38 results about "Threshold estimation" patented technology

The invention discloses a low signal-to-noise ratio conditional modulation identification method based on wavelet transform and channel attention. The method comprises the following steps: acquiring a received signal; and inputting the received signal into an identification model of a pre-trained neural network, and outputting a classification identification result and classification accuracy. In the recognition model based on signal reconstruction and the neural network, a wavelet threshold estimation module and a wavelet reconstruction module form a signal reconstruction path, a multi-scale feature extraction module and a prediction classification module form a recognition path, and the two paths are mutually enhanced, so that the combination of the advantages of a digital signal processing technology and the neural network is realized; a wavelet threshold estimation module is introduced to predict a denoising threshold through a neural network, and network parameters of the wavelet threshold estimation module are updated according to back propagation; a signal reconstruction path is combined with a recognition path, the interpretability of the network is enhanced, and the accuracy of modulation recognition in a low signal-to-noise ratio scene is remarkably improved by using multi-scale modulation characteristics.

The invention is suitable for the field of pain assessment, and particularly provides a pain threshold estimation method and system based on multi-modal data fusion, and the method comprises the steps: constructing a cross-modal weak feature enhancement fusion network, carrying out the processing of collected multi-modal data, and extracting physiological weak features and behavior weak features; introducing a time-varying attention Transform model, dynamically allocating weights for the processed physiological weak features, behavior weak features and environment auxiliary data by using a pain dynamic tag, and outputting cross-modal fusion features; constructing a double-branch pain threshold estimation model, wherein the model comprises a supervised estimation branch and an unsupervised estimation branch; and introducing a pain feature element library, carrying out online calibration on parameters of the double-branch model in combination with a meta learning algorithm, and outputting a real-time pain threshold value and a short-term pain threshold value change trend of the target object. According to the embodiment of the invention, the clinical operability of pain assessment is remarkably improved, and ineffective intervention caused by inaccurate assessment is reduced.

The invention discloses an arc SAR imaging method and system, and the method comprises the steps: S1, building a radar signal model and a motion geometric model, and deducing the distance history of a target point at different moments; s2, converting the range history into a Doppler frequency domain, analyzing a corresponding relation between range migration and Doppler frequency, designing a phase compensation coefficient, and correcting the range migration of the Doppler frequency domain in a time domain to obtain corrected range data; s3, analyzing the spectral characteristics of the azimuth response function, and estimating an effective support domain based on an energy approximation threshold; s4, performing azimuth accumulation on the spectrum data in the support domain by using Chirp-Z transformation to obtain an azimuth accumulation result; and S5, outputting an SAR imaging result in combination with the corrected range direction data and the azimuth direction accumulation result. The method has the advantages that the calculation complexity and the memory required by calculation can be reduced, and the like.

An apparatus for data processing and related non-transitory computer-readable medium are provided. In the method, the apparatus computes a burst threshold for a data stream. The burst threshold is associated with a throughput of the data stream. The apparatus further identifies a set of bursts in the data stream based on the burst threshold, and detect a pattern in the data stream based on the set of bursts. The apparatus further estimates at least one subsequent burst in the data stream based on the pattern and the burst threshold, and outputs an indication of the at least one subsequent burst in the data stream. The method enables a device to predict one or more future data bursts based on characteristics of existing data. The predicted data bursts allow the transmission resource and power to be adaptively arranged to significantly reduce power consumption and improve transmission efficiency.

An electronic circuit includes a fast-settling threshold estimator for determining a fast-settling threshold voltage which includes a maximum peak detector that receives an analog signal containing digital data and produces an estimate of a maximum value of the analog signal; a differentiator that produces an output voltage responsive to a change in the estimate of a maximum value of the analog signal; a preselected voltage coupled to a comparator that outputs a pulse when the output voltage is greater than the preselected voltage; a minimum peak detector that produces an estimate of a minimum value of the analog signal; and an averaging circuit that produces a weighted average of the estimate of a maximum value of the analog signal and the estimate of a minimum value of the analog signal. In one embodiment, the fast-settling threshold estimator is coupled to a bit slicer and to a RF receiver.

The present disclosure relates to a non-parametric optimal read threshold estimation scheme for a memory system. The memory system includes a memory device having pages and a controller having a neural network. The controller performs read operations on selected pages using a read threshold set; obtains a read threshold set, a checksum value, and an asymmetric ratio of 1 counts to 0 counts associated with decoding of the selected page based on each read operation; provides the obtained read threshold set, checksum value, and asymmetric ratio as input information to the neural network; and estimates an optimal read threshold voltage by the neural network based on the input information and weights, wherein the weights include a combination of multiple matrices and a bias vector.

