575 results about "Network processing" patented technology

The invention provides a falling detection method and system based on millimeter wave radar fusion human body postures, and relates to the technical field of falling detection, and the method comprises the steps: collecting human body echo signals, and constructing three-dimensional point cloud data through distance, angle and speed estimation; the point cloud is processed to generate a dynamic sequence, and skeleton features and key point coordinates are extracted in combination with a graph convolutional network. A skeleton connection relation is constructed based on the key points, attitude features are calculated, key change features are extracted through a self-attention mechanism, and the key change features are fused with a point cloud sequence to construct multi-modal features. And the double-branch state recognition network processes the fusion features, and when abnormality is detected, further analysis is carried out through residual attention and space-time diagram convolution, and finally, the falling state is recognized and the risk level is evaluated.

The invention relates to the technical field of medical image intelligent analysis and artificial intelligence auxiliary diagnosis, in particular to a medical image disease course prediction system based on an industrial neural network, and the system comprises a reference generation module which is used for obtaining static image data; processing the static image data by using a physical perception neural network to generate a pure ideal state reference; a perturbation simulation module; the industrial kinetic parameters are used as perturbation terms to be superposed to a pure ideal state reference, and a theoretical damaged state is generated; the projection verification module is used for acquiring real multi-modal observation data; generating a real residual error; generating a theoretical residual error based on the theoretical damaged state and the pure ideal state reference; calculating a manifold coupling confidence coefficient; the closed-loop correction module is used for performing inversion optimization on the industrial kinetic parameters; outputting a disease course prediction result according to the manifold coupling confidence coefficient; according to the method, the problem that a traditional medical model lacks physical consistency explanation is solved, and the credibility of artificial intelligence auxiliary diagnosis is remarkably improved.

The invention relates to the technical field of deep learning and computer vision, and discloses a histogram interactive attention-based pavement crack segmentation network processing method and system, so as to enhance the edge detail fidelity and improve the crack segmentation precision. The method comprises the steps of image preprocessing, up-sampling, down-sampling, feature fusion and image reconstruction processing. Wherein global feature modeling in the intensity sub-boxes and among the sub-boxes is realized by constructing a histogram interactive attention module (HIA); a double-branch detail enhancement feedforward module (DDEF) is introduced to enhance spatial detail and high-frequency edge information expression; meanwhile, a Fourier jump enhancement module (FFSM) is adopted to jointly refine jump connection features in a spatial domain and a frequency domain. Through the synergistic effect of the modules, the network can realize continuous recovery and structural consistency modeling of a crack boundary in a complex pavement environment, so that the accuracy and the stability of a segmentation result are remarkably improved.

The invention provides a structure health early warning method and system based on space-time correlation characteristics and digital twinning, and relates to the technical field of data processing. The method comprises the following steps: acquiring multi-source heterogeneous data of a target structure; performing dynamic sampling alignment and wavelet packet decomposition on the multi-source heterogeneous data to extract energy features to obtain synchronous data, and performing abnormal data filtering on the synchronous data to obtain cleaned fusion data; performing wavelet decomposition on a high-frequency vibration signal in the fused data to obtain a damage impact feature, performing time sequence processing on low-frequency temperature data in the fused data to obtain a temperature time feature, and performing dynamic graph convolutional network processing on strain data in the fused data to obtain a spatial correlation feature; constructing an input vector; calculating a damage degree index; and according to the damage degree indexes, early warning grades are divided, and corresponding control instructions are triggered for different early warning grades. By implementing the technical scheme provided by the invention, the accuracy of structural health early warning can be improved.

