787 results about "Feature Dimension" patented technology

The invention belongs to the field of environment monitoring and computer vision, and particularly relates to a water body color recognition regression method and system based on space-time causality and manifold learning, and the method mainly comprises the steps: carrying out the detection of a current target water body video sequence, extracting a water body region, carrying out the high-dimensional feature dimension reduction of the water body region, and obtaining a water body color recognition result; and performing feature extraction on the water body region through a space-time causal feature learning model, fusing the flow shape learning features and the space-time causal features to obtain fused features, and outputting a finally predicted water body color value. According to the method, end-to-end assembly line design of preprocessing-segmentation-feature modeling-regression is adopted, manual intervention is not needed from video input to color prediction, and through cascade cooperation of five core modules (video preprocessing, water body segmentation, manifold learning, time sequence causal modeling and color recognition), the real-time performance of the system is improved. Full-link automation from environmental interference suppression, feature extraction to result output is realized, information loss of intermediate links is avoided, and recognition efficiency and robustness are improved.

The invention discloses an underground engineering lining disease detection system based on point cloud, and belongs to the technical field of underground engineering detection. In order to solve the technical problems that an existing underground engineering lining disease detection method is low in detection precision, low in automation degree and the like, underground engineering point cloud data to be detected and corresponding position information are collected, and an improved PointNet + + model is adopted for disease recognition. According to the method, original three-dimensional coordinates of a point cloud are expanded into seven-dimensional point cloud data containing coordinates, normal vectors and reflection intensity, the normal vector standard deviation of points in a neighborhood of each candidate point is calculated to serve as local geometric complexity, a local geometric complexity index is fused into sampling distance measurement, then a multi-scale local neighborhood is constructed by combining sphere query, and therefore the multi-scale local neighborhood is obtained. And extracting geometric features, texture features and deformation features by using a PCA feature dimension reduction technology, carrying out feature fusion based on a normal vector weighting mechanism, finally obtaining disease type classification based on a network model, and calculating the size and position of the disease.

The application discloses a feature editing method for large model content security, which compares and analyzes the sparse coding features of a chat assistant constructed based on a large language model under positive user input and negative user input, extracts the internal response differences of the model to different semantic directions, and the mechanism can automatically and accurately identify the key feature dimensions highly related to the semantic direction of the target attribute. The model activation is mapped to a sparse feature space by using a sparse autoencoder, and each dimension of the feature has independent and interpretable semantic meaning. By injecting a feature guide vector in the space, the interference of the control process on the text grammar, fluency and information density is significantly reduced. The sparse representation mechanism is introduced to structure the intermediate activation features in the reasoning process of the large language model and to intervene in a targeted manner, so that the reply of the chat assistant to the user input conforms to the preset safety specification, and the safety and controllability of the chat assistant in the interaction with the user are improved.

The invention discloses a quantum fuzzy neural network adaptive to high-dimensional input and a classification method, and relates to the field of quantum calculation and fuzzy neural networks and the field of computer vision. The network input layer receives high-dimensional data, amplitude coding, forward and reverse enhanced chain entanglement layer, parameterized quantum transformation and fuzzy set mapping are carried out through a quantum fuzzy feature extraction module, and dynamic dimension fuzzy features are output; high-dimensional neural features are extracted through a DNN feature extraction module to adapt to quantum fuzzy feature dimensions; dynamically distributing the weights of the quantum fuzzy features and the classic neural features through an adaptive feature fusion module; and carrying out Softmax classification on the fusion features through a classifier, and outputting a category probability. According to the method, the high-dimensional data coding efficiency can be effectively improved, the complex fuzzy logic relation learning capability of the quantum part and the quantum state correlation stability are enhanced, the uncertainty of the data is represented, and accurate classification of high-dimensional uncertainty images is realized while noise interference is reduced.

