41results about How to "High discrimination accuracy" patented technology

The invention relates to an RNA-protein binding site discrimination method based on a convolutional block attention mechanism dense network, and belongs to the field of bioinformatics. Firstly, a public RNA-protein binding site data set is acquired, then, a feature space is constructed by adopting four feature extraction methods, and finally, a dense connection convolutional network improved based on a convolutional block attention mechanism is used for learning a long-term dependency relationship in a combination capture sequence of advanced local features, a bidirectional gating circulation network and a self-attention mechanism. The RNA-protein binding site is discriminatively analyzed by creatively using the dense network improved based on the convolutional block attention mechanism in combination with the idea of the bidirectional gating loop network and the self-attention mechanism, richer feature information can be learned and extracted, and then the model discrimination accuracy is improved.

The invention discloses an escalator fault diagnosis method based on multi-sensor data input. The escalator fault diagnosis method comprises the four steps of escalator original data collection, data preprocessing, fault data set setting, escalator fault recognition and cloud model updating. According to the data preprocessing, escalator signal data are filtered through an improved wavelet threshold noise reduction method; the quality of escalator signal data is improved; the fault data set is set by fusing fault data of all parts and all types of the escalator, full utilization of the fault data of the escalator is achieved, and the richness of fault category signals is improved. According to fault identification, multi-sensor data is input into a CNN-GRU network, and the CNN-GRU network extracts features in the input data and obtains a final identification result. According to cloud model updating, the CNN-GRU model is optimized by using CFOA, a network model with the optimal CNN-GRU is constructed, the recognition precision is high, the generalization performance is good, and the method can be widely applied to the field of escalator fault diagnosis.

The invention belongs to the field of computer vision game anti-cheating, and discloses an FPS anti-cheating method based on deep learning and vision, and the method comprises the following steps: 1, game data preprocessing: zooming an input image to the input size of a network, and carrying out the normalization operation; 2, performing feature extraction on the input image: introducing a focus module into a backbone network part to cut the image, and splitting the high-resolution image into a plurality of low-resolution images or feature maps; 3, constructing a Neck module, wherein the Neck module is responsible for further optimizing and fusing the features on the basis of the features extracted by the backbone network; step 4, constructing a Head module; and 5, constructing a classifier module: analyzing the difference of the duration ratio, the frequency ratio and the open mirror killing ratio of effective aiming of the normal player and the plug-in player, and quantifying a data standard to judge whether the player is the plug-in player or not. According to the method, the plug-in players can be effectively detected.

The invention discloses an intelligent correction method and system for correcting paper homework, and the method comprises the steps: constructing a knowledge base according to the examination line content and course knowledge, and obtaining the scanning data of the to-be-corrected paper homework from the outside; performing hyper-parameter fine tuning on the plurality of large models, adding role setting of a corrector to the large models in training, and using an RAG retrieval enhancement technology according to the constructed knowledge base; correcting the paper homework scanning data by using the fine-tuned large models, checking the consistency of a plurality of large models for generated answers by using an LLM confrontation communication method, and performing multiple rounds of challenges to obtain a correction result meeting a consistency requirement; and integrating a final correction result according to the correction results of the plurality of large models, and feeding back the final correction result to the paper job through the printer. According to the method, based on specially designed LLM countermeasure communication, RAG related technologies are integrated, and the model discrimination accuracy is greatly improved; and during exporting, a hardware printing exporting mode is adopted, the method is simple and direct, and examination paper marking correction can be realized.

The invention discloses a construction environment safety belt identification method based on multi-scale fusion, and the method comprises the following steps: S100, constructing an improved yolov8 safety belt identification model based on IAFPN, namely introducing an AIFI mixed feature encoder and a BiFusion module on a high-level feature map for calculation; s200, deploying a reasoning strategy based on an SAHI algorithm, a t-iou matching strategy and a double-camera linkage strategy; s300, collecting and training data, and comparing and judging the tracking target; the method comprises the following steps: improving a yolov8 model through an IAFPN, fusing a Transform layer, optimizing an up-sampling structure and a feature fusion structure, increasing the feature fusion capability of the model for a small target on the basis of reducing the parameter quantity, and improving the recognition precision; based on an SAHI algorithm, a t-iou matching strategy, a safety belt judgment strategy and a double-camera linkage strategy, slice reasoning is carried out on the image, and the judgment accuracy of an angle shielding image target image is improved; and a feature data set is constructed through a large amount of data acquisition and training, so that the overall discrimination accuracy is improved.