The invention discloses a video crowd counting method based on a cascaded cross-domain feature interaction network. The method comprises the following steps: carrying out data enhancement processing of random cutting and horizontal flipping on a current frame and front and back frames of the current frame; and constructing a cross-domain feature interaction network composed of a spatial domain branch and a frequency domain branch. The frequency domain branch extracts frequency domain feature output of different stages through a high and low frequency signal aggregation module and a feature encoder based on adjacent frames; the spatial domain branch is based on a single-frame image, and static spatial semantic features are extracted through a feature encoder. Cascade fusion is carried out on the double-branch features on multiple scales, two-way channel cross attention is utilized to reconstruct time sequence correlation frequency domain features of a current frame, and fusion and reconstruction of the two domain features are achieved through a cross-domain feature mutual modulation module. And after the reconstructed double-branch features are processed by the fusion network, outputting a crowd density map of the current frame by a density regression head. And after training is completed, storing the optimal model for video crowd counting. According to the invention, through cross-domain feature cascade and bidirectional time sequence modeling, the accuracy and robustness of crowd counting in a video scene are effectively improved.

The invention relates to a solid-state laser radar ranging method and system based on neural network processing, and belongs to the technical field of laser radar ranging, and the method comprises the steps: obtaining a laser echo signal when laser radar detection is carried out on a surrounding target object, and obtaining a photon counting histogram based on the laser echo signal; preprocessing the photon counting histogram, and inputting the preprocessed photon counting histogram into a pre-trained neural network model to obtain a peak confidence map containing an echo peak value and a peak time offset map; on the basis of the peak confidence map and the peak time offset map, obtaining one or more echo peak values of which the confidence exceeds a predetermined threshold and time position information of the one or more echo peak values of which the confidence exceeds the predetermined threshold; and based on the time position information, the distance information with one or more target objects is determined, so that the time position information measurement accuracy of one or more echo peak values is improved, and the laser radar ranging accuracy is improved.

The invention discloses a multi-unmanned aerial vehicle cooperative inspection trajectory control method for optimizing medical area coverage and service efficiency. The method comprises the following steps: firstly, constructing a medical multi-unmanned aerial vehicle auxiliary inspection mobile edge computing system model, defining an unmanned aerial vehicle and mobile user set, and establishing a communication model containing A2G and A2A links, an unmanned aerial vehicle mobile model and an energy consumption model; then taking a joint function of a coverage score, a system throughput and an emergency task completion rate as an optimization target, under energy and communication connectivity constraints, proposing an LT-MADDPG algorithm, adopting a CTDE framework, processing a time sequence state by an actor network integrated with LSTM, fusing global information by a commentator network embedded with Transform through multi-head attention, and finally obtaining an emergency task. And modeling a cooperative relationship between the unmanned aerial vehicles and a medical task priority. Experiments show that the method is superior to a traditional algorithm in the aspects of coverage, service fairness and system throughput, and the medical inspection efficiency and reliability are effectively improved.

The invention discloses a blood glucose prediction method based on multi-scale data processing, and belongs to the technical field of blood glucose prediction. The method comprises the following steps: firstly, setting three sliding windows, generating small-scale, medium-scale and large-scale sub-sequence sets on a normalized historical blood glucose sequence, and extracting corresponding feature vectors; constructing two types of attention sequences by calculating the difference between the small-scale feature vector and the medium-scale and large-scale feature vectors; a scale feature fusion unit is combined with the attention degree sequence to obtain small-medium and small-large scale weighted features; and finally, processing the weighted features by adopting a two-channel feature blood glucose prediction network, and outputting a predicted blood glucose value. According to the method, through a fusion strategy of multi-scale feature extraction and attention guidance, the accuracy of blood glucose prediction is improved.

The invention discloses an oil chromatography data trend classification method and system based on depth feature enhancement and an attention mechanism, and the method comprises the steps: carrying out numeralization conversion, deletion detection and grouping trend calculation on oil chromatography original gas component data, and generating a basic feature vector; executing multi-scale sliding statistics, change rate and subsequence feature enhancement, and calculating comprehensive similarity and attention weight based on a template library to generate a weighted similarity vector; splicing the enhanced feature and the weighted similarity vector into a time sequence input sequence, and outputting an oil chromatogram trend classification result after attention expansion and long and short term memory network processing. According to the method, structured processing and basic trend extraction of data are realized, adaptive matching and weighted aggregation of historical operation modes are realized, and a multi-dimensional dependency relationship and time sequence dynamic change are captured, so that accurate classification of oil chromatogram trends is realized.

