a pig image to be recognized; S2, inputting the pig image into a trained pig behavior recognition model to obtain a pig behavior recognition result; the pig behavior recognition model comprises a backbone network, a neck network and a detection head; the backbone network is used for multi-scale feature extraction on the input pig image; the neck network is used for feature fusion and enhancement on the multi-scale features extracted by the backbone network to generate a multi-scale enhanced feature map; the detection head comprises a key point prediction branch, a visual classification detection branch and a behavior classifier. The application deeply fuses the skeleton space topology features between key points and visual priori, so that the behavior discrimination can explicitly utilize the geometric dependency relationship between key points, and effectively solves the problem of accurate pig behavior monitoring in a complex pigsty environment.

The invention relates to the technical field of identity authentication, and discloses an identity authentication method and system based on text-image multi-modal feature fusion. The method comprises the steps that character interval time and input behavior characteristics are encoded into embedded vectors, and text behavior representation vectors are obtained through processing of an encoder; calculating inter-frame dynamic characteristics based on the image time sequence data, and identifying a micro-expression triggering moment to obtain an image behavior anchor point; extracting a text behavior anchor point from the text time sequence data, and calculating the time sequence matching degree of the text behavior anchor point and the image behavior anchor point to obtain a cross-modal consistency score; and adjusting the fusion weight of the text behavior representation vector and the image feature vector according to the cross-modal consistency score, and outputting a verification result after feature fusion. According to the method, a cross-modal behavior anchor point time sequence alignment mechanism and a self-adaptive fusion strategy are constructed, so that the detection capability of time decoupling attacks and the overall robustness of a verification system are improved.

The application discloses a self-supervised contrast learning method and device, electronic equipment and a storage medium. The method comprises the following steps: constructing a point cloud data set according to positive samples and negative samples in a training stage, and obtaining spatial geometric features of point cloud data by using self-supervised contrast learning based on a laser point cloud 3D identification model framework; and performing target detection of 3D point cloud data according to the spatial geometric features of the point cloud data in an inference stage. The application fully utilizes spatial priori and improves the performance of the model.

The invention discloses a visible light-infrared target detection method based on structured feature extraction and dynamic weight prediction. The method comprises the following steps: carrying out target detection through a visible light-infrared target detection model based on structured feature extraction and dynamic weight prediction; the target detection model comprises a thermal radiation-motion feature extraction module, a multi-granularity visual representation module and a collaborative weight decision module; the thermal radiation-motion feature extraction module is used for capturing thermal radiation information in an infrared image and motion features of a target so as to generate infrared feature representation; the multi-granularity visual representation module is used for performing structured multi-granularity visual feature extraction on the visible light image so as to construct visible light feature representation; and the collaborative weight decision module is used for performing cross integration of infrared feature representation and visible light feature representation through a dynamic weight prediction mechanism. The target detection model provided by the invention has better performance in a visible light-infrared target detection task.

The invention discloses a multi-view bearing fault diagnosis method based on noise, edge and difficulty perception, and the method constructs a multi-view perception framework of an input view, an operator view and a training view, and specifically comprises the steps: constructing a noise perception multi-scale expansion gating modulator at the input view; designing an adaptive noise perception graph to generate directional Gaussian perturbation so as to actively enhance the noise distinguishing capability; constructing a boundary edge sensing module at an operator perspective, and designing boundary adaptive filling and edge adaptive gradient sensitive masks to effectively suppress convolution boundary artifacts and enhance fault edge features; an inverse distillation strategy guided by difficulty perception clustering is put forward from a training perspective, a sample learning state is dynamically recorded by constructing a difficulty memory mechanism, and an intra-class multi-state soft label is generated by designing online clustering and inverse neighborhood sampling, so that a model is guided to carry out targeted optimization on a difficult sample, and the fault discrimination capability is remarkably improved. Experiments prove that the method has higher fault diagnosis performance.

The invention provides a high-missing-rate feature processing method and device, and the method comprises the steps: carrying out the preprocessing of original features in original feature data, obtaining the preprocessed transformation features, carrying out the multi-path embedding transformation of the original features based on a plurality of embedding transformation rules, and obtaining the transformation features of the original features; fusing a plurality of embedded vectors obtained by conversion into a target embedded vector as a first derivative feature; obtaining a comprehensive similarity matrix based on comprehensive similarity obtained by fusing multiple similarities between the target embedding vectors of any two original features; for each original feature, selecting the original feature of which the comprehensive similarity with the original feature meets a preset similarity condition, and generating at least one second derivative feature; and splicing the transformation features, the first derivative features and the second derivative features to obtain a feature set, thereby deeply mining the relationship of the high-missing-rate features, improving the quality of the feature set, and enhancing the model training effect.