The application discloses an intelligent scheduling control method for an automobile part automatic assembly line, and relates to the technical field of automatic control, which comprises the following steps: obtaining the task instruction response time length value and the node execution action state change value of each execution unit in a current scheduling period, and calculating the response delay value and the state fluctuation value of the corresponding node; generating a response efficiency value according to the response delay value and the state fluctuation value of each node, and generating a feedback control level label in combination with the node task attribute information; judging whether to perform compression processing based on the feedback control level label and the available idle time value of each node in the current scheduling period, and if the compression condition is met, encoding the response delay value and the state fluctuation value into a compressed feedback data frame; and completing the relay transmission of the compressed feedback data frame in a time period that meets the scheduling idle condition, so that the scheduling delay pressure control, task state accurate maintenance and instruction issuing sequence dynamic optimization of the automobile part automatic assembly line are realized.

The invention belongs to the field of social network derived topic analysis, and particularly relates to a time-varying network-oriented key influence user discovery method, which comprises the following steps of: preprocessing original topic network data to obtain topic network data; slicing the topic network data along the time dimension to obtain topic network data at different time points; calculating dynamic importance of the user nodes at different time points, and obtaining a transfer sequence of each user node according to the dynamic importance; constructing a feature vector of each user node by adopting an attention mechanism; obtaining an accumulated interaction sequence and a recent interaction sequence of each user node, and obtaining an interest vector of each user node through a dual Transform structure; splicing the feature vector and the interest vector of each user node to obtain a comprehensive feature; inputting all the comprehensive features into a full connection layer and a softmax function layer to obtain a key influence user identification result; according to the method, the key influence elements in the topic space can be effectively found.

The invention discloses a colorectal early cancer infiltration depth prediction model construction method based on mixed cutting enhancement, and the method comprises the steps: carrying out the preprocessing of a historical white light endoscopic image, increasing the sample size on the premise of highlighting a focus region, obtaining a preprocessed sample set, obtaining a focus region ROI in the preprocessed sample set through a doctor and / or employing a feature recognition algorithm, and carrying out the prediction of the colorectal early cancer infiltration depth. The method comprises the following steps: marking category labels of shallow infiltration and deep infiltration on a focus region ROI, enhancing and improving data diversity and model generalization by utilizing CutMix to obtain enhanced data with labels, taking a historical white light endoscopic image, the enhanced data with labels and category labels as training data, and performing enhanced training on a basic model by virtue of a mixed cutting enhancement strategy, so as to obtain an enhanced white light endoscopic image, an enhanced white light endoscopic image, an enhanced white light endoscopic image, an enhanced white light endoscopic image, an enhanced white light endoscopic image, an enhanced white light endoscopic image and an enhanced white light endoscopic image; after a white light endoscope image is input into the colorectal early cancer infiltration depth prediction model obtained through training, the shallow infiltration / deep infiltration type is predicted, and objective judgment of colorectal early cancer infiltration depth under a conventional white light endoscope is achieved.

The invention belongs to the technical field of robot control, and discloses a multi-dimensional touch decoupling method and system based on displacement field streamline and phase consistency, and the method comprises the steps: obtaining displacement or strain data of a flexible surface in continuous time, constructing a time sequence, and generating a displacement vector field and a displacement field streamline; establishing a joint motion topological stability reference through streamlines in continuous time, and comparing the current streamlines to detect topological anomaly so as to determine a touch candidate area; a displacement signal instantaneous phase is extracted in the candidate area, a phase gradient direction field is constructed, and finally the shear touch or torsion touch type is judged in combination with the cooperative relation between the streamline direction and the phase gradient direction. According to the method, stable decoupling of touch and movement can be realized under the background of strong joint movement, and the distinguishing accuracy of high-similarity touch types is improved.