The invention discloses a low-light image enhancement method based on multilevel feature fusion. The low-light image enhancement method comprises the steps of acquiring a data set, dividing the data set, extracting features, constructing a synchronous multi-scale network, training the synchronous multi-scale network and testing the synchronous multi-scale network. According to the synchronous multi-scale low-light image enhancement method in combination with the Laplacian pyramid, the input image is processed in parallel by adopting a double-path structure: the preliminary enhancement image is obtained through the local-global convolutional neural network, and the detail and texture information of the image is enhanced based on the Laplacian pyramid decomposition network. A multi-scale network is adopted to process images in scenes of deblurring, defogging, rain removal, low light enhancement and the like, and details and features are extracted in a layered manner, so that the definition, color and contrast ratio of the images are effectively improved. Comparison experiments prove that the method has the advantages that noise is effectively suppressed, and remarkable effects are achieved in the aspects of detail recovery and color restoration. The method is suitable for image enhancement processing under various complex illumination conditions.

The invention provides a landslide mass recognition method and device, equipment, a storage medium and a program product, and can be applied to the field of landslide mass recognition. The method comprises the following steps: based on a landslide mass identification model, respectively carrying out feature extraction on images from a plurality of modalities to obtain extraction features corresponding to the images; converting the extracted features into corresponding frequency domains and carrying out frequency domain compression to obtain frequency domain features corresponding to the image; converting the frequency domain features into weight factors, and obtaining weighted features corresponding to the images according to the weight factors and extraction features corresponding to the same image; splicing the weighted features of the images to obtain fusion features; processing the fused features through a multilayer dense network, and carrying out multiple interpolations on the processed features by using a bilinear interpolation method; and obtaining the category of each pixel in the features after multiple interpolations by using a maximum parameter value decision method, and determining the landslide mass.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating an output using a generative neural network. One of the methods include: obtaining a prompt input; determining, for each of a plurality objectives that correspond to different aspects of an output to be generated by a generative neural network, a weight to assign to the objective; generating a prefix input that defines the weight assigned to each of the plurality of objectives; and processing, using the generative neural network, the prefix input and the prompt input to generate the output.

The invention relates to the technical field of network security, in particular to a security monitoring method and system for network traffic. The method comprises the following steps: capturing a network flow data flow in real time, and extracting a network entity and a direct communication relationship to construct a basic communication graph; identifying and quantifying a high-order interaction mode between network entities, and taking the high-order interaction mode as an implicit feature enhanced basic communication graph to generate an enhanced security graph; processing the enhanced security map by using a multi-scale time sequence diagram neural network, and capturing a short-term burst mode and a long-term evolution mode at the same time; a dynamic anomaly score is calculated based on the network entity historical behavior baseline and the current network situation, and a security alert is generated when an adaptive threshold is exceeded. The system correspondingly comprises a flow capture module, a feature extraction module, a high-order mode analysis module, a security map construction module, a multi-scale analysis module, a dynamic risk assessment module and an intelligent alarm module, and comprehensive and accurate network threat detection is realized.

The invention discloses a Raman spectrum joint classification and quantification method based on a prior fusion double-flow network, and belongs to the technical field of spectral analysis and chemometrics. The method comprises the following steps: acquiring Raman spectrum original data, and preprocessing to obtain standardized spectrum data; raman spectrum prior information is extracted, and a prior information double-flow network containing classification and quantitative flow sub-networks is constructed; standardized data and prior information are input, classification preheating-combined fine tuning training is adopted, conflict is relieved by combining dynamic weighting and PCGrad, and samples are expanded through a physical consistency strategy synchronously; processing unknown samples by using the trained network, and outputting categories and component contents; according to the method, weak peak information is mined through double-flow architecture and cross-flow attention, derivative noise is suppressed, classification-quantification performance is balanced through double-task optimization, and small sample overfitting is solved through physical expansion; on a polycyclic aromatic hydrocarbon data set, classification and regression indexes are superior to those of a traditional method, the low-concentration / weak-peak scene precision and robustness are better, and the engineering application prospect is good.

