364 results about "Feature transformation" patented technology

A multimodal metadata retrieval-augmented generation method and system, relating to the technical field of artificial intelligence. In the present invention, the method comprises: determining data modalities of query content, performing feature extraction on data of each modality on the basis of the determined data modalities, converting extracted features into feature vectors, and fusing the feature vectors of the modalities to generate a multimodal feature vector; extracting, from data of each modality in the query content, key metadata of the data of each modality, converting the key metadata into a key metadata vector and connecting same to the multimodal feature vector, and constructing an augmented vector; and performing retrieval by means of the augmented vector. The method helps improve knowledge coverage and information utilization efficiency, and enhances the accuracy, relevance, and credibility of generated content.

The invention discloses a visual detection optimization control method, device and equipment based on digital twinning and a storage medium, and relates to the technical field of visual detection, and the method comprises the steps: obtaining an analog image and a real image in a digital twinning environment, and carrying out the frequency domain feature transformation of the analog image and the real image, thereby obtaining an image spectrum feature; a frequency domain alignment model is established based on multi-band spectrum envelope guide residual mapping, and the structure of virtual and real image spectrum features is kept aligned; further extracting features through multi-scale convolution and channel dependence mapping to obtain virtual-real fusion features; quantifying channel similarity and establishing a covariance regularization constraint, performing channel correction on the cross-domain features, and eliminating feature drift to obtain second virtual-real fusion features; and finally, performing visual detection and micro defect identification based on the features. The problem that a virtual sample and a real sample are different in local texture structure and channel distribution is solved.

The invention relates to the technical field of data fusion, in particular to a multi-source electrocardiosignal correction method and system based on adaptive fusion, and the method comprises the following steps: constructing a multi-channel input tensor, extracting local features through a weight calculation network, carrying out the adaptive weight fusion and dimension reduction of multiple paths of signals, and carrying out the correction of the multi-source electrocardiosignal. A nonlinear mapping relation is established through a deep reconstruction network, a standard waveform is reconstructed, and network parameters are optimized based on reconstruction error reverse iteration. According to the method, local neighborhood features of multichannel signals are extracted by constructing a weight calculation network, a dynamic channel weight sequence reflecting the real-time contribution degree of a signal source is constructed, the amplitude intensity is adaptively adjusted according to the signal quality, unstable channel noise interference is effectively inhibited, and high-quality signal components are enhanced; a deep reconstruction network is used for carrying out nonlinear feature transformation on a fusion sequence, accurate mapping from non-standard input to standard lead waveforms is established, and weight distribution and optimization of signal reconstruction parameters are achieved in combination with an error back propagation mechanism.

The invention relates to the technical field of remote sensing image processing, in particular to an image super-resolution system and method based on high and low frequency separation perception Mama, and the method comprises the steps: firstly carrying out the shallow convolution feature extraction of a low-resolution image; then entering a plurality of frequency sensing Mama groups, performing frequency separation and enhancement on each group through a high and low frequency feature adaptive enhancement module, and performing depth feature transformation through a plurality of frequency sensing Mama blocks; the extracted depth features are refined through a global channel-space attention module, and finally a high-resolution image is reconstructed through up-sampling. Through organic combination of the modules, the defects of insufficient frequency perception, low global modeling efficiency, insufficient feature optimization and the like are effectively overcome, and high-quality collaborative reconstruction of remote sensing image structures and textures is realized.

A method and system for decoding speech based on recorded brain signals is provided. The method can include receiving recorded brain signals via a microelectrode array. The method can include extracting one or more features from the recorded brain signals. The method can include converting the one or more extracted features into one or more feature embeddings. The method can include transforming, by one or more encoders, the one or more feature embeddings. The method can include predicting, by one or more decoders, phonemes based on the one or more transformed feature embeddings. The method can include predicting speech based on the predicted phonemes.