The invention relates to a three-component seismic data random noise suppression method and device based on quaternion wavelet transformation, a medium and equipment. The method comprises the following steps: representing seismic data of X, Y and Z components as quaternion signals in a quaternion array form in a time-space domain; performing quaternion wavelet transform on the quaternion array, and performing multilayer decomposition on the quaternion signal in the form of the quaternion array by using the quaternion wavelet transform to obtain a quaternion wavelet coefficient corresponding to each layer; calculating a threshold value by using a new threshold value estimation expression, performing threshold value operation on quaternion wavelet coefficients of each layer by adopting a semi-soft threshold value function, reserving the wavelet coefficients greater than the threshold value, and setting the wavelet coefficients smaller than the threshold value to be zero; and performing quaternion wavelet inverse transformation on the quaternion wavelet coefficient after threshold processing, reconstructing a denoised quaternion signal, and extracting an X component, a Y component and a Z component from three imaginary part positions of the denoised quaternion signal to complete vector denoising.

The present application relates to a kind of coated optical element laser damage dominant mode determination and threshold estimation method, belong to laser optical element damage evaluation technical field.The implementation method of the present application is as follows: based on the calculation of film optical field distribution of transmission matrix method or finite difference time domain method;Temperature and stress field of film and substrate are predicted by combining finite element analysis;Compare damage threshold to determine dominant mode;And verify the accuracy of model through experiment.The present application quantifies the damage contribution of film electric field-thermal coupling model and substrate thermal analysis model by constructing them respectively, and verifies the model through small sample experiment, aims to solve the problems of complex damage competition mechanism of coated optical element film and substrate, difficult to distinguish dominant damage mode, and low efficiency and high cost of traditional experimental method.

The invention relates to threshold estimation and calibration for anomaly detection. This document is machine learning (ML) and extreme value theory (EVT) techniques for normalizing and thresholding anomaly scores without assuming value distributions. In an embodiment, a computer receives a number of non-normalized anomaly scores and selects a highest subset of the non-normalized anomaly scores that exceeds a tail threshold according to a peak over-threshold (POT). Parameters of a probability density function are trained according to EVT based on the highest subset of non-normalized abnormal scores. After training and in a production environment, a normalized anomaly score is generated based on the non-normalized anomaly score and trained parameters of the probability density function. Anomaly detection compares the normalized anomaly score to an optimized anomaly threshold.

The invention discloses an intelligent self-service pure tone threshold test method and device and a medium, and relates to the technical field of threshold test, and the method comprises the steps: executing a pure tone threshold test process based on an information gain selection mechanism according to the recognition difficulty level in noise, the matching degree score and a pure tone threshold test starting strategy, the key pressing behavior of the testee is judged in real time in combination with the personal operation rhythm template, the hearing threshold estimation result of each frequency band and the corresponding uncertainty index and behavior event mark are output, and fixed-point deviation correction complementary measurement and weight correction are carried out based on the hearing threshold estimation result of each frequency band, the uncertainty index and the behavior event mark; obtaining a final hearing threshold result and an overall credibility grade; by establishing the candidate test set and introducing the pure tone threshold measurement process based on the information gain selection mechanism, the dynamic adjustment of the test path according to the hearing threshold estimation result and the uncertainty index of each frequency band is realized, and the test efficiency and the result stability are improved.

The application discloses a wavelet denoising method and system for SAR images and electronic equipment, and relates to the technical field of remote sensing image processing. The method comprises the following steps: determining a wavelet and a wavelet decomposition level, and calculating a wavelet coefficient matrix obtained by wavelet decomposition of a to-be-processed SAR image; determining an initial adaptive threshold and a corresponding estimated wavelet coefficient matrix according to the wavelet coefficient matrix based on a preset adaptive threshold estimation function; removing a wavelet coefficient when any wavelet coefficient is less than the initial adaptive threshold; when any wavelet coefficient is greater than or equal to the initial adaptive threshold, constructing a risk function based on the wavelet coefficient and the estimated wavelet coefficient matrix; performing minimum calculation on the risk function based on a minimum maximum estimation criterion to obtain optimal wavelet coefficients after threshold processing; and performing wavelet reconstruction on the optimal wavelet coefficients to obtain a denoised SAR image. The application effectively removes noise in the image while better maintaining the image edge, thereby improving the denoising effect.