The application discloses a kind of aspect-level sentiment analysis method based on multilevel knowledge enhancement, belong to sentiment analysis field. Including the following steps: S1: text word vector is obtained using GloVe word embedding tool, the syntax dependency tree of sentence is constructed using Stanza, and the dependency graph is constructed accordingly;S2: the context representation of sentence is extracted by inputting word vector into BiLSTM, and the dependency graph is updated using sentiment dictionary and sensitive relationship set, to realize the sentiment and syntax enhancement of sentence;S3: the enhanced dependency graph is input into GCN to model node features, to obtain specific aspect representation;S4: aspect word is enhanced using concept atlas to obtain aspect word representation, which is fused with specific aspect representation to obtain aspect representation;S5: aspect representation and context representation are coordinated and optimized using interactive attention, to obtain the final representation of sentence, so as to determine aspect sentiment tendency.The application can effectively improve the accuracy of specific aspect sentiment classification, and help merchants accurately locate problems in products or services.

The application discloses an allergen discrimination system based on an allergen family characteristic peptide group, wherein the system obtains an allergen characteristic peptide group through an allergen characteristic peptide group determination module; through a distribution characteristic acquisition module, according to the allergen and the allergen characteristic peptide group, the enrichment of the allergen characteristic peptide group, the enrichment of the allergen, the functional characteristic and the distribution characteristic are determined; through an allergen recognition characteristic set construction module, according to the enrichment of the allergen characteristic peptide group, the stable allergen characteristic peptide group, the multi-targeting characteristic and the allergen recognition characteristic set are determined; through an unknown allergen characteristic vector generation module, according to the unknown allergen sequence and the stable allergen characteristic peptide group, the unknown allergen characteristic vector is obtained; through a discrimination result determination module, the unknown allergen characteristic vector is input into a preset allergen discrimination model, and then the discrimination result of the unknown allergen sequence is determined. Through the implementation of the application, the discrimination accuracy of unknown allergens is improved.

The invention discloses an industrial pump interpretable intelligent fault detection method based on anti-fact learning, and relates to the technical field of industrial pump fault diagnosis, and the method comprises the steps: constructing a fault detection model; constructing an anti-fact search space for limiting a physical feasible range for generating an anti-fact sample; in the anti-fact search space, an anti-fact sample is generated through an optimization process, the optimization process minimizes the disturbance amplitude between the original working condition and the anti-fact sample, and meanwhile, the output judgment result of the fault detection model on the anti-fact sample is forced to be opposite to the output judgment result of the original sample; and outputting a structured interpretable result, the structured interpretable result including physical variable difference analysis of the anti-fact sample and the original sample, sensitivity ranking of the key physical variables and abnormal source inference, to provide engineering executable maintenance suggestions. According to the method, high-reliability and engineering executable pump fault diagnosis and explanation capability is realized.

The invention provides a landslide area extraction method and device, electronic equipment and a storage medium, and belongs to the technical field of image data processing. The method comprises the following steps: respectively inputting a first remote sensing image of a time phase before occurrence of a landslide, a second remote sensing image of a time phase after occurrence of the landslide and a slope map into a first feature extraction network, a second feature extraction network and a third feature extraction network of a landslide extraction depth model, a first feature map, a second feature map and a third feature map which are correspondingly output are obtained; inputting the first feature map, the second feature map and the third feature map into a feature fusion module of a landslide extraction depth model to obtain a fused feature map output by the feature fusion module; and inputting the fused feature map into a decoder of the landslide extraction depth model to obtain a landslide extraction result output by the decoder. According to the method, the accuracy, robustness and efficiency of landslide automatic detection and mapping are improved, the slope map is introduced as a supplement of topographic information, and the identification capability of the model for the potential landslide area is enhanced.

