17results about How to "The solution accuracy is not high" patented technology

The invention discloses a method and a system for establishing a low-frequency model based on multi-attribute fusion. The method for establishing the low-frequency model based on the multi-attribute fusion comprises the following steps: establishing a work area longitudinal wave velocity compaction trend body; based on the longitudinal wave velocity compaction trend body and the seismic horizon velocity body, constructing a pre-stack parameter elastic very low frequency model; establishing a well interpolation model based on the three-dimensional constructive framework; and performing frequency domain merging on the pre-stack parameter elastic very-low-frequency model and the well interpolation model to obtain a multi-attribute fusion low-frequency model. According to the method, seismic velocity information, regional compaction trend information, logging information and geologic model information are fully fused to participate in establishment of the low-frequency model, the'bullseye 'phenomenon of a conventional well interpolation low-frequency model is improved, the overall rule is matched with the structure background, and the problems that well drilling distribution is not uniform, reservoir heterogeneity is high, and the reliability is high are solved. And the low-frequency model precision is not high due to fast transverse change.

This invention discloses a method for locating epileptogenic zones based on high-frequency oscillations and connectivity, comprising the following steps: 1. Acquiring SEEG data and selecting representative channels; 2. Filtering to acquire 54-200Hz signals and selecting baseline and target data; 3. Acquiring normalized high-frequency energy; 4. Calculating time and energy coefficients to obtain a high-frequency epileptogenicity index; 5. Filtering to acquire 12-45Hz signals; 6. Calculating nonlinear regression analysis; 7. Calculating total intensity; 8. Defining and calculating the connectivity high-frequency epileptogenicity index and evaluating performance. This invention can accurately locate epileptogenic zones in patients with different seizure patterns, and has potential application value in epilepsy treatment.

The embodiment of the invention provides a low-complexity signal adaptive detection method and system based on an FPGA, and the method comprises the steps: obtaining a signal envelope which comprises a plurality of data samples; performing amplitude multi-point smoothing processing on the signal envelope to obtain a to-be-judged signal comprising a plurality of judgment samples; obtaining an adaptive threshold value based on the data sample; sequentially comparing the judgment samples with the self-adaptive threshold value; when M continuous judgment samples in the to-be-judged signal are greater than the adaptive threshold values corresponding to the judgment samples, outputting a high-level detection mark; and when N continuous judgment samples in the to-be-judged signal are smaller than the adaptive threshold values corresponding to the judgment samples, outputting a low-level detection mark. Amplitude multi-point smoothing processing is firstly carried out on a signal envelope, then rising edge pulse judgment and falling edge pulse judgment are respectively carried out by adopting multi-data-point accumulation (M-N method), and a relatively high correct judgment probability can also be achieved when the signal-to-noise ratio is relatively low.

This invention discloses a calibration method, apparatus, equipment, and medium for a fracture model of metallic materials. The method includes: determining initial fracture parameters and initial instability parameters of stress triaxiality to match calibration experiments; wherein the calibration experiments include at least one of uniaxial tensile tests, notched tensile tests, tension-shear tests, shear tests, compression tests, and biaxial tensile tests; each calibration experiment is matched with at least one stress triaxiality; updating the initial fracture parameters and initial instability parameters until both fracture parameters and instability parameters meet preset strain conditions; and determining the calibration result of the fracture model based on the fracture parameters and instability parameters of the stress triaxiality. This technical solution solves the problem of low accuracy in the calibration of fracture models of metallic materials, effectively improving calibration efficiency while increasing accuracy, and decoupling multi-condition simulations.

The invention relates to the technical field of power distribution network intelligent control, in particular to a power distribution network fault self-healing distributed resource collaborative reconstruction method, system and device and a storage medium. The method comprises the following steps: collecting multi-source fault information of a power distribution network, constructing a fault information set, determining a fault section based on the fault information set through confidence-weighted information consistency verification, and generating a hierarchical isolation strategy to minimize a power failure range of a non-fault area; generating a fault isolation operation sequence according to the position of the fault section, and evaluating the photovoltaic output capacity, the energy storage charge and discharge capacity and the connection line transmission capacity; establishing a coupling relationship between network topology adjustment and distributed resource output adjustment, and determining a switching operation sequence and a resource scheduling scheme by solving a multi-constraint optimization problem; and calculating a dynamic importance score according to the load type, the current power utilization state and the social influence factor, sorting the loads according to the dynamic importance score, sequentially accessing the loads, and verifying the voltage and the line load rate.