The invention discloses a multi-modal operation and maintenance data method and system based on a large model, and the method comprises the steps: carrying out the preprocessing of obtained multi-modal operation and maintenance data, so as to obtain data feature sequences with the same feature dimension corresponding to different-modal operation and maintenance data; fusing the data feature sequences with the same feature dimension corresponding to the operation and maintenance data in different modes; processing the fusion vector, inputting the processed fusion vector into the fine-tuned large model, and outputting a semantic reasoning result; analyzing the semantic reasoning result into data in a structured form; and secondary verification is carried out on the abnormal event in the data in the structured form to determine final fault information, and the fault information is expressed in the structured form. A multi-source data fusion strategy is combined with large model reasoning, so that storage, retrieval and analysis of massive heterogeneous operation and maintenance data are both real-time and accurate. Depending on the reasoning ability of a large model, the system can timely give a problem analysis result, the fault processing speed is accelerated, and the service quality is ensured.

The invention relates to the technical field of low-altitude equipment, in particular to a low-altitude resource intelligent scheduling method and system based on deep learning, and the method comprises the steps: collecting the real-time state and network load data of a low-altitude flight equipment group, and constructing a dynamic operation data set; generating an operation mode feature set through multi-dimensional airspace situation awareness and analysis, and performing sparse clustering division based on the feature set to form a network resource demand priority mapping table; traversing the mapping table to dynamically calculate the resource demand, determining a multi-dimensional weight coefficient, and performing high-dimensional feature dimension reduction and optimization through a mixed integer nonlinear programming solver to obtain a resource demand feature vector; constructing a resource scheduling strategy optimization model by adopting a deep reinforcement learning algorithm based on the vector; inputting real-time data into the model to execute a resource scheduling decision, and outputting a dynamic allocation strategy; simulation deduction and compliance verification are carried out on the strategy in the digital twin simulation platform, and cooperative intelligent scheduling of communication, calculation and spectrum resources is achieved.

The invention discloses an intelligent scheduling method for unmanned workshop production, and the method comprises the steps: collecting the feature data of task orders, equipment loads and the like in real time, carrying out the normalization and feature dimension reduction processing, and constructing a standardized task state vector; constructing a task-resource-state causal graph based on a Bayesian network, and identifying and quantifying key factors influencing the scheduling performance; in combination with causal reasoning and anti-factual reasoning, multi-strategy generation and screening under new tasks and abnormal working conditions are realized, and online fine adjustment is performed on selected strategies through a reinforcement learning model; feedback data are collected after scheduling is executed, the causal model is dynamically corrected, continuous self-evolution and generalization ability improvement of the strategy are achieved, and the production flexibility, the resource utilization rate and the scheduling robustness of an automatic workshop can be effectively improved.

The invention discloses a power distribution network transient characteristic prediction method based on supervised learning, and relates to the technical field of power distribution network state prediction, and the method comprises the steps: collecting historical operation data through a power distribution network monitoring system, carrying out the data preprocessing, and obtaining standardized multi-dimensional time series data; carrying out transient feature extraction, constructing a high-dimensional feature set, and carrying out feature dimension reduction according to a transient event tag to generate a feature subset; inputting the feature subset into a mixed supervised learning model of a gradient boosting decision tree GBDT and a long short-term memory network LSTM for joint training to obtain a transient feature prediction result; and calculating a root-mean-square error according to the transient characteristic prediction result and the real-time monitoring observation value of the power distribution network, and dynamically adjusting hyper-parameters of the supervised learning model based on a Bayesian optimization algorithm. According to the method, the detection accuracy can be improved, the calculation complexity can be reduced, and the discrimination capability and the time sequence prediction capability of the model are considered.