The application discloses a kind of semi-supervised multi-label image classification method, system, equipment and storage medium.The method utilizes feature extraction network to obtain the spatial feature representation of input image, and respectively constructs class-independent global feature and class-specific global attention feature;Feature map is divided into local area, local average feature and local class-specific attention feature are generated, the fusion of global feature and local feature is realized by residual form, attention enhancement prediction of teacher network is constructed based on fusion feature, and the optimal threshold value of each class is searched on labeled data using class adaptive threshold estimation method, to generate high-quality pseudo label on unlabeled data;Semi-supervised collaborative optimization is realized by combining network training again.The application can effectively improve the accuracy and stability of multi-label image classification under limited annotation conditions, and is especially suitable for medical image recognition, retail commodity detection, remote sensing scene analysis and other multi-label complex scenes.

The invention discloses a model training method and device based on a multi-channel detection system, equipment and a medium, and the method comprises the steps: obtaining environment parameters of the multi-channel detection system; simulating a radiation source signal and a receiving signal based on the environmental parameters, and determining a detection threshold as a first label based on the receiving signal and a preset false alarm probability; based on the detection threshold, detecting the generated receiving signal through a Monte Carlo simulation experiment, and counting the detection probability that the radiation source signal exists in the receiving signal as a second label; constructing a first training data set based on the environmental parameters and the first labels, and constructing a second training data set based on the environmental parameters and the second labels; and iteratively updating the threshold estimation model based on the first training data set, and iteratively updating the detection probability prediction model based on the second training data set. According to the embodiment of the invention, the effectiveness and robustness of the method in the aspects of predicting the detection performance and determining the threshold value are realized.

The present invention discloses a circular arc SAR imaging method and system, comprising the following steps: S1. establishing a radar signal model and a motion geometry model to derive the range history of a target point at different times; S2. converting the range history to the Doppler frequency domain, analyzing the correspondence between range migration and Doppler frequency, designing a phase compensation coefficient, and correcting the range migration in the Doppler frequency domain in the time domain to obtain corrected range data; S3. analyzing the spectral characteristics of the azimuth response function and estimating the effective support domain based on an energy approximation threshold; S4. using a Chirp-Z transform to perform azimuth accumulation of the spectral data within the support domain to obtain an azimuth accumulation result; and S5. combining the corrected range data with the azimuth accumulation result to output a SAR imaging result. The present invention has the advantages of reducing computational complexity and memory requirements.

The invention relates to the field of material analysis, and relates to a DIC-based irregular crack self-adaption and multi-target synchronous width calculation method. The method comprises the following steps: step S101, data reading: obtaining a displacement component of each point of a DIC displacement field; reading images frame by frame to form a two-dimensional array; recording input metadata for each frame of image; s102, carrying out ROI (Region of Interest) self-adaption and masking; step S103, pixel-physical quantity outline calibration is carried out, and a displacement field is converted to millimeter in a unified manner; step S104, estimating a self-adaptive threshold value; step S105, carrying out discontinuous displacement detection; step S106, carrying out candidate segment aggregation and multi-target grouping; step S107, carrying out local trend regression fitting; and S108, geometric normal correction is carried out. According to the method, the crack boundary is identified by the displacement discontinuity of the DIC displacement field, the slope k of the center line is obtained by combining local trend regression, the row width is geometrically corrected into the normal real width, and the system error caused by orientation is eliminated.

Various embodiments provide for performing a read voltage level calibration process with selective skipping collection of calibration-related information on a memory device of a memory system, such as a memory sub-system. For example, some embodiments provide for performing a coarse threshold estimate (CTE) calibration process, with selective skipping of failed bit count (CFBit) information collection, on a memory system.

Various embodiments provide for performing a read voltage level calibration process with selective skipping collection of calibration-related information on a memory device of a memory system, such as a memory sub-system. For example, some embodiments provide for performing a coarse threshold estimate (CTE) calibration process, with selective skipping of failed bit count (CFBit) information collection, on a memory system.