This invention discloses a method and system for locating and analyzing atmospheric pollution sources based on monitoring stations, belonging to the field of environmental monitoring and pollution source apportionment. The method involves acquiring pollutant concentration sequences and spatial coordinates of monitoring stations; extracting the main disturbance characteristic parameters of each station to construct a disturbance time-series vector and a propagation delay matrix; generating a candidate vector set of pollution propagation paths based on the relationship between delay and space, and performing path consistency assessment to select the path with the highest coherence; combining meteorological data to infer the initial location of the pollution source; establishing a local anti-diffusion model based on the estimation results; reconstructing concentration data from neighboring stations; and verifying the effectiveness of source location through error feedback. This invention can achieve high-precision identification of pollution source locations in complex pollution scenarios, possessing technical advantages such as strong timeliness, high adaptability, and verifiable location results.

The present application relates to the technical field of spectral detection, and discloses a crop canopy light field spectrum data correction method and a nutrient detection device. The method comprises the following steps: collecting crop canopy hyperspectral data, removing the background by using an optimized soil to adjust the vegetation index after radiation correction; constructing a light field heterogeneity parameter model combined with a red edge response factor, determining a red edge modulation coefficient through iterative optimization of a clustering evaluation index; calculating the Euclidean distance from a pixel to the clustering center of strong and weak light to generate a fusion weight coefficient; constructing a differentiated benchmark correction model, and performing linear weighted fusion twice on the spectral data and the nutrient content calculated by the double regression model by using the fusion weight. Through the construction of the light field heterogeneity representation parameter and the soft segmentation strategy based on the distance weight, the smooth transition of the strong and weak light regions is realized, the spectral distortion caused by uneven light is eliminated, and the accuracy and robustness of the crop nutrient detection under complex light field are improved.

The invention discloses a support vector machine circuit breaker state discrimination method and system based on improved genetic algorithm optimization and a storage medium, and the method comprises the steps: S1, obtaining a coil current signal in a circuit breaker, and extracting related feature data; s2, collecting a vibration signal time sequence generated by mechanical interaction in the circuit breaker; s3, designing selection, crossover and variation links of an improved genetic algorithm through a tournament selection mechanism, hybrid crossover and Gaussian variation, and taking a support vector machine discrimination result as a fitness function; s4, selecting a radial basis kernel function, and searching an optimal parameter of a penalty coefficient and kernel function parameter combination based on an improved genetic algorithm to construct an optimized support vector machine model; s5, training the optimized support vector machine model by using the training set data set; and S6, obtaining a current signal and a vibration signal of a circuit breaker coil, extracting features, and inputting the features into the trained and optimized support vector machine model.

The invention discloses a method and a system for identifying a developmental fault of a power transmission line based on Clark transform, and the method comprises the steps: determining an initial fault type based on the amplitude break variable in fundamental frequency signals of a normal operation state and an initial fault state of the power transmission line, which are extracted through Fourier transform; determining a special phase based on the initial fault type, determining a phase data input sequence of a three-phase current signal in the Clark transformation based on the special phase, and performing the Clark transformation to obtain a normal operation state, an initial fault state and a line mode component and a zero mode component after the initial fault; and judging a developmental fault type which is about to be evolved from the initial fault type by comparing amplitude changes of the line mode component and the zero mode component respectively. By comparing the amplitude change of the line mode component and the zero mode component during normal operation, the initial fault and the amplitude change after the initial fault, which fault the initial fault is about to develop into can be quickly and accurately judged, so that the accuracy of relay protection and fault phase selection is improved.

The application provides a power distribution network fault type judgment method and system, comprising: obtaining a fault voltage signal, a standard comprehensive feature matrix and a standard voltage signal characteristic value; based on the fault voltage signal and the standard voltage signal characteristic value, using Hilbert transform, discrete Fourier transform and Hilbert-Huang transform to generate a comprehensive feature matrix corresponding to the fault voltage signal; comparing the comprehensive feature matrix with the standard comprehensive feature matrix to obtain a power grid fault type, the application uses three algorithms of Hilbert transform, discrete Fourier transform and Hilbert-Huang transform to determine the fault type, improves the accuracy of short-circuit fault discrimination, especially improves the discrimination accuracy of long-distance power transmission faults, nonlinear load cut-in interference, noise interference faults and high-impedance grounding faults which are difficult to identify in the power distribution network, thereby ensuring the safety and reliability of the power distribution network.

