50 results about "Fuzzy classifier" patented technology

The invention discloses a method for identifying benign and malignant lung nodules based on multi-dimensional information. The method comprises the following steps: 1, representing a three-dimensional nodule in a two-dimensional way; 2, building a feature model; 3, building a fuzzy classifier; 4, evaluating the classifying performance. The method has the beneficial effects that accurate feature modeling is crucial to the identification of benign and malignant lung nodules. More objective bases are laid for the identification of the benign and malignant lung nodules by adopting imaging diagnosis features, common shapes and textual features in image processing, and patient information, feature extraction is performed on a two-dimensional image generated on the basis of a helical scanning technology, and a novel method is adopted for feature modelling, so that the extracted features are more accurate. In the method, a fuzzy C-means (FCM) clustering algorithm is adopted for identifying benign and malignant status of a suspected nodule, and a probability indicating the suspected nodule is benign or malignant is given, so that the method is more accordant with the thinking mode of a doctor.

The invention, which belongs to the technical field of remote sensing application, particularly relates to a high-resolution-image-based regional adaptive cultivated land extraction method. The method comprises: pretreatment is carried out on data; multi-scale segmentation is carried out on a high-resolution image, an optimal segmentation scale of cultivated land is determined, and an optimal segmentation layer of cultivated land classification is obtained; a plurality of cultivated land sample sets are constructed at the optimal segmentation layer to form an attribute feature space of each cultivated land sample set; optimal image feature space selection is carried out; according to the optimal image feature spaces, fuzzy classifier of cultivated land extraction are constructed; and on the basis of the cultivated land sampling sets and the fuzzy classifiers, cultivated land extraction is carried out. According to the invention, automatic identification of a cultivated land object is realized by combining the cultivated land sample sets and the fuzzy classifiers. The uniformity of the segmented object of the high-resolution image and the extraction precision of the fuzzy classifier are improved by selecting the optimal segmentation scale and the different cultivated land sample sets; and on the basis of construction of the cultivated land sample sets, the restriction requirements on the time phase and the number of the data source are reduced.

The embodiment of the invention provides an automatic paper marking method and system and relates to the technical field of information education. The method comprises processing a test paper image and obtaining a paper marking image according to a fuzzy classifier; extracting the information of each question in a machine paper marking area and comparing the information with the stored information of each question; outputting a score corresponding to the machine paper marking area and recording knowledge point information corresponding to mistakes in the machine paper marking area; generating a comprehensive score from the score of a manual paper marking area and the score of the machine paper marking area, and generating mistake knowledge point information according to the knowledge point information corresponding to the mistakes in the machine paper marking area and the knowledge point information corresponding to the mistakes in the manual paper marking area; finally, generating and storing a student test result report according to student information, the comprehensive score and the mistake knowledge point information. The method and system can not only achieve score statistics on the test results, but also carry on statistics and analysis of the knowledge points.

The invention relates to a cotton foreign fiber classification system and method thereof. Multi-featured vector is formed by selecting color feature, shape feature and texture feature, and corresponding vague classifier will be structured, therefore the class recognition of foreign fiber is achieved. The invention detects the content of foreign fiber in the cotton with convenience, rapidity and precision, and it also classifies the cotton. The invention make the cotton seller reduce the content of foreign fiber in cotton.

The invention relates to a ceramic wall and floor tile surface defect detection device and method based on multi-feature information fusion. The method is characterized by adopting two brightness-adjustable LED white backlight sources to illuminate the ceramic tile surface in an illuminating mode of being inclined to the tile surface at a low angle to obtain a ceramic wall and floor tile with uniform light distribution; collecting image data of the ceramic wall and floor tile through a CCD camera; extracting a suspected defect area of the image data; adjusting the brightness, height and angleof the LED white backlight sources and controlling the height of the CCD camera; collecting image data of a clearer suspected defect area; and finishing multi-feature calculation and information fusion through a fuzzy classifier based on a fuzzy C-mean algorithm, and identifying the defect type corresponding to the suspected defect area. The method can fuse various characteristic quantities, distributed in a grayscale feature space and a morphological feature space, of the suspected defect area of the ceramic wall and floor tile, can accurately and effectively determine existence of the defectand the type of the defect, and is good in defect detection effect.

The invention relates to an active forewarning method for an express way maintenance construction area, the method specifically comprises the following steps: S1, collecting various basic data of a express way maintenance construction section; S2, selecting an original sample by using a K-means clustering algorithm; S3, preprocessing the original sample and calculating an adjoint vector; S4, preprocessing input data to obtain an initial input vector and a traffic state judgment result; S5, judging a fuzzy classifier by combining with weather and static data design forewarning based on the traffic state judgment result obtained by synergetic neural network; and S6, judging data collected in a new period in real time, combing the data into historical data, and performing real time adjustment on parameters of a traffic state judgment model continually by using the synergetic neural network. According to the invention, factors influencing safety around the express way maintenance construction area is considered synthetically, real time judgment and forewarning are performed on the security state of the express way maintenance construction area, and the method has very strong practicability and wide application prospect.