The invention discloses an electromagnetic field intelligent calculation method based on deep learning, and the method specifically comprises the steps: inputting a space-time input vector into an MFF-PINN neural network, the MFF-PINN neural network comprises parallel sub-networks and a linear superposition module, the sub-network comprises a scale transformation module, a Fourier feature transformation module and an MLP processing module, and the MFF-PINN neural network comprises a linear superposition module; firstly, scale transformation is carried out on a space-time input vector, then Fourier feature transformation is carried out on the vector after scale transformation, and the Fourier feature transformation module carries out Fourier transformation on the vector after scale transformation based on an effective frequency matrix; all the sub-networks share the Fourier feature transformation, and the output obtained by the Fourier feature transformation is input to the MLP processing module in the first sub-network; and carrying out linear superposition on the output of the sub-networks. According to the invention, the expression capability of the network on the high-frequency component and multi-scale characteristics of the electromagnetic field is obviously enhanced.

The invention discloses an indoor positioning method and system based on 5G channel characteristics and pedestrian dead reckoning, and belongs to the technical field of indoor positioning. The method comprises the following steps: firstly, converting input complex 5G channel impulse response (CIR) data into a two-dimensional characteristic matrix; inputting the two-dimensional feature matrix into a deep convolutional neural network to perform feature transformation and obtain a 5G positioning position track, and then performing time sequence smooth optimization on the 5G positioning position track by using a Kalman filter to obtain an optimized 5G position track; according to the invention, a deep convolutional network is adopted to fully extract amplitude-phase composite features of 5G channel impact response CIR, and machine learning step length measurement and calculation of multi-dimensional gait features are combined to accurately position an indoor position track, so that the utilization efficiency of positioning data is improved; and intelligent complementation can be carried out between the optimized 5G position trajectory and the dead reckoning trajectory, and the indoor positioning precision and positioning stability in a complex indoor environment are improved.

The invention relates to the technical field of computer vision, and discloses a visual language processing method and device based on multi-encoder fusion, and the method comprises the steps: obtaining a to-be-processed visual input image, and determining a corresponding visual coding strategy; based on the visual coding strategy, distributing the visual input image to at least two visual encoders, and generating initial visual features corresponding to the visual encoders; performing consistency processing and fusion processing on all the initial visual features in sequence to generate fusion features; performing feature conversion on the fusion feature to obtain a target visual feature matched with the visual language model input layer; and inputting the target visual features and the text input data into a visual language model for visual text association processing, and generating a visual language task result. According to the method, the multi-source visual encoder is fused, the visual features of different granularities in the image are comprehensively extracted and fused, and the comprehensive perception and understanding ability of the visual language model to the image content can be improved.

An airspace flow prediction method and device based on an ensemble learning algorithm are provided. The method includes the steps: collecting historical airspace flow data and related spatial structure data, and preprocessing; constructing a GNN model, and calculating an influence degree of each node and an influence degree between the nodes in an airspace network by using the GNN model, the node being any airport or any waypoint; performing, by the GNN model, feature conversion and attention fusion on the influence degree of the node, the influence degree between the nodes and time series data to acquire a fused feature vector; inputting the fused feature vector into an LSTM model to acquire a predicted airspace flow of the node; and applying the predicted airspace flow of the node to manage navigation of traffic in the airspace network.

The invention discloses an electric vehicle charging demand prediction method and system and a medium, and the method comprises the steps: calculating loss values corresponding to each alternative source domain and a target domain according to a plurality of alternative source domain covariance matrixes and target domain covariance matrixes, and screening the alternative source domain with the minimum loss value as a source domain; performing feature conversion on the source domain feature matrix, and aligning the source domain feature matrix with the target domain feature matrix to obtain an aligned source domain feature matrix; inputting the aligned source domain feature matrix into an XGBoost model for processing to obtain a first predicted value sequence; performing charging prediction on the historical charging time sequence data after source domain preprocessing by using an LSTM model to obtain a second prediction value sequence; dynamic weights of the XGBoost model and the LSTM model are calculated; performing weighted calculation on the first predicted value sequence and the second predicted value sequence by using the dynamic weight to obtain a final charge quantity demand prediction sequence of the target domain; according to the invention, the problem of poor model prediction precision caused by cross-regional data distribution difference can be solved.