The invention relates to the technical field of big data analysis, and discloses a prospecting target area positioning method and system based on big data, and the method comprises the steps: collecting multi-source exploration data in real time through distributed nodes, completing coordinate normalization, semantic alignment and time synchronization through a spatial heterogeneous data flow engine, and generating a standardized incremental data block; performing local feature sensitivity analysis based on the historical model library, identifying a newly added feature dimension, and performing parameter increment updating by adopting a sliding window gradient descent method; inputting the updated model into a target evolution model driven by a Bayesian space-time probability field, and dynamically calculating the metallogenic probability of each space grid in combination with a stress field, an element migration path and historical verification data; and generating high, medium and low three-level target area maps according to probability sorting, and pushing the high, medium and low three-level target area maps to a three-dimensional visual decision terminal. According to the method, minute-level dynamic response of the target region under triggering of newly-added data is realized, computing resource consumption is reduced to be less than 5% of that of an original system, and prospecting efficiency and abnormal region identification timeliness are improved.

The invention discloses an image data management method and system for high-precision size visual inspection, and relates to the technical field of image data, the image data management method for high-precision size visual inspection comprises the following steps: S1, collecting a data set, and carrying out region segmentation processing; s2, correcting and enhancing the data set, and extracting contour feature parameters; s3, mapping association is carried out, and screening classification is carried out; s4, parameters are optimized and adjusted, and a standardized size detection feature set is generated; s5, constructing a size detection model to obtain a target size quantification detection result; and S6, performing accuracy verification, and establishing an image feature size parameter association database. According to the invention, through synchronous acquisition of a multi-view image data set and a physical size calibration data set, and in combination with calibration reference of a geometric quantity standard appliance, the problems of incomplete information and large physical size mapping deviation of a traditional single-view image are solved.

The invention relates to the field of electric power monitoring, in particular to an electric power monitoring system intrusion detection method and system based on flow analysis. The method comprises the following steps: collecting network traffic, analyzing and recombining to obtain structured session data; time sequence behavior features and function code distribution features are extracted to construct a multi-dimensional feature set; inputting the feature set into a compliance rule base and a behavior baseline model in parallel, and respectively outputting a rule matching result and an abnormal deviation degree score; generating a comprehensive threat index by adopting a weighted decision fusion strategy; and when the index exceeds a dynamic threshold value, intrusion is determined and an alarm is given. According to the invention, the problem of insufficient precision and adaptability caused by single feature dimension and isolated detection mechanism is solved.

The invention relates to the technical field of image analysis, in particular to a vegetation fine classification and recognition method and system based on an unmanned aerial vehicle, and the method comprises the following steps: obtaining a multispectral image through the unmanned aerial vehicle, extracting red edge reflectivity, NDVI and gray-level co-occurrence contrast, generating a feature vector in a standardized manner, calculating neighborhood offset to obtain a dynamic weight, and combining the dynamic weight into a weighted vector; high discrete features are screened as effective channels, multi-scale clustering is carried out, center and region growth extension recognition is optimized, and a vegetation classification atlas is generated. According to the method, a neighborhood pixel feature offset dynamic weight mechanism is introduced, multi-spectral feature dimension contribution degree is adjusted in a self-matching mode, effective channels are screened based on full-image dispersion, redundant interference is eliminated, image pyramid multi-scale clustering and consistency constraint are fused, the complex vegetation boundary recognition capability is improved, dynamic weight and multi-scale optimization are coordinated, and the method is high in robustness and high in robustness. Sample dependence is reduced, and accurate distinguishing of spectrum similar vegetation is achieved.

The invention discloses an interpretable load prediction method based on knowledge distillation and a time self-attention additive neural network, and the method comprises the steps: collecting the historical load and meteorological data of a power grid as the input characteristics of a model, carrying out the detection of a data quartile abnormal value, dividing the data quartile abnormal value into a training set, a test set and a verification set, and carrying out the detection of the data quartile abnormal value; standardization and abnormal value filling are carried out through Z-shaped orthogonalization and linear filling, and finally, a tensor form meeting the model input requirement is converted through a sliding window; designing a knowledge distillation'teacher-student 'framework based on multiple scales and multiple cycles; constructing a time self-attention additive neural network TSA-NAM as a student model; calculating a shape function representing the contribution degree and the characteristic value in the sub-network to obtain the interpretability of the characteristic dimension; exporting the attention weight of the time self-attention module to obtain the interpretability of the time dimension; performing simulation verification; according to the method, high reliability and high precision are guaranteed, and meanwhile, multi-dimensional interpretability is brought to power load prediction.