The invention provides a research and development investment project discrimination method and system based on a large language model, and relates to the technical field of artificial intelligence, and the system comprises a data collection module which obtains data through the calling of an API interface, and the data comprises project document data, scientific research report data, market data and historical project data. And the data preprocessing module is used for preprocessing the data, and the preprocessing comprises formatting processing, denoising processing and standardization processing. According to the research and development investment project discrimination method and system based on the large language model, external data source information can be acquired in real time and discrimination standards and rules can be adjusted by introducing an adaptive updating module, the discrimination standards can be flexibly adjusted in the face of rapidly changing market demands and technical progress, and the discrimination efficiency is improved. And the discrimination accuracy of emerging fields or interdisciplinary projects is improved.

The invention discloses a simple moraine lake outburst susceptibility quantitative discrimination method, system and device and a storage medium, and belongs to the technical field of hydraulic engineering, and the method comprises the steps: determining the position of a moraine lake and the physical properties of moraine lake mother glacier; acquiring a current moraine lake area, a moraine lake area expansion rate within seven years, a mother glacier slope degree, a moraine dam backwater gradient, a mother glacier average ice tongue gradient and a distance between the moraine lake and the tail end of the mother glacier ice tongue; judging the area of the moraine lake, the expansion rate of the area of the moraine lake within seven years and the distance between the moraine lake and the tail end of the tongue of the mother glacier; providing moraine lake outburst probability P; and dividing the moraine lake outburst susceptibility through the moraine lake outburst probability P. According to the method, quantifiable terrain parameters are obtained, subjective judgment is not needed, the needed cost is extremely low, time-consuming and labor-consuming field investigation does not need to be carried out, the judgment accuracy of moraine lake outburst susceptibility is greatly improved, and the disaster prevention and reduction requirements of moraine lake outburst can be met.

This invention belongs to the technical field of bearing fault diagnosis. To address the inaccuracy of existing bearing fault diagnosis methods, this invention proposes a fault diagnosis method and system for press bearings based on conditional generative adversarial networks (GANs). A generator is trained using operating condition vectors and Gaussian noise. A discriminator is trained using the health state features generated by the generator and the actual health state features. The generator and discriminator are then subjected to adversarial training to prevent the discriminator from distinguishing between the health state features generated by the generator and the actual health state features, as well as from distinguishing the matching between the health state features generated by the generator and the corresponding operating condition vectors. This results in a well-trained health state model. The health state model is then used to generate a dynamic health baseline for the press shaft to be diagnosed, thereby obtaining the fault diagnosis result for the press shaft and achieving high-precision fault diagnosis and early warning.

The invention discloses a tectonic altered granite strength prediction method based on multiple linear regression, and belongs to the technical field of altered rock engineering geology. The method comprises the following steps: step 1, determining altered granite with different alteration degrees from a macroscopic view, and sampling; step 2, performing slice identification, SEM microstructure feature analysis and XRD mineral composition analysis on the rock sample, and performing alteration type classification in combination with macroscopic features; step 3, carrying out a dry density test, a porosity test, a water absorption test and a longitudinal wave velocity test on the collected rock sample; and 4, constructing a multivariate regression-based altered granite uniaxial compressive strength prediction model. According to the method, on the basis of research on the engineering geological characteristics of the altered granite, the multi-parameter indexes are selected and the multi-parameter variable indexes are input, so that the rock mass strength of the altered granite can be comprehensively predicted, and the prediction accuracy can be improved.

The application discloses a cabinet article access recognition method based on machine vision and relates to the technical field of intelligent cabinet article recognition, and comprises the following steps: acquiring a cabinet video and a cabinet door state, completing cabinet compartment registration and intercepting a current access event image sequence; extracting a foreground response and determining texture consistency, generating a short-time occlusion responsibility graph sequence; screening an occlusion recession frame to obtain a cabinet state image before and after access; combining a cabinet compartment calibration graph and a historical state snapshot to generate a compartment responsibility embedding feature; generating a directed change feature through a compartment occlusion responsibility enhanced SNUNet-CD network; according to the directed change feature and the historical snapshot, attributing an access type and closing a state, and outputting a recognition result and a current snapshot. The application utilizes machine vision and an occlusion responsibility enhanced change detection network, realizes cabinet article access recognition and state closure, and has the advantages of less occlusion misjudgment, accurate compartment attribution and stable state updating.