The invention discloses a plasma power supply intelligent regulation and control method based on reinforcement learning, and the method comprises the following steps: collecting the operation state data of a plasma power supply, and carrying out the preprocessing; feature conversion and constraint screening are carried out, and a feasible control domain is constructed and aggregated; feature extraction, index calculation and weight distribution processing are carried out, and a multi-target reward function is constructed; feature modeling is carried out on the improved reversible residual network, and normalization constraint and parameter iteration updating are carried out in combination with a multi-target reward function; collecting multi-source disturbance data to carry out feature extraction and fusion modeling, and carrying out dynamic updating, feature calculation and output mapping on the adaptive strategy network; the control action sequence is issued to the power converter to execute current and voltage inner loop control, a control strategy is evaluated in real time and optimized in a closed loop mode, and a final control strategy is generated. According to the invention, reinforcement learning and the reversible residual network are fused, multi-target adaptive regulation and control of the plasma power supply are realized, and the method has the advantages of high stability, excellent energy efficiency and accurate response.

The invention discloses a meat product fresh-keeping state monitoring and dynamic evaluation system, and belongs to the field of food safety and quality monitoring. Comprising an acquisition module integrating multi-source sensing of gas, temperature, humidity, optics and the like, and multi-dimensional data related to meat product preservation is acquired in real time. The system is provided with a data preprocessing and fusion unit which is used for performing synchronous acquisition, de-noising and feature conversion on original signals and outputting high-quality feature data; and the intelligent monitoring and dynamic evaluation unit realizes high-precision discrimination and deterioration risk prediction of the fresh-keeping state based on a fusion algorithm. Monitoring and evaluation results are uploaded to a cloud analysis platform in a wired or wireless mode, archiving, trend analysis and decision optimization of historical and real-time data are achieved, and quality tracing of the whole process is supported. And the central control and display terminal manages each node in a centralized manner to realize intelligent early warning and visual display.

The invention relates to the technical field of intelligent control and reinforcement learning technologies, in particular to a trajectory tracking control method fusing multi-objective optimization and a physical sensing network, which comprises the following steps: acquiring a real-time state vector of a to-be-controlled object, decomposing the acquired state vector into a sphere dynamic flow and a platform attitude flow, coding features of different attitude flows are extracted, a fusion feature vector is constructed, and at the same time, an attention mechanism is used to carry out feature transformation to determine control decision features; and for the determined control decision features, utilizing a multi-objective optimization function to carry out cooperative constraint on the generated actions, carrying out feature training in combination with an experience playback mechanism and a self-adaptive stable learning mechanism, and after training is completed, determining a trajectory tracking control instruction of the object to be controlled through dynamic adjustment of learning parameters. According to the invention, by establishing an adaptive stable learning mechanism, the learning rate and exploration noise are dynamically adjusted based on performance stagnation detection, and the training stability and convergence speed are improved.

The invention relates to the technical field of industrial simulation, in particular to an industrial simulation optimization method based on multi-modal feature fusion. The method comprises the steps of obtaining multi-source industrial data, performing standardization processing through a data synchronization and timestamp alignment strategy, and performing feature extraction and vectorization by adopting a modal specific algorithm to generate a feature mapping relationship; modal features are obtained from the feature mapping relation for normalization, dimensionality reduction and deep learning model construction processing, and a multi-modal feature library is generated; generating a cross-modal attention weight matrix based on an improved attention mechanism; carrying out feature correlation evaluation, screening and weighted fusion by adopting a graph neural network and Bayesian optimization to generate a fusion feature vector; the fusion features are converted into simulation parameters, and a multi-physics field engine is loaded to execute an optimization algorithm to generate a simulation result; and finally, verifying, evaluating and adjusting parameters through space-time alignment, and generating optimal configuration. According to the invention, the precision, efficiency and adaptive optimization capability of industrial simulation are effectively improved.