The invention discloses a large model content security-oriented feature editing method, which comprises the following steps of: comparing and analyzing sparse coding features of a chat assistant constructed on the basis of a large language model under positive user input and negative user input, and extracting internal response differences of the model in different semantic directions; the mechanism can automatically and accurately identify key feature dimensions highly related to the semantic direction of the target attribute. A sparse auto-encoder is utilized to activate and map the model to a sparse feature space, and each dimension of feature has an independent and interpretable semantic meaning. By injecting a feature guide vector into the space, the interference of a control process on text grammar, fluency and information density is remarkably reduced. A sparse representation mechanism is introduced, structural modeling and targeted intervention are carried out on intermediate activation features in a big language model reasoning process, replies input by a chat assistant to a user are guided to conform to a preset safety specification, and the safety and controllability of the chat assistant in interaction with the user are improved.

The invention provides a clinical test data security sharing method and device based on a block chain. The method is applied to the technical field of data processing, and comprises the following steps: extracting feature dimensions corresponding to various types of data, and calculating corresponding data sensitivity coefficients; determining a data sensitivity level of the corresponding data category; extracting corresponding associated metadata; selecting a corresponding encryption algorithm to encrypt the data, and generating corresponding encrypted data and a data abstract; uploading the associated metadata, the encrypted data and the data abstract to a block chain node, constructing a clinical test data sharing account book, and recording a timestamp and a node signature of data operation; extracting identity authentication information and historical access records of visitors, and calculating access credibility; judging whether the visitor is allowed to acquire the decryption key or not; after the visitor obtains the encrypted data, the integrity in the data transmission process is verified based on the non-tampering property of the block chain. Therefore, the security of clinical test data sharing based on the block chain is improved.

The invention discloses a tractor transportation operation working condition construction method based on improved particle swarm optimization and KMeans fusion, and relates to the technical field of agricultural machinery working condition analysis. The method comprises the following steps: acquiring original data of tractor transportation operation through a plurality of data acquisition modes, and carrying out preprocessing and three-stage screening to obtain an effective kinematics fragment; selecting multi-dimensional characteristic parameters to construct a characteristic matrix, and performing data dimension reduction by adopting principal component analysis; optimizing a KMeans clustering initial center by using an improved particle swarm optimization (IPSO) algorithm which introduces a dynamic inertia weight and a Gaussian mutation strategy, and performing clustering analysis on the feature space after dimension reduction; and selecting representative fragments based on feature similarity, and synthesizing a standardized working condition curve by taking the sum of average relative errors of all feature dimensions as a target function. According to the method, the problems that a traditional clustering algorithm is prone to falling into local optimum and the working condition construction accuracy is insufficient are solved, the constructed working condition can truly and comprehensively reflect the actual transportation operation characteristics of the tractor, and a reliable basis is provided for tractor power system optimization, operation efficiency improvement and energy consumption reduction.

The invention relates to a low-dimensional subspace clustering method based on projection matrix guidance, and the method comprises the steps: extracting a light response non-uniformity PRNU noise residual error from input image data through employing a denoising filter, and constructing a PRNU feature data set of an image; performing feature dimension reduction on the feature data set by adopting a projection matrix method, constructing a projection matrix maintaining a geometric structure, mapping the projection matrix to a low-dimensional potential subspace, and further constructing a model for the subspace by utilizing a sparse self-representation method; constraint is applied to sparse self-representation in the low-dimensional potential subspace, and joint optimization is carried out through an augmented Lagrange multiplier method ALM and an alternating direction minimization ADM strategy to be used for efficient clustering of data in the low-dimensional potential subspace. According to the method, the projection matrix maintaining the geometric structure is constructed, the high-dimensional PRNU features are mapped to the low-dimensional potential subspace, the local neighborhood relation and the global distribution structure are reserved in the dimension reduction process, the calculation cost is reduced, and the clustering robustness and performance are effectively improved.