The invention discloses a fixed-wing unmanned aerial vehicle distributed formation and obstacle avoidance method based on diffusion reinforcement learning, and the method comprises the steps: processing single observation data through a graph attention network, and enabling the single observation data to serve as a condition for generating an optimal action in a reverse denoising process; the method is suitable for distributed cluster control of a fixed-wing unmanned aerial vehicle cluster in an unknown and disordered environment, treats huge challenges brought by complex dynamics and non-integral constraints in the cluster, and effectively solves a series of problems that an existing method is difficult to capture multi-modal action distribution necessary for robust obstacle avoidance under an uncertain condition. A large number of numerical simulation results show that the method is obviously superior to an existing baseline method in the aspects of flight stability and collision rate, the potential of a diffusion model in the aspect of extensible and robust unmanned aerial vehicle flight control is highlighted, and the method has good theoretical popularization value and engineering application prospects.

The invention discloses an SAR small target ship detection method and system based on a three-branch attention module and a mixed measurement mode, and belongs to the technical field of target detection. In order to solve the problem that the effect is poor when an existing network model detects a small target in an SAR image, a small target enhancement detection network is adopted for detection, a backbone network is used for feature extraction and sending feature maps of different levels into a Neck network for processing, and the backbone network comprises a plurality of convolution modules and at least four CCA modules. The CCA module comprises three parallel branches, the first branch and the second branch comprise a CBAM module and a CBS module, the output of the first branch and the output of the second branch are spliced and then sent to the CBS module and the convolution module, and then are spliced with the third branch storing the original feature map to obtain the output; the Neck network carries out bidirectional cross-layer feature fusion, the detection network carries out complete target detection based on three detection heads, and in object detection and bounding box regression tasks, an MIoU is adopted to carry out target positioning.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating content items based on source document metadata using a generative neural network. One of the methods include: receiving, from a user, a request to generate a content item using a generative neural network conditioned on a context input, wherein the context input comprises content derived from a source electronic document; obtaining metadata associated with the source electronic document; generating a prompt for the generative neural network based on the context input and the metadata associated with the source electronic document; processing the prompt using the generative neural network to generate the content item; and providing the content item for presentation to the user.

An image tampering detection method, applicable to the technical field of artificial intelligence (AI). In the image tampering detection method, when a feature of an image in which a tampered region needs to be detected is extracted, the image feature is processed by means of a gating network, so as to obtain a confidence level of each expert network among multiple expert networks processing the current image feature, and then which expert networks among the multiple expert networks are used for targeted processing of the image feature is determined. Thus, the image processing strategy is dynamically adjusted on the basis of the tampering characteristics of images themselves, avoiding the use of a network having a fixed structure to uniformly process images of various tampering types, improving the effect of image tampering detection.

According to one example of the present disclosure, a system may be provided. The system may comprise at least one processing core configured to process computations of the at least one neural network model comprising at least one tensor, at least one memory circuit configured to store the at least one tensor, a bus circuit, electrically coupled to the at least one processing core and the at least one memory circuit, configured to transmit the at least one tensor based on a memory access operation instruction, and a controller configured to control a priority of a memory access operation for each tensor of the at least one processing core.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating answers to medical questions using neural networks and other components. In one aspect, a method includes: obtaining question data representing a medical question; obtaining a plurality of document snippets from a medical database that stores medical documents; for each document snippet in the plurality of document snippets, determining a relevance score for the document snippet by using a ranking neural network based on the document snippet and the medical question; selecting, based at least in part on the relevance scores, a subset of the plurality of document snippets; generating a prompt that includes (i) the medical question and (ii) the subset of the plurality of document snippets; and generating an answer to the medical question based on processing the prompt using a generative neural network.