A method, according to one approach, includes: causing a received query to be translated from a first language to a second language. The method also includes generating potential answers for the translated query, and extracting features from the translated query. The method also includes causing the extracted features to be converted into feature vectors. The method also includes causing the feature vectors to be compared against existing vectors in a knowledge base that correspond to past question-answer pairs. The method also includes causing the potential answers to be ranked based at least in part on an outcome of comparing the feature vectors against the existing vectors in the knowledge base. Furthermore, the method includes causing a final answer to be generated based at least in part on the ranked potential answers.

The invention discloses a malicious code control flow feature extraction method based on a graph neural network. The method comprises the following steps: constructing a control flow graph, a data flow graph and a function call graph; designing a drawing neural network architecture; training a graph-level classifier; performing graph interpretation by using a GNNExplainer algorithm, attention mechanism analysis and a gradient analysis method; converting the extracted control flow mode into a structured detection signature, and mapping the structured detection signature to an original binary code; and integrating with a static analysis tool through a standardized interface. The invention further discloses a malicious code control flow feature extraction system based on the graph neural network. Multi-level graph structure representation is constructed, important information such as a control flow structure and a data dependency relationship is fully reserved, the deep structure similarity of malicious codes can be recognized, the deformation resistance is higher, and therefore the malicious code detection precision is improved; according to the method, key sub-graphs can be recognized, graph structure features are converted into detection rules, then the detection rules are integrated with existing static analysis tools, and practicability is improved.

The invention discloses an industrial network traffic anomaly detection and source tracing method, device and equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: converting a traffic data package of an industrial network into a process characteristic analysis software package, extracting a plurality of dimension statistical features from the process characteristic analysis software package, and storing the extracted statistical features in a database; converting the statistical characteristics of each dimension into table data; performing classification coding processing on connection behavior features in the table data to obtain a target coding result, normalizing statistical features to obtain a normalization result, converting time features to obtain a conversion result, obtaining a mixed feature vector, constructing a DAGMM model comprising a self-codec and a Gaussian mixture model by using an unsupervised mode, and obtaining a mixed feature vector; performing anomaly detection on the mixed feature vector to obtain an anomaly detection result; and if the traffic is abnormal, performing reverse decoding on the mixed feature vector by using a tracing algorithm to obtain a tracing result, thereby improving the efficiency of performing anomaly detection and tracing on the industrial network traffic.

The invention discloses a light-weight gating characteristic polymerization steel surface defect detection method and device, and relates to the technical field of steel surface defect detection. The method comprises the following steps: acquiring a to-be-detected steel surface image; inputting the steel surface image into a defect detection model obtained by training an improved YOLO11s detection network, and outputting a bounding box, confidence and category of a defect; wherein the improved YOLO11s detection network comprises a backbone network, a detection head and a neck network arranged between the backbone network and the detection head, the backbone network is provided with a gating multi-branch aggregation module, and the gating multi-branch aggregation module is used for enhancing direction texture response and introducing different receptive field contexts through multi-branch feature transformation and gating weighted aggregation in a feature extraction stage; the neck network is provided with a scale perception fusion module which is used for adaptively distributing different scale source feature contributions during cross-layer fusion; the input end of the detection head is provided with a lightweight channel attention enhancement module which is used for re-calibrating channel response before prediction so as to suppress noise textures and highlight defect-related channels. Compared with the prior art, the detection precision and the positioning stability of the multi-type steel surface defects are improved while the light weight of the network is kept.

The invention discloses a railway power distribution network fault information automatic diagnosis and positioning method and system, and belongs to the technical field of power distribution fault diagnosis and positioning, and the method comprises the steps: employing an auxiliary classification generation adversarial network, obtaining a class-balanced standardized training data set, and carrying out the preprocessing; mapping the preprocessed data in the standardized training data set into a snowflake-shaped SDP image under a two-dimensional polar coordinate system, inputting the snowflake-shaped SDP image into a fault diagnosis model based on a deep convolutional neural network, and outputting a fault section and a preliminary fault distance; constructing a physical circuit model of railway power distribution network distribution parameters, and calculating theoretical simulation voltage and simulation current by using a transmission line equation; and starting iterative correction based on the Pearson's correlation coefficient of each data, and outputting a diagnosis result. According to the method, high-precision, high-reliability and self-adaptive fault information automatic diagnosis and positioning are realized by fusing data enhancement, feature transformation, intelligent diagnosis, physical inversion check and topology self-adaptive adjustment.