The invention discloses a power load prediction method based on a dynamic expert pool and a load balancing mechanism MoE, and belongs to the field of power load prediction, and the method comprises the following steps: collecting multi-dimensional input data needed by power load prediction, detecting and repairing an abnormal value in the input data, and obtaining a power load prediction result; constructing a time-feature matrix by using the repair data; projecting the time feature matrix into three subspaces of Q, K and V, calculating attention weights among feature dimensions, and obtaining enhanced features through attention weighting; the enhanced features are input into a routing layer of the MoE model, selection probability distribution of experts is calculated, the MoE model adopts a dynamic expert pool and introduces a load balancing mechanism, and a Top-2 expert selection strategy is randomly distributed in the reasoning stage; and when the MoE model is migrated to a new scene, updating the attention weight of the MoE model by adopting a meta-learning strategy. According to the method, complex factors influencing the power load can be comprehensively captured, and the power load prediction precision and the cross-scene adaptive capacity are remarkably improved.

The invention discloses a sea-air target trajectory prediction method and device based on dynamic graph learning and a generative adversarial network, and a medium, and belongs to the technical field of sea-air target trajectory prediction, and the method comprises the steps: S1, data preprocessing; s2, dynamic graph learning and generative adversarial network (GAN) training are carried out, feature dimension relevance is modeled in real time through a dynamic graph structure, and a GAN module is utilized to generate adversarial trajectory features; s3, CNN-Transform hybrid model prediction is carried out, local feature extraction is carried out through a CNN module, global dependency modeling is carried out through a Transform module, and a decoder part is associated with a historical trajectory and future prediction through a cross attention mechanism; and S4, performing visualization and evaluation based on dimension reduction. According to the invention, the precision, robustness and interpretability of trajectory prediction in a complex scene are significantly improved.

The invention discloses a landslide displacement double-layer fusion prediction method and model, and the method comprises the steps: carrying out the decomposition of an original displacement time sequence through employing an ICEEMDAN algorithm, and obtaining a plurality of IMF components; performing feature engineering on each IMF component, representing a displacement trend by adopting a trend slope and a window mean value, representing mutation early warning by adopting kurtosis and a frequency spectrum entropy, representing a period rule by adopting a main frequency and a zero-crossing rate, representing system stability by adopting a sample entropy and a standard deviation, and constructing a three-dimensional feature space fusing a time domain and a frequency domain; data standardization is carried out on the extracted features, the interference effect of dimensions on the model is eliminated, and it is ensured that all feature dimensions are within a unified calculation scale range; a CNN-BiLSTM model is constructed for each IMF component; a CPO algorithm is used to optimize the CNN-BiLSTM model; according to the method, the data acquisition difficulty during model training and use can be reduced, and the usability of the model in actual deployment is enhanced; and the prediction precision and the accuracy of landslide displacement prediction are improved.

The invention discloses a multi-modal time sequence fusion voice-driven gesture generation method, which comprises the following steps of: firstly, learning compact discrete representation of gesture motion through a vector quantization variational auto-encoder model, and constructing a quantized potential space for a subsequent generation task; extracting audio features of the voice audio through an audio encoder; performing low-dimensional embedding on the identity of the speaker to obtain an identity feature; performing time sequence alignment on the audio features and the historical gesture sequence and then splicing the audio features and the historical gesture sequence along feature dimensions to form multi-modal initial representation; deep feature fusion is carried out through a multi-modal time sequence fusion module integrating a self-attention mechanism, a cross attention mechanism taking identity features as conditions and a Mamba module; and finally, reconstructing a target gesture sequence through a pre-trained decoder. According to the method, the technical problems of insufficient multi-modal fusion, low calculation efficiency and single generated action are solved, and the gesture animation which is natural, smooth and personalized and meets the real-time interaction requirement can be generated.