The invention discloses a network intrusion detection feature selection method based on a multi-strategy improved neural network algorithm, and the method comprises the steps: firstly carrying out the cleaning, normalization and feature coding processing of original network flow data, constructing an initial feature space, and representing a feature selection scheme in a binary vector form; then Tent chaotic mapping is introduced to initialize a population, and the diversity of an initial solution is enhanced by using the ergodicity and uniformity of a chaotic sequence; in the feature optimization process, the Levy flight strategy is adopted to update the position of a leader, the global search ability is improved through a search mechanism combining long and short steps, local optimum is avoided, meanwhile, a follower update strategy based on the adaptive inertia weight is designed, global exploration and local development ability is dynamically balanced, and the algorithm convergence speed is increased; and further constructing a weighted fitness function taking a classification error rate and a feature subset scale as double targets, evaluating candidate feature subsets, and determining a global optimal feature subset through an elitist retention and iteration termination mechanism.

The invention relates to the technical field of image processing, in particular to an equipment cluster target information positioning method based on a neural network model, and the method comprises the following steps: 1, obtaining an operation instruction of target equipment and video information corresponding to the operation instruction, and generating an operation sample; collecting a plurality of operation samples for each type of operation instruction, and generating an operation sample data set corresponding to the operation instruction; 2, acquiring image information of an operation panel of the target equipment, and generating three-dimensional modeling information of the operation panel based on a point cloud algorithm; 3, inputting the operation sample data set into the feature extraction network model; and 4, correspondingly mapping the targets in the target instruction set into the three-dimensional modeling information, and generating target mapping information of the operation instruction in the real world. According to the scheme, the intelligent identification and positioning of the equipment operation state and the target information can be realized without carrying out any physical transformation on the equipment or obtaining the authorization of an original factory interface.

The invention provides a hyperspectral image classification method. Comprising the following steps: extracting spatial-spectral joint features of an image by using two-layer 3D convolution; a spatial-spectral feature extraction module based on a pre-activated residual network is constructed in combination with 2D and 3D convolution, the image advanced spatial semantic feature extraction capability is enhanced, and the convergence speed of the model is increased; a plurality of residual modules are connected to fully mine different forms of features extracted by each convolutional layer, and meanwhile, multi-feature fusion is carried out between blocks to realize feature complementation; effective fusion of shallow and deep features is realized by using long-distance residual connection, and the expression ability of the features is further enhanced; and a Softmax classifier is adopted to realize image classification. According to the method, the feature reuse of the hyperspectral image can be effectively realized, the deep and advanced spatial spectrum features with better identification performance and robustness are obtained, and the classification precision is improved.

The invention relates to a carsickness recognition method, an electronic device and a vehicle, and the method comprises the steps: obtaining the environment features of the vehicle and the visual features of a target passenger in the vehicle; the vehicle environment characteristics comprise vehicle state parameters and environment parameters associated with inducing carsickness of passengers in the vehicle, and the visual characteristics are behavior performance characteristics related to the carsickness state; constructing a main feature vector according to the visual features; according to the visual features and the vehicle environment features, auxiliary feature vectors are constructed; based on the correlation degree between the main feature vector and the auxiliary feature vectors, performing dynamic weighted fusion on each auxiliary feature vector to generate a multi-modal feature vector; and carsickness identification processing is carried out on the multi-modal feature vector, and a carsickness identification result for the target passenger is generated. According to the invention, the problem of low carsickness identification accuracy is solved.

The application discloses a kind of stone surface image defect intelligent detection and real-time grading method, by the RGB image of the stone surface to be detected, thermal radiation image and depth image are synchronously collected, three modal images are respectively input corresponding feature extraction network and extract multi-scale hierarchical features, with cross-modal gate fusion module Channel inter-attention weight and spatial attention weight of each modal feature map are calculated respectively for each feature level and dynamically weighted fusion, generate unified cross-modal defect representation, then through grading decision network based on four semantic attributes mapping to continuous quality score space, configurable grading threshold output discrete quality level.The application utilizes the complementary characteristics of multi-modal information, solves the problem that single visible light image is difficult to accurately distinguish natural texture from real defect under complex texture background, while completely decouples defect objective attribute description and grading subjective standard application, significantly reduces deployment and maintenance cost.

The application discloses a social network link prediction method based on a space-time feature perception time sequence diagram network, which divides key features in a time sequence diagram into two categories of time sequence features and structure features, for the time sequence features, through adaptive fusion of a continuous time method and a discrete method, long sequence dependence and recent dependence are effectively captured and weighed, for the graph structure features, through co-occurrence neighbor coding, graph structure information is effectively explicitly coded, and a hash-based method is used to improve co-occurrence neighbor retrieval efficiency. Such a method not only exceeds previous methods in prediction accuracy by fully capturing time sequence features and structure features, for example, the average accuracy is improved by 14.18%, 15.89% and 32.48% respectively on USLegis, UNtrade and Unvote data sets compared with previous optimal methods, and better trade-off is achieved in inference efficiency. And it is simple to realize and easy to reproduce.