The invention discloses an original message binary feature extraction method based on a time sequence convolutional network. The method comprises the following steps: generating an embedded vector sequence of a message; constructing a time sequence convolutional network architecture; pre-training and classifier training are carried out through self-supervised pre-training, supervised fine tuning and multi-task learning; model interpretation is carried out, and quantized behavior features are extracted based on interpretation results; converting the features into a Snort / Suricata rule format, and establishing a mapping relation from the features to original message segments; and integration to an IDS / IPS engine is realized. The invention also provides an original message binary feature extraction system based on the time sequence convolutional network. A closed loop from feature discovery to automatic rule deployment is constructed, and the detection response efficiency and accuracy of complex network threats are remarkably improved.

The application discloses a kind of based on timing characteristic analysis one section converter operating abnormality detection method, to be in by analyzing the timing characteristic of one section converter sampling data, complete the real-time abnormality detection of one section converter operating state.Specifically, the method of the present application carries out timing characteristic analysis on multiple time nodes for the sample data collected when one section converter is in normal operation, establishes data model for online real-time abnormality detection after learning corresponding timing characteristic, and then detects whether the operation of one section converter is abnormal through a comprehensive abnormality detection index.Compared with prior art, the method of the present application obtains corresponding feature transformation matrix through timing characteristic analysis process capable of minimizing the difference between score vectors in time sequence, and completes the real-time online abnormality detection of one section converter on the basis of monitoring through the comprehensive abnormality detection index of three different abnormality detection indexes.

The invention provides a component detection method and device, electronic equipment and a storage medium. The method comprises the following steps: acquiring an original image of a component, and preprocessing the original image to obtain a target image; performing feature extraction processing on the target image by using a pre-trained convolutional neural network to obtain multi-scale features; converting the multi-scale features into a plurality of feature sequences, and adding position codes to the plurality of feature sequences to obtain a plurality of coded sequences; performing feature extraction processing on the plurality of coding sequences to obtain a plurality of single-head classification results, and performing fusion processing on the plurality of single-head classification results to obtain a fusion classification result; and determining a target classification result based on the plurality of single-head classification results and the fusion classification result, and determining a detection result of the component based on the target classification result. Therefore, defects existing in the component can be automatically identified, and the problem that the detection result of the component is inaccurate due to inconsistent manual detection standards is avoided.

The invention provides a machine vision coding method based on feature distillation, and relates to the technical field of image processing.The method comprises the steps that an image to be processed is input into a machine vision coding model, and the model extracts first potential feature representation of multiple channels through an analysis encoder; the method comprises the following steps: quantitatively dividing into basic layer quantitative features containing semantic and spatial structure features and enhancement layer quantitative features containing detail and texture features; the hyper-priori correlation module encodes hyper-priori information and generates enhanced auxiliary features and basic auxiliary features, and the conditional entropy coding network realizes encoding and decoding of the basic layer quantization features and the enhanced layer quantization features based on the enhanced auxiliary features and the basic auxiliary features. A machine vision task result is obtained through a feature transformation and task processing module, and after splicing is conducted through a splicer, a reconstructed image is output through a synthesis decoder. According to the invention, both machine vision and image reconstruction can be considered.