The present application relates to the technical field of industrial equipment sign prediction, and discloses a data-driven industrial equipment sign prediction method, device and medium. The method comprises: real-time sampling of the operating status of industrial equipment to obtain multivariate time series data; encoding and decoding the multivariate time series data through a time series reconstruction model to obtain a reconstruction output; performing difference calculation on the multivariate time series data and the reconstruction output to obtain a reconstruction error set including multiple univariate feature reconstruction errors and multiple multivariate feature reconstruction errors; scoring the reconstruction error set to obtain anomaly scores, wherein the anomaly scores include: multiple multivariate feature anomaly scores and multiple univariate feature anomaly scores; threshold estimation of the anomaly scores to obtain anomaly thresholds, comparing and judging the anomaly scores and the anomaly thresholds to obtain an anomaly determination result of the current state. The present application improves the accuracy of data-driven industrial equipment sign prediction.

The invention discloses a multi-target detection method based on a millimeter-wave radar, and the method comprises the steps: collecting the echo data of multiple targets in a complex scene through the millimeter-wave radar, and carrying out the preprocessing of the echo data, and obtaining the distance-Doppler two-dimensional data to be processed; based on the obtained two-dimensional data to be processed, adopting a CFAR detection algorithm to obtain a coarse estimation CFAR mark detection result; the maximum module value square of each column of pulses on the two-dimensional data to be processed is stored, and a false target interference mark is obtained through dynamic adaptive threshold estimation; and based on the obtained CFAR mark and the false target interference mark, accurate detection of the target is realized through a comprehensive mark. According to the method, a CFAR-based comprehensive marking algorithm is adopted, target detection is realized by dynamically estimating background noise and a self-adaptive threshold value, multi-dimensional information such as distance, speed and time-frequency characteristics is fused, a real target and clutter are effectively distinguished, and the target can be accurately identified and related information of the target can be extracted in a complex scene.

A method for reading data from a solid-state drive (SSD) configured to store data in a plurality of memory cells arranged in memory blocks comprising rows, the method performed by a controller in communication with the plurality of memory cells. The method comprises retrieving data from a target row of memory cells of the plurality of memory cells associated with a read request received from a host using initial threshold voltages. The method also includes decoding the data using a hard decision stage. Additionally the method comprises estimating read threshold voltages of the target row of memory cells based on a transformation of a distribution of threshold voltages of cells in a memory block containing the target row when the hard decision decoding stage fails. The method further includes retrieving data from the target row using the estimated read threshold voltages.

The application discloses a low signal-to-noise ratio condition modulation identification method based on wavelet transform and channel attention, comprising: acquiring a received signal; inputting the received signal into a pre-trained neural network identification model to output a classification identification result and a classification accuracy rate. Based on the signal reconstruction and the identification model of the neural network, the wavelet threshold estimation module and the wavelet reconstruction module constitute a signal reconstruction path, the multi-scale feature extraction module and the prediction classification module constitute an identification path, the two paths are mutually enhanced, the advantages of the digital signal processing technology and the neural network are combined, the wavelet threshold estimation module is introduced to predict a denoising threshold through the neural network, and the network parameters of the wavelet threshold estimation module are updated according to the back propagation; the signal reconstruction path and the identification path are combined, the network interpretability is enhanced, the multi-scale modulation features are utilized, and the modulation identification accuracy rate in the low signal-to-noise ratio scene is significantly improved.

The invention belongs to the technical field of short-time burst signal capturing, and particularly relates to a synchronous capturing and collecting method for multi-node short-time burst signals. Each node performs data synchronous acquisition in a time triggering mode; the signal-to-noise ratio of a target signal is improved through weighted moving average and data self-correlation processing of data collected by the nodes, and signal capture and extraction are completed through a peak threshold estimation method. The method effectively solves the problems of low timeliness of synchronous acquisition of the short-time burst signals under multiple nodes, high requirements on equipment and the like, has the characteristics of accuracy, high efficiency, high speed and the like, and provides technical support for subsequent research on positioning and tracking of the type of signals and the like.

The invention relates to the technical field of radar signal processing, and provides an angle classification threshold estimation method and device, a medium and a program product, and the method comprises the steps: calculating the change of an included angle between the projection of a wave beam center point in an X-axis direction and a Z axis in a signal processing period according to a satellite position, a satellite flight speed and a wave beam scanned by a satellite; and the change of the included angle between the projection of the beam center point in the Y-axis direction and the Z-axis is used as the two-dimensional threshold of the angle classification of all targets in the beam. Compared with a fixed threshold, the self-adaptive angle sorting threshold can achieve higher sorting accuracy and has better engineering practical value.