The invention discloses a laying hen genetic disease molecular marker screening system based on data fusion and AI prediction, the system comprises six modules, a multi-omics data acquisition module obtains laying hen genome and transcriptome data, and a FineDataLink data fusion module carries out feature alignment and association mapping to generate a fusion feature matrix; the dynamic time sequence diagram neural network processing module constructs a time sequence association diagram and outputs a time sequence feature vector, and the attention enhancement deep forest analysis module evaluates feature importance and outputs a screening result; the federal variation auto-encoder modeling module constructs a federal training framework to generate a molecular marker probability distribution model, and finally the molecular marker screening output module extracts key molecular markers. The system realizes deep fusion of multi-omics data and efficient application of an AI algorithm through multi-module cooperation, improves the molecular marker screening efficiency and accuracy, and provides technical support for disease-resistant breeding of laying hens.

A non-transitory computer-readable medium storing instructions which, when executed by a processor, cause performance of a method of notifying a user of a consolidated event, the method including: receiving a first datum from a first source and a second datum from a second source, wherein the first and second sources are both connected to a network; processing information associated with the first datum to identify an event information associated with the first datum; processing information associated with the second datum to identify an event information associated with the second datum; comparing at least a portion of the event information associated with the first datum with at least a portion of the event information associated with the second datum; determining occurrence of a consolidated event based on the comparing; determining a notification based on the determined consolidated event; and causing transmission of the determined notification to a user device.

The invention discloses an ocean variable prediction method and system based on space-time coherence, and the method comprises the steps: carrying out the frequency spectrum transformation of an input tensor, obtaining a frequency spectrum transformation result, fusing the frequency spectrum transformation result with a Coriolis parameter, and obtaining an enhanced frequency domain feature; performing multi-scale wavelet decomposition on the enhanced frequency domain features to obtain multi-scale features; performing convolution processing on a depth variable in the input tensor to obtain a vertical mixing feature, and encoding different depth layers of the depth variable into depth embedding; fusing the vertical mixing feature, the depth embedding feature and the multi-scale feature to obtain a depth fusion feature; calculating a multi-scale evolution feature based on the constructed physical constraint item and the deep fusion feature; performing weighted fusion and fusion network processing on evolution characteristics of different scales in the multi-scale evolution characteristics to obtain space-time coherent characteristics; and predicting target ocean variables based on the space-time coherent features. The ocean variable prediction precision can be improved.

The invention discloses a low-altitude defense scene low-slow small target identification method, a terminal, a medium and a product. According to the method, firstly, a video frame is processed through a multi-scale feature adaptive target detection network, the network integrates a global self-attention mechanism to capture a remote dependency relationship, shallow details and deep semantic information are fused by adopting a dynamic weighting multi-scale feature fusion structure, and a target bounding box and a category are output in combination with an optimized loss function. And then, a time sequence level multi-target identity keeping and trajectory generating module is used, stable association of cross-frame target identities is realized through fusion of Kalman filtering motion prediction and appearance feature matching, and a target life cycle is managed in cooperation with a trajectory maintenance mechanism. According to the method, the problems of low-speed small target feature weakening, complex background interference, unstable multi-target tracking and the like are effectively solved, the recognition precision, the anti-interference capability and the tracking continuity are remarkably improved, and meanwhile, the low-altitude defense real-time requirement is met.

The invention relates to the technical field of video analysis and processing, and discloses a key frame extraction method based on hierarchical attention and potential gating and related equipment. The method comprises the following steps: analyzing an original video stream and separating frames, adjusting the resolution and normalizing to obtain a standardized sequence; inputting the standardized sequence to a multi-stage convolutional neural network encoder, generating multi-stage spatial features, and obtaining a spatial attention map through attention mapping; pooling the lowest resolution feature to obtain a frame-level embedded vector sequence, and performing time sequence conversion to obtain time sequence attention association information; constructing a potential gating sub-network containing a strategy network, wherein the strategy network outputs spatial feature weights of all levels; the fusion unit sums the spatial features and the corresponding attention maps according to weights, and fuses time sequence results to obtain a fusion feature sequence; and outputting a frame importance score after the scoring network processing, and screening according to the importance score to obtain a target key frame. According to the method, key frame extraction accuracy is improved, and redundant frames are effectively reduced.