The invention discloses a remote sensing image small target detection method and system based on improved Swin Transform, and the method comprises the steps: obtaining a high-resolution remote sensing image data set, carrying out the preprocessing, constructing an improved Swin Transform backbone network, and enhancing the small target feature capture through a dynamic window mechanism and cross-window residual connection; then, designing a multi-scale adaptive feature fusion module, unifying multi-level feature dimensions and utilizing learnable weights to dynamically weight and fuse, and optimizing small target semantic expression; constructing a lightweight three-branch detection head, and respectively executing target classification, bounding box regression and small target existence prediction; and finally, performing model training by utilizing the training set, optimizing hyper-parameters in the verification set, and outputting a detection result in the test set. The method can effectively solve the problems of feature loss, high calculation complexity and background interference in small target detection of a traditional method.

The invention provides a time sequence prediction method and device for performing multi-level text alignment by using a large model, and belongs to the technical field of time sequence prediction of a rail transit system based on the large model. Comprising the following steps of splitting multivariable time sequence input into a plurality of univariate time sequences according to feature dimensions, performing additive decomposition on each univariate time sequence, and performing fragmentation processing on each decomposed time sequence component; embedding the fragments into a text embedding space of a pre-training language model and aligning the fragments with the text embedding space; combining the structured prompt with the aligned time sequence representation to form input of a large model; and feeding the input combining the prompt and the alignment representation into the frozen large language model, obtaining an output representation of the model, and mapping the output representation into a final prediction result through a linear projection layer. According to the method, the time sequence data and the natural language modality are effectively aligned and fused, and the prediction accuracy and interpretability are remarkably improved.

The invention belongs to the technical field of fault diagnosis, and discloses a thermal defect identification method and system for a high-voltage switchgear, and computer equipment, and the method comprises the steps: firstly segmenting an infrared image through employing a transfer learning optimized Mask R-CNN model, reducing the dependence of annotated data through sharing pre-training parameters, and achieving the region extraction of pixel-level equipment; secondly, multi-dimensional temperature information is extracted in combination with a gray histogram and a gray co-occurrence matrix, and key features are screened through PCA to enhance noise immunity; and finally, the LSSVM is adopted for classification, so that the training efficiency is remarkably improved. According to the method, the equipment area is automatically segmented through deep learning, temperature distribution is quantified in combination with multi-dimensional features, man-made misjudgment is reduced, pre-training model parameter sharing is utilized, new tasks are rapidly adapted in a small sample scene, the generalization ability and efficiency are improved, feature dimensions are compressed through principal component analysis, the real-time monitoring requirement is met, and the monitoring efficiency is improved.

The invention discloses a document image tampering detection method based on text aggregation and multi-frequency enhancement, and relates to the technical field of computer vision, image forensics and deep learning, and the method comprises the steps: obtaining a to-be-detected original RGB image, carrying out the multi-mode decomposition of the to-be-detected image, and obtaining a DCT coefficient graph, a corresponding quantization table, a high-frequency view and a low-frequency view; and inputting the original RGB image, the DCT coefficient graph, the corresponding quantization table, the high-frequency view and the low-frequency view into a document image tampering detection model for processing, and outputting a final detection result. The document image tampering detection model carries out feature dimension reduction and preliminary text aggregation through the vision-frequency fusion module, carries out coding and fusion through the multi-frequency feature extractor, generates comprehensive frequency features, carries out wavelet transform decoupling through the direction perception frequency decoupling enhancement module, and outputs tampered area masks based on the decoding prediction module. According to the invention, hidden tampering artifacts can be revealed more comprehensively.