The present application relates to the technical field of unmanned driving, and discloses a vehicle control method and device, computer equipment and a storage medium, comprising: acquiring the pose of a target pallet to determine the target parking pose and reference line of a target vehicle; determining an angle parameter based on the reference line and a preset angle reference line, and determining a distance parameter of the target vehicle to the reference line based on the angle parameter and the current pose of the target vehicle; determining a path point on the reference line based on the distance parameter, the angle parameter and the target parking pose; controlling the target vehicle to move to the path point until the target vehicle reaches the target parking pose. The present application does not need to perform path planning in advance, reduces the time required for planning based on the target parking pose, improves overall efficiency, controls the vehicle based on the target parking pose, and realizes high-precision autonomous positioning of the vehicle depot distance driving.

This application discloses a method, apparatus, and computer program product for correcting genome assembly. The method includes: correcting the initial assembly result of a sample based on long-read sequencing data to obtain a first corrected result; correcting the first corrected result based on short-read sequencing data to obtain a second corrected result and a correction record; and filtering out erroneous corrections from the second corrected result based on the short-read sequencing data and the correction record to obtain the target assembly result of the sample. This application solves the problem in related technologies where the accuracy of the assembly result is low after first correcting the initial genome assembly result with long-read data and then with short-read data.

This invention discloses a method, apparatus, device, and medium for segmenting blood vessel images. The method includes: preprocessing and segmenting a blood vessel image to be segmented to obtain multiple blood vessel image blocks of a preset size; inputting each blood vessel image block to be segmented into a target image segmentation model to obtain corresponding blood vessel segmentation image blocks; and stitching the images based on the relative positional relationship of the blood vessel segmentation image blocks to obtain a target blood vessel segmentation image. The target image segmentation model includes an encoder, a first decoder, and a second decoder. The second decoder is a variational self-decoder trained using a blood vessel enhancement segmentation reference image corresponding to the model training sample image as sample labels. This invention solves the problem of low accuracy in blood vessel image segmentation using basic blood vessel segmentation models, improves the generalization ability of neural network models used for blood vessel image segmentation, and enhances the accuracy of blood vessel image segmentation.

The application relates to a test waveform generation method and a digital board card. The method is applied to a digital board card of an integrated circuit test device, and the digital board card comprises a first encoding module and a second encoding module. The method comprises the following steps: the first encoding module encodes the waveform format of an original test stream based on waveform format information of the original test stream, and first encoding data of each period is obtained; the second encoding module updates the first encoding data of each channel of each period based on a mask enable signal corresponding to each channel, and second encoding data is obtained; the second encoding data is respectively subjected to differential encoding and cumulative flip integral processing, and third encoding data is obtained; and the third encoding data of each channel is subjected to delay control and serial XOR processing, and test waveforms of the entire test process are obtained. The method can solve the problem that the existing test waveform generation mode has low precision due to dependence on manual operation.

The application provides a tooth width calculation method of a mouth scanning dental mold, a storage medium and an electronic device, and the method comprises the following steps: acquiring a mouth scanning dental mold image, segmenting the mouth scanning dental mold image according to single teeth to obtain a plurality of tooth images; acquiring a certain tooth image, calculating the centroid coordinates corresponding to the tooth image, and calculating the reference point coordinates corresponding to the tooth image according to the centroid coordinates; taking the connecting line of the centroid coordinates and the reference point coordinates as a reference axis, determining two pixel points with the largest distance to the reference axis on the current tooth image, and taking the width between the two pixel points as the tooth width of the current tooth image; repeating the above steps until the tooth widths of all tooth images are calculated, and outputting the sum of the tooth widths of all tooth images. Since the reference axis of each tooth image is adaptively adjusted according to the centroid and reference point coordinates of each tooth image, the calculation of the tooth width is more accurate.

The application belongs to the technical field of emotion recognition, and in particular to an emotion recognition method based on left and right brain emotion lateralization; the method specifically comprises the following steps: S1, electroencephalogram signal acquisition electrode and processing: obtaining standard electroencephalogram signals; S2, time domain feature and frequency domain feature extraction of electroencephalogram signals of left and right brains: extracting Hjorth and differential entropy features; S3, calculating edge weights of a dynamic graph convolutional neural network based on feature differences of left and right brains: calculating weights by using a CRITIC weight method; S4, calculating an adjacency matrix of the dynamic graph convolutional neural network based on inter-channel correlation of electroencephalogram signals: calculating phase locking values to construct the adjacency matrix; S5, emotion recognition by using an improved dynamic graph convolutional neural network: iteration and training; the application extracts the same features from the time domain and the frequency domain of the left and right brains respectively, thereby better recognizing that the left brain is more sensitive to positive emotions and the right brain is more sensitive to negative emotions, and solving the problem of low accuracy of emotion recognition in the two dimensions AV of valence and arousal.