The invention relates to an AMT signal denoising method and device based on self-adaptive multi-level U-Net. According to the method, noise segment classification is performed on a noisy AMT signal, a corresponding position mask of the clean AMT signal is recorded, and noise positioning information is provided for training of a denoising model, so that the training effect of the denoising model is improved, and loss of non-noise segments can be avoided; the method also uses two U-Net networks to form a U-Net cascade structure, the first network performs preliminary denoising, the second network processes fragment data output by the first network, generates a noise intensity probability by using a residual error, uses the noise intensity probability as a driving signal, distributes a corresponding weight value, and outputs the driving signal to the U-Net cascade structure. According to the method and the system, the second network is dynamically controlled to selectively absorb and fuse fragment data output by the first network, an area with remarkable residual noise can be adaptively concerned, complementation and optimization of a feature level are realized, and finally, the denoising effect of the denoising model on the noisy AMT signal is improved.

The density and power consumption of modern integrated circuits, such as Graphic Processing Units (GPUs), Central Processing Units (CPUs), and Network Processing Units (NPUs) is growing rapidly, which necessitates designing advanced cooling systems. Existing solutions for characterizing and validating these cooling system are inadequate. A flexible, scalable Thermal Test Vehicle (TTV) is disclosed which is based on an array of power transistors, measurement / control circuitry, and onboard computer. The TTV is configured for characterizing the performance of electronic cooling solutions under a variety of operating conditions.

The invention discloses a filler ratio optimization modeling method for removing tire particle dissolved substances in a rainwater system, and the method comprises the steps: carrying out the structural processing of obtained parameters, and constructing a structural parameter set; calling a mechanism library and a medium fingerprint library, processing data by adopting a mechanism constraint neural network (PINN), and generating a multi-component dissolution spectrum; the method comprises the following steps: describing migration and removal processes of pollutants in a packed bed layer through a one-dimensional convection-dispersion-reaction bed layer model, treating a competitive effect among the pollutants by adopting a multi-component competitive adsorption model, and outputting a dynamic penetration curve, a removal rate and pressure drop data of the pollutants; then carrying out multi-objective optimization and robustness analysis to obtain an optimal filler combination structure and operating parameters; and finally, outputting and visualizing a result. According to the method, the multi-component competition effect can be accurately quantified, filler proportion optimization modeling is carried out on the basis, the limitation of the prior art can be overcome, and the reliability, economical efficiency and environmental safety of treatment facility design are improved.

The small sample anti-migration prediction method suitable for the metallurgical process end point component under the zero expansion characteristic comprises the steps that smelting report data of a large sample steel grade and a small sample steel grade in the metallurgical process are collected to serve as source domain data and target domain data; adopting median to fill and restore abnormal values existing in the report data; dividing the source domain data and the target domain data into continuous feature variables and classification feature variables; inputting the continuous feature data and the classification feature data of the source domain data into a TabNet coding and decoding self-supervising network for self-supervising training, and inputting the target domain data into the trained self-supervising network for feature extraction and reconstruction; introducing an adversarial network, and gradually aligning feature distribution of a source domain and a target domain; and inputting the high-dimensional reconstruction features processed by the TabNet coding and decoding self-supervised network in the source domain into the deep table network model for training, and inputting the small sample steel grade features aligned by the adversarial network distribution into the trained deep table network model for transfer learning.

The invention discloses an online training course personalized recommendation method and system, and relates to the technical field of online education. The method comprises the following steps: acquiring multi-modal learning interaction data of a user, and constructing a course knowledge graph comprising courses, knowledge components and relationships between the courses and the knowledge components; processing the course knowledge graph by adopting a heterogeneous graph attention network so as to track the dynamic knowledge state of the user; estimating the cognitive load level of the user through a pre-trained cognitive load classification model based on the multi-modal learning interaction data; and based on the dynamic knowledge state and the cognitive load level, generating a personalized learning path in a deep reinforcement learning framework by maximizing a cumulative reward function combining knowledge gain reward and cognitive load balance reward.