The invention discloses a ship engine component cross-working-condition life prediction method and system, and the method comprises the steps: collecting the operation data of key components of a ship engine in real time, and carrying out the preprocessing of the operation data of each component, so as to guarantee the data synchronism and data validity of the operation data; inputting the preprocessed operation data into the corresponding trained life prediction network model, and outputting a life prediction result corresponding to each component; the life prediction network model comprises a feature extraction module used for extracting time sequence features under different scales, splicing all the time sequence features on a feature dimension, performing feature enhancement on a splicing result based on a Transform model, and outputting a global feature vector after compression of a global average pooling layer; the life prediction module is used for outputting a life prediction result according to the global feature vector; the domain discrimination module is used for outputting a domain discrimination result according to the global feature vector; in training, the feature extraction module and the domain discrimination module carry out adversarial training.

The invention provides a full-link data tracing processing method and system for a cross-border home supply chain, and relates to the technical field of supply chain management, and the method comprises the steps: 3, dividing a plurality of data unit sets which are correlated with each other based on the process data of the supply chain, determining a reference data unit set, and obtaining two groups of different data correlation rules; analyzing an interaction relationship between the two groups of data association rules, establishing a data state identification range, and selecting two data verification reference points in the data state identification range; and according to the evolution relationship of the data verification reference point on the time sequence, the evolution span is quantified by calculating the comprehensive distance between two points on the time and data feature dimensions, a data state evolution path is constructed to generate a data calibration coefficient, and a data link track comprising a timestamp and link information is formed. According to the invention, full-link data tracing of the cross-border home supply chain is realized, and the collaboration efficiency of each link of the supply chain is improved.

The invention discloses a quantization method, a medium, computer equipment and a program product. The method comprises the following steps: acquiring a difference matrix between a current parameter matrix and a pseudo quantization parameter matrix of a neural network; the pseudo quantization parameter matrix is obtained by sequentially performing quantization processing and inverse quantization processing on the current parameter matrix; projecting the difference matrix to a plurality of target feature spaces to obtain a plurality of projection difference matrixes of the difference matrix; wherein any target feature space is a feature space composed of at least one feature dimension with the importance from high to low in the original feature space, and the number of the at least one feature dimension is smaller than the total number of the feature dimensions in the original feature space; determining a target projection difference value matrix from the plurality of projection difference value matrixes through grid search, and compensating the pseudo quantization parameter matrix based on the target projection difference value matrix to obtain a compensation parameter matrix; and performing quantization processing on the compensation parameter matrix.

The invention relates to the technical field of cutter residual life prediction, and discloses a cutter residual life prediction method based on line angle attention and contrast drive aggregation, which comprises the following steps of: extracting six types of statistical characteristics including a mean value, a standard deviation, a median, an absolute maximum value, a root mean square and skewness by using original data of a multi-channel sensor in a cutter cutting process; a dual feature dimension reduction strategy of Pearson's correlation coefficient and grey correlation analysis is adopted, key features strongly related to the wear state are screened, and standardization processing is carried out; and constructing a deep learning architecture fusing line angle attention and contrast driving feature aggregation. According to the method, the model has higher recognition capability on the characteristic mode of numerical jump but consistent trend in the tool wear process, the problem that a traditional attention mechanism is prone to losing key time sequence association in the nonlinear degradation process is solved, and the modeling precision of non-stationary sensor data under the complex cutting working condition is remarkably improved.

The invention discloses an image semantic segmentation mask generation method based on non-classifier guidance, and the method specifically comprises the steps: carrying out the standardized preprocessing of an input remote sensing image, and unifying the data format and feature dimension; a remote sensing image is used as input, basic features are extracted through a backbone network, the backbone network selects a lightweight network MobileNetV2, feature expression is enhanced in combination with a filtering mixed attention mechanism and a multi-head attention mechanism, and finally fusion features are output; the fusion features serve as condition input, progressive generation and optimization of semantic segmentation masks are completed through a discrete diffusion model, and finally a high-precision semantic segmentation result is output; and training the model, and outputting a semantic segmentation test result. According to the method, a filtering mixed attention mechanism is introduced, and two-dimensional Fourier transform, CBAM and SE modules are combined, so that the capturing capability of the model on local detail features is remarkably enhanced, and feature extraction effects on different scales are improved through fusion of global context information.