The present application provides a human physiological parameter monitoring component and a wearable device, the component includes a mainboard and a light source module and a light detection module arranged on the mainboard, wherein the light source module includes a first light emitting component for emitting a first light ray and a second light emitting component for emitting a second light ray, which are arranged at intervals; the light detection module includes a first light detector for receiving a reflected light ray of the first light ray and a second light detector for receiving a reflected light ray of the second light ray, and the first light detector and the second light detector are respectively located on both sides of the light source module; the distance between the second light emitting component and the second light detector is greater than the distance between the first light emitting component and the second light detector, and the distance between the second light emitting component and the second light detector is greater than the distance between the first light emitting component and the first light detector. Through the above method, the accuracy of human physiological parameter monitoring can be improved, and it is conducive to the miniaturization of the product.

The invention relates to the technical field of computers, and discloses a sentiment analysis method and device, a medium and a program product, and the method comprises the steps: dividing a data set containing a plurality of sentiment categories into different subtasks according to a set regression category; traversing each sub-task, randomly selecting a set number of samples in the current sub-task when the current sub-task is adopted to train the spiking neural network, and calculating the weight importance corresponding to each sample; forming a fuzzy set containing a membership degree according to weight importance, and selectively performing gradient shielding on part of neurons in the full connection layer and part of convolution kernels in the convolution structure by using the fuzzy set; and after all the subtasks are traversed, obtaining an observation model for sentiment analysis, and performing sentiment analysis on the to-be-tested data by using the observation model. The gradient shielding mode based on the fuzzy set theory can improve the continuous learning capability in the sentiment analysis task, reduce the data transmission quantity, improve the hardware processing speed and further improve the sentiment prediction accuracy.

The application discloses a multi-intent determination method and device, a storage medium and an electronic device, relates to the technical field of smart homes, and comprises the following steps: performing data preprocessing on interaction information input by a target object to obtain preprocessed target interaction information; performing intent feature extraction on the target interaction information through a deep learning model to obtain a plurality of intent features corresponding to the target interaction information; determining the confidence of an intent recognition result corresponding to each intent feature through a target classifier, and determining a multi-intent recognition result according to the confidence of the intent recognition result corresponding to each intent feature; and determining an intent set corresponding to the interaction information in the multi-intent recognition result according to a logical relationship in the interaction information. By using the technical solution, the problem that the accuracy of an intent recognition technology is not high when the intent recognition technology processes multi-intents in a complex context in the related art is solved.

The application provides a forest stand volume growth prediction method and micro-plot correction process based on Taylor series, a simple calculation method of forest stand volume growth is derived according to Taylor series expansion, a growth model of national scale is established by using national continuous inventory data and environment data obtained by remote sensing satellites, tree core data is obtained by arranging a small amount of micro-plot, the accurate measurement value of forest stand volume growth is obtained through the analysis of annual ring data, the prediction value of forest stand volume growth is obtained by applying the national scale model, then a correction model is researched and built, a correction process of the region is established, and the accurate prediction of the future forest stand volume growth of any plot in the region is realized.

The invention relates to the technical field of costume identification, and discloses a national costume identification method and system, and the method comprises the steps: obtaining portrait image data of national costumes worn in all regions; processing the portrait picture data to obtain portrait picture data with the same format, size and resolution; carrying out segmentation processing on the processed portrait image data through an FCN model trained by a human body analysis data set based on national costume feature labeling to obtain a portrait region feature map containing feature labels labeled on different body parts; inputting the portrait region feature map into a DenseNet convolutional neural network for training to obtain a trained clothing identification model; and inputting the ethnic costume picture into the trained costume identification model, and outputting an ethnic costume identification result. According to the invention, the feature attributes of the portrait region feature map sample are enhanced, the learning and training effect of the DenseNet convolutional neural network is improved, and higher ethnic costume identification accuracy is realized.

This invention discloses a method, apparatus, and medium for predicting lymph node metastasis. The method involves acquiring a three-dimensional image of a lymph node to be detected and determining the region of interest (ROI) within the image. Feature extraction is performed on the ROI to obtain local features corresponding to the three-dimensional lymph node image. The ROI data is then expanded to obtain pixel blocks to be processed, and feature extraction is performed on these pixel blocks to obtain global features corresponding to the three-dimensional lymph node image. Based on the local and global features, the lymph node metastasis prediction result for the three-dimensional lymph node image is determined. This invention achieves a comprehensive description of the features of the three-dimensional lymph node image through the complementarity of local and global features, solving the problem of low accuracy in lymph node metastasis prediction and improving its reliability. Furthermore, the local features only include the features of the lymph node itself, reducing redundant features.