The invention discloses a particle board surface defect identification method and system based on a twin network and supervised comparative learning, and the method comprises the steps: (1) collecting a particle board surface image, carrying out the preprocessing, constructing a small sample data set containing various defects, and dividing the small sample data set into a training set and a test set; (2) constructing an improved twin supervised contrast network model, wherein the model comprises a feature extraction backbone network, an LDFPN and a classification contrast learning head; (3) training the twin supervised comparison network model by using the training set, and adopting joint optimization of cross entropy classification loss and supervised comparison loss during training; and (4) performing defect identification and classification on the shaving board surface image by using the trained twin supervision comparison network model. The method and the system aim at solving the problems of low particle board surface defect identification accuracy and weak model generalization ability under the condition of small samples, efficient and accurate automatic quality detection is realized, and the method and the system have the potential of real-time deployment on a production line.

This invention relates to the field of railway engineering schedule management and intelligent decision-making technology, and in particular to a method and system for ranking railway engineering schedule indicators. The method includes constructing a multi-dimensional schedule indicator system, establishing a hesitant fuzzy decision matrix, introducing a β threshold to construct a covering approximate space, calculating attribute dependencies to reduce redundant indicators, reconstructing the matrix, and calculating a comprehensive evaluation value, ultimately achieving a scientific ranking of schedule indicators. This application can effectively handle uncertainties, incompleteness, and attribute redundancy while preserving the multiple hesitancy and fuzziness of evaluation information, thereby improving the efficiency and accuracy of schedule indicator screening and ranking.

The application discloses a Yi ancient book character detection method based on visual-language guidance, which comprises the following steps: Yi ancient book image preprocessing; constructing a Yi ancient book character detection model, wherein the Yi ancient book character detection model comprises an image encoder, a coarse text positioning module, a positioning guide visual attention module, a text semantic encoder, an instance-language matching module and a detection head; training the Yi ancient book character detection model based on a Yi ancient book image dataset; inputting a to-be-detected Yi ancient book image into the trained Yi ancient book character detection model, and outputting a character detection result image. The application has high robustness and detection precision under the conditions of degradation, dense distribution and complex background interference, thereby providing new technical support for the digital protection, character detection and intelligent recognition of Yi ancient books.

The invention discloses a gesture recognition method and system for self-vehicle traffic police command based on multi-modal fusion. According to the method, firstly, the scale of an intersection is judged by integrating multi-source radar sensing data, lane information is positioned by using a double-branch model, a theoretical attention angle geometric priori area is constructed, the relation between the body orientation of a traffic police and the space is mapped into the attention intensity of the traffic police, and whether a command gesture faces a vehicle or not is determined; detecting the output traffic police identity confidence through YOLOv8; extracting skeleton key points by using OpenPose and modeling an action time sequence to obtain a preliminary gesture probability; and finally, carrying out multi-modal joint coding on lane information, traffic police attention intensity, traffic police identity confidence and traffic police command gesture probability values, and inputting the coded information into a gesture recognition model for global feature fusion and decision making. According to the invention, illumination change, fuzzy traffic police command intention affiliation, traffic police height difference and complex background interference can be effectively overcome, and command gesture instructions facing the vehicle can be accurately distinguished.

The application discloses a set of bile acid marker combinations, kits and applications for female sarcopenia patients with liver fibrosis risk stratification. The bile acid marker combinations are GCA, GDCA, GCDCA, TDCA, TDCA-3S, TCDCA, GLCA-3S and GDCA-3S in serum or plasma. Through quality control, difference screening and multi-model verification, a joint scoring model is constructed to distinguish the basic bile acid metabolism type and the liver fibrosis susceptible type subgroup. The experimental results show that the AUC of the model verification queue reaches 0.982; the median of the susceptible type subgroup LSM is 6.0 kPa, and the high-risk proportion is 20.7%, which is significantly higher than that of the basic type. The application can realize early and accurate stratification of the liver fibrosis risk of female sarcopenia patients, and is especially suitable for clinical screening and risk assessment.

The experiment operation atomic action timing positioning and scoring method based on contrast learning belongs to the technical field of computer vision. The method comprises the following steps: constructing an atomic action semantic library; obtaining training videos and labeling the atomic action categories and start and end time stamps of each video segment; performing multi-modal feature extraction on the video segments to obtain initial feature representations; performing contrast representation learning on the initial feature representations by using hierarchical contrast learning to obtain optimized feature representations; constructing a timing positioning network and training the network with the optimized feature representations and boundary labels as supervision; inputting a target experiment operation video into the trained timing positioning network to output the atomic action category, start and end time stamps and confidence of each video segment; and comparing the output results with a standard operation specification knowledge graph to generate an operation specification score. The method can realize fine-grained action positioning and automatic scoring of experiment operations, reduces labeling costs and improves scoring accuracy.