The invention provides a data security encryption method and device based on a block chain, relates to the technical field of data security, and solves the technical problems that in the prior art, an encryption strategy is separated from data content, and security and efficiency are difficult to balance. The method comprises the following steps: identifying a modal type of to-be-encrypted original data through an artificial intelligence model, and extracting modal features of the original data; uploading the modal type and the modal feature to a block chain network, triggering a corresponding cross-modal encryption strategy according to a preset rule, converting the modal feature into conversion data of a target modal, and performing hash operation on the conversion data to generate a dynamic encryption key; encrypting the original data by using the dynamic encryption key to obtain a first ciphertext, and generating a second ciphertext based on an attribute-based encryption algorithm; and storing the first ciphertext and the second ciphertext into a distributed storage system, and recording the storage address hash and the second ciphertext into a distributed account book of the block chain network together.

The invention discloses a remote sensing image semantic segmentation method and device based on frequency enhancement and uncertainty perception contrast refinement, and the method comprises the steps: extracting features of an input remote sensing image, and obtaining a patch token sequence; a multi-anchor frequency enhancement module is adopted for processing, and the processing comprises the steps that the features are converted into a frequency domain, a low-frequency component is reserved through a low-pass filter, the features are fused with original features after being inversely transformed back to a spatial domain, anchor point interaction is conducted through cls tokens corresponding to multiple categories, and enhanced feature representation is generated; processing the enhanced feature representation through an uncertainty perception comparison module, including: providing priori knowledge by using a pre-trained teacher model, calculating an uncertainty weight based on a prediction entropy, and constructing a confidence-weighted contrast learning loss to refine the features; and generating a semantic segmentation result. According to the method, noise is suppressed by introducing a frequency enhancement mechanism, category distinguishing is enhanced in combination with multi-anchor interaction, feature representation is optimized by using uncertainty perception contrast learning, and finally more accurate semantic segmentation is realized.

This application provides an adaptive semantic joint source-channel coding method, system, electronic device, and storage medium. The method includes: Step S1: acquiring raw input data and real-time channel state information, wherein the real-time channel state information includes at least the signal-to-noise ratio (SNR); Step S2: using a lightweight semantic coding module to extract multi-scale semantic features from the raw input data, and incorporating the SNR embedding vector in a feature modulation manner during the coding process to obtain channel-adaptive semantic features; Step S3: using an SNR embedding and channel-adaptive attention module to adjust the attention weights of the semantic features to obtain attention features; Step S4: using a dynamic codebook generation and vector quantization module to convert the attention features into a discrete codebook index sequence; Step S5: using a joint source-channel coding module to modulate and map the codebook index sequence into complex channel symbols and transmit them.

The invention discloses a road snow-blown visibility recognition method and system based on image recognition, and the method comprises the steps: carrying out the sequential processing of an original snow-blown image through initial feature extraction, multi-level down-sampling and feature enhancement, and multi-scale context fusion, and directly outputting a visibility interval and confidence, according to the method, downsampling, feature transformation and feature enhancement processing processes are repeatedly executed for multiple times, deep feature maps with gradually reduced scales are sequentially obtained, automatic and objective recognition of the visibility of the whole road line blown snow is achieved, the adaptability to the scene of the blown snow which is high in burstiness and non-uniform in space is improved, and the visibility of the whole road line blown snow is improved. Therefore, the model can more accurately capture the key depth of field and texture degradation characteristics which influence the visibility, and the recognition result is more accurate.

The present application belongs to the field of CT tomographic reconstruction technology and artificial intelligence, and discloses a limited-angle CT reconstruction artifact suppression method based on a multi-domain feature fusion network. In view of the problem that the reconstruction result of traditional CT scanning under limited-angle conditions is prone to artifacts and structural distortion, thereby affecting the image quality and defect detection accuracy, the present application constructs a multi-domain feature fusion artifact suppression network, takes the limited-angle reconstruction result as input, and realizes artifact suppression and detail recovery through the synergistic effect of the encoder part, the decoder part, the feature enhancement part and the feature conversion part. The present application can obtain high-quality tomographic images under limited-angle conditions, effectively reduces the scanning angle and time of industrial CT detection, improves the imaging clarity and reliability without increasing the radiation dose, is suitable for industrial detection of complex structure workpieces, and has important industrial application value.