This invention provides a geostress testing device and method based on sleeve-induced fracturing, belonging to the field of rock mass engineering monitoring technology. The device includes a sleeve fixing frame, a piezoelectric ceramic acoustic transducer, a shielding wire, sleeve rubber, an annular plug, a hose, a sleeve support, a mounting rod, an angle disc, a high-pressure hand pump, an oscilloscope, and an acoustic drive board. Waveform signals are emitted through the acoustic drive board and one piezoelectric ceramic acoustic transducer, while the other transducers receive the waveform signals. Pressure is slowly increased by the high-pressure hand pump, and the pressure gauge value and the waveform information displayed on the oscilloscope are recorded in real time. By analyzing the changes in the amplitude parameters of the transducer waveform information and the pressure gauge value, the initial fracturing pressure and the fracturing re-tension pressure of the rock mass are determined. After fracturing, the changes in the amplitude parameters of the waveform information are measured again by rotating the device by a certain angle each time. The location of the fracturing can be determined based on the received waveform signals at different angles, and the presence or absence of rock mass fracturing can be determined by the changes in the waveform signal amplitude.

The present application relates to the technical field of audio and video player equipment, and specifically to an intelligent control method and system for audio and video players. The method includes: collecting motion speed data and temperature data; obtaining a preset time before the current time as a statistical time, obtaining a first possibility that the user wears headphones according to the characteristics of the motion speed data at the statistical time; obtaining a second possibility that the user wears headphones according to the characteristics of the temperature data at the statistical time; obtaining the probability that the user wears headphones at the current moment according to the first possibility and the second possibility that the user wears headphones; and intelligently controlling the audio and video player according to the probability that the user wears headphones at the current moment. The present application provides more precise control of the audio and video player.

The invention provides a multi-instance learning-based grape tire risk hierarchical prediction method. The method comprises the steps of obtaining suspected grape tire pathological image data and clinical parameters; preprocessing the image and cutting the image into image blocks; extracting image block features by using a feature extraction network, and aggregating to generate packet-level features; high-discrimination features are obtained through a multi-instance feature aggregation module; coding the coding network into features, and generating multi-source fusion features with fusion; inputting into the classification network, outputting prediction probabilities of non-grape tyres, partial grape tyres and complete grape tyres, and calculating disease risk scores according to the prediction probabilities of the non-grape tyres, the partial grape tyres and the complete grape tyres; dividing into four risk levels of low, medium, medium and high and high according to the scores, and automatically triggering an auxiliary diagnosis prompt for cases falling into classified easy-to-confuse levels. Pixel-level labeling is not needed, clinical safety and interpretability are considered while classification accuracy is improved, and the method is suitable for grape tire intelligent auxiliary diagnosis of primary and superior medical institutions.

The present application relates to the technical field of irrigation prediction, in particular to a kind of irrigation water consumption prediction method based on soil moisture condition.The method is by synchronously obtaining photosynthetic radiation intensity and matrix water content in sliding time window, calculates water consumption rate;Further, the mean ratio of radiation and consumption rate is used to determine the load index representing water extraction resistance, and the ratio of the variation coefficient of the two is used to construct the fluctuation transmission index reflecting the matrix hydraulic connectivity;Combined with the two indexes, the hydraulic block risk coefficient is generated to identify the hidden drought caused by capillary fracture;Finally, according to the risk coefficient and the preset maximum irrigation amount, the actual water demand is predicted.The method can effectively distinguish crop physiological regulation and matrix physical water shortage, significantly improve the discrimination accuracy of matrix micro-hydraulic connectivity state, and then make the irrigation water consumption prediction more targeted and timely, avoid misirrigation or missed irrigation.