The invention belongs to the technical field of traditional Chinese medicine analysis, and particularly relates to a construction method and application of a specific chromatogram of an achyranthes aspera throat-clearing mixture. According to the invention, the spanning from single index control to overall quality evaluation is realized; the specific characteristic spectrum method for the achyranthes aspera throat-clearing mixture is established for the first time, and the limitation that only the content of a single component is measured in the prior art is broken through. By simultaneously monitoring the five key characteristic peaks of chlorogenic acid, caffeic acid, bergenin, geniposide and catechin and the consistency of the overall spectrum of the five key characteristic peaks, the basic material feeding condition and the matching stability of the monarch, minister, assistant and guide medicines in the preparation can be comprehensively reflected, so that the overall quality and the uniformity of the curative effect of the product are more scientifically guaranteed.

The present application relates to oil filling equipment technical field, provide a kind of large oil filling equipment typical fault defect identification method and system, it includes the following steps: step 1, acquisition large oil filling equipment in the ultrasonic signal of running process, and the ultrasonic signal is preprocessed, obtain the voiceprint data to be analyzed;Step 2, extract the multi-dimensional acoustic feature in the voiceprint data to be analyzed;Step 3, construct multi-dimensional feature fusion coding mechanism, the multi-dimensional acoustic feature is weighted and fused, obtain fusion feature vector;Step 4, the fusion feature vector is input into the deep learning model that is pre-constructed and trained, and the fault type and confidence corresponding to the current running state of equipment are output;Step 5, based on the fault type and confidence, calculate fault risk index, and combine preset multi-level early warning threshold, generate equipment state evaluation result and operation and maintenance suggestion.The present application can preferably identify fault defect.

The invention belongs to the technical field of concrete construction quality detection, and particularly discloses a concrete chiseling roughness illumination correction detection method and system, and the method comprises the steps: collecting a color image and a depth map of a chiseling surface, and calculating a roughness saliency map and an exposure quality map; performing adaptive exposure correction on each image to obtain an enhanced image; taking the enhanced image, the depth map, the roughness saliency map and the exposure quality map as common input, and extracting initial texture features and geometric structure features to obtain an initial fusion feature map; respectively constructing a high-resolution texture flow and a low-resolution semantic geometric flow based on the initial fusion feature map, and fusing the high-resolution texture flow and the low-resolution semantic geometric flow to obtain a second fusion feature map; and performing frequency domain enhancement on the second fusion feature map to obtain a frequency domain enhanced feature map, and performing scabbling roughness grade classification to realize concrete scabbling roughness illumination correction detection. According to the invention, the problems of insufficient image reliability, limited model discrimination precision and poor environmental adaptability in the prior art are solved.

The application relates to a CSP-CPSO-SVM electroencephalogram decoding algorithm and a brain-controlled lower-limb exoskeleton system, wherein the algorithm adopts a CSP algorithm to extract features of electroencephalogram signals to obtain an initial feature set; adopts a CPSO algorithm to intelligently search and filter and optimize the initial feature set, and obtains a feature subset which is strong in discrimination and low in redundancy; the feature subset is input into an SVM classifier for recognition, and an intention category label is output, the combination of CSP-CPSO-SVM has obvious advantages in improving feature discrimination and classification accuracy, and provides a reliable method for efficient recognition of motor imagery electroencephalogram signals. The system comprises a sensing unit, an intention transmission unit and an execution unit, the system integrates the CSP-CPSO-SVM electroencephalogram decoding algorithm, establishes a complete mapping link from intention recognition of electroencephalogram signals to motion control of a lower-limb exoskeleton, and realizes closed-loop control from intention recognition to trajectory following. The system can stably follow an expected trajectory according to a motion intention, and has good robustness.

The invention discloses a meta-learning-based knowledge graph small sample node classification method, which is used for realizing efficient and stable small sample node classification with generalization ability on graph structure data. Firstly, a task difficulty-driven meta-training scheduling mechanism is provided, and a model is guided to gradually learn from easy to difficult by evaluating task complexity and dynamically sequencing and scheduling training tasks, so that training stability and generalization performance are improved. Secondly, a Poisson learning pseudo-label generation method based on a graph structure is designed, high-confidence pseudo-labels are generated by using unlabeled nodes and adjacency relations thereof, and KL divergence constraint optimization inter-class prototype distribution is introduced, so that the adaptability of a model to unseen classes is enhanced. And finally, a discriminative negative sample enhancement strategy is constructed, and the recognition capability of the model at the category boundary is enhanced by generating a discriminative negative sample pair in the feature space, so that the prototype confusion problem is effectively relieved.