The application relates to a three-component seismic data random noise suppression method, device, medium and equipment based on a quaternion wavelet transform, which comprises the following steps: representing X, Y and Z three-component seismic data as a quaternion signal in the form of a quaternion array in a time-space domain; implementing quaternion wavelet transform on the quaternion array; using the quaternion wavelet transform to perform multi-layer decomposition on the quaternion signal in the form of the quaternion array, so as to obtain corresponding quaternion wavelet coefficients of each layer; calculating a threshold value by using a new threshold value estimation expression, and performing threshold value operation on the quaternion wavelet coefficients of each layer by using a semi-soft threshold value function, wherein the wavelet coefficients greater than the threshold value are reserved, and the wavelet coefficients less than the threshold value are set to zero; performing quaternion wavelet inverse transform on the threshold value processed quaternion wavelet coefficients, reconstructing a denoised quaternion signal, taking out X, Y and Z components from three virtual parts of the denoised quaternion signal, and completing vector denoising.

The present application relates to the field of high-voltage engineering and insulation technology, and particularly relates to a method for artificial contamination test under high-voltage environment, which comprises the following steps: firstly, constructing a target deposition distribution function and a target response set on a parameterized grid on the surface of an insulator, and merging an environmental boundary condition time sequence set to form a target deposition and response data set; then, implementing layered deposition of soluble contaminants and insoluble particles under the electrode system time-varying electric field and particle charging control in the on-power state, collecting surface state field to generate an in-situ deposition state data set; then, obtaining electrical and surface responses and estimating a median flashover voltage according to segmented loading and stepwise flashover test to form an electrical and physical response data set; finally, based on the kinetic model, closed-loop optimization is performed on the electrode system time-varying trajectory, the environmental program and the deposition spectrum within the leakage current constraint. The present application realizes reproducible configuration, stable threshold estimation and fast test convergence.

Various embodiments provide for performing a read voltage level calibration process with selective skipping collection of calibration-related information on a memory device of a memory system, such as a memory sub-system. For example, some embodiments provide for performing a coarse threshold estimate (CTE) calibration process, with selective skipping of failed bit count (CFBit) information collection, on a memory system.

The invention discloses a multi-parameter satellite component health assessment and prediction method based on joint distribution evolution. The method comprises the following steps: S1, acquiring initial reference data of normal work of a satellite component and current operation data of a latest complete period; s2, carrying out joint distribution difference analysis on the current data and the reference data, and calculating the distribution difference by adopting a slice Wasserstein distance; s3, generating a health state index (HSI), and smoothing the HSI through index moving average; s4, constructing a monotonic constraint Gaussian process regression model; and S5, estimating the residual life of the satellite component in combination with the early warning and failure threshold values. According to the method, the multi-parameter coupling degradation trend is captured, non-physical rising of a prediction result is avoided, a large number of samples and complex mechanism models are not needed, calculation is efficient, prediction is stable, in-orbit deployment is easy, and the reliability of health assessment and life prediction of satellite components is remarkably improved.

Various embodiments provide for performance of a coarse threshold estimate (CTE) read on a memory device of a memory system under multi-plane mode, such as a memory sub-system.

The invention relates to a method for judging a laser damage dominant mode and estimating a threshold value of a coated optical element, and belongs to the technical field of laser optical element damage evaluation. The implementation method comprises the following steps: calculating thin film light field distribution based on a transmission matrix method or a finite difference time domain method; predicting temperature and stress fields of the film and the substrate by combining finite element analysis; comparing damage threshold values to judge a dominant mode; and the model accuracy is verified through experiments. According to the method, a film electric field-thermal coupling model and a substrate thermal analysis model are respectively constructed, damage contributions of the two models are quantified, and small sample experiment verification is combined, so that the problems that a film-coated optical element film and substrate damage competition mechanism is complex, damage dominant modes are difficult to distinguish, and a traditional experiment method is low in efficiency and high in cost are solved.

The present disclosure relates to a flash memory system including a non-volatile memory and a circuit for performing operations on the non-volatile memory. The circuit can determine, by a machine learning model at a first point in time, a plurality of voltage thresholds for a read operation on the non-volatile memory based on one or more stress conditions at the first point in time. The plurality of voltage thresholds can correspond to respective durations since the first point in time. The circuit can obtain, from the plurality of voltage thresholds, one or more voltage thresholds associated with a particular duration of the respective durations. The circuit can perform the read operation on the non-volatile memory with the one or more voltage thresholds associated with the particular duration at a second point in time later than the first point in time by the particular duration.