The embodiment of the invention discloses a target tracking method and device based on a TSK fuzzy classifier and a storage medium, and the method comprises the steps: firstly constructing a multi-output regression data set for a feature set of a stable track, and calculating the fuzzy membership degree of each feature relative to a fuzzy rule; secondly, training posterior parameters of a TSK fuzzy classifier based on the motion features and the HOG features based on the multi-output regression data set and the fuzzy membership degree, and constructing a corresponding classifier; inputting theobservation set into a classifier to obtain a label vector matrix, and performing data association on the label vector matrix to obtain correct association of the target and observation; and finally,carrying out filtering and trajectory management on the target to obtain a final trajectory of the target. Through the implementation of the method, the TSK fuzzy classifier is trained by using the multi-frame information, and a multi-feature learning mechanism is added in the training process, so that the learning ability of the classifier is improved, the uncertainty in the data association process can be effectively processed, and the accuracy of target tracking is improved.

The invention provides a crowdsourcing quality improvement method based on an integrated TSK fuzzy classifier. The method comprises the following steps: constructing a plurality of TSK fuzzy classifiers TSK-noise-FC with anti-noise capability by utilizing two types of classification data containing label noise, and generating a plurality of enhanced attributes on a selected verification set without noise data, thereby generating an enhanced verification set; clustering the enhanced verification set by using a classic fuzzy clustering algorithm FCM; forming label-containing multi-representative-point data, recording the data as a data dictionary, quickly correcting the label noise-containing data by utilizing the generated data dictionary, finding the nearest representative point for each noise-containing crowdsourcing data, and taking the label of the representative point as a prediction result. According to the method, the crowdsourcing cost can be reduced, the operation speed is high, a plurality of weak classifiers with the anti-noise capability can operate in parallel, for prediction of crowdsourcing data, only the nearest representative point containing the label needs to be found, and the label of the representative point is the correction result of the crowdsourcing data.

The invention provides a classifier using method comprising: initializing an input sample set and computing the initial value of a similarity matrix of the sample set; selecting and adding a sample to a training set according to the initial value of the similarity matrix, and retraining a classifier until not a hyperbox is expanded. By introducing a new concept of the similarity matrix, the method finds an approach most effectively utilizing the training samples and uses a fuzzy min-max neural network classifier as an application object. All the samples are input at one time and an algorithm depends on the overall structure of the samples. Therefore, a purpose is achieved that the input sequence of the samples does not influence the performance of a fuzzy classifier.

The invention relates to the technical field of data classification, and discloses a data classification method and system based on an improved TSK fuzzy classification model, and the method comprises the steps: obtaining a to-be-classified data set, and dividing the to-be-classified data set into a training set and a test set; based on the training set, utilizing a supervised enhanced soft subspace clustering algorithm and a deep learning algorithm to respectively optimize a fuzzy rule consequent and a fuzzy rule consequent of a fuzzy rule of a pre-constructed TSK fuzzy classification model; and inputting the test set into the optimized TSK fuzzy classifier model for classification processing to obtain a data classification result. According to the method, the fuzzy rule consequent and fuzzy rule consequent parameters of the TSK fuzzy classification model are optimized respectively, so that the adaptability of the model to high-dimensional data and complex data sets is improved, and the precision and efficiency of the model in data classification are further improved.

The invention provides a method for identifying digits in blurred images based on a blur classifier. The method is suitable for identifying the digits in degraded images. The method includes the steps that the degree of truth value that characters in each image are the digits is distinguished according to the distribution characteristics of foreground pixels and external features of the digits, slight slant of the images can be tolerated, the digits in the images can be identified through comprehensive judgment, and the identifying accuracy is high. The method can automatically set parameters needed by the blur classifier based on characteristic value distribution of a small number of training samples, and multi-feature distinguishing and automatic setting of the parameters enable the method to have good universality. The method effectively solves the problems that the accuracy of an existing digit identifying method aiming at the blurred images needs to be improved, and a large number of training samples are needed.

The invention discloses a defrosting control method, device and equipment and an air conditioner, and aims at solving the problem of how to accurately control defrosting time. According to the method,after first defrosting is finished for preset duration, the change rate of outer pipe temperature is obtained, and the defrosting result is evaluated based on a fuzzy classifier according to outdoorenvironment temperature and the change rate of the outer pipe temperature, for example, the defrosting result is evaluated to be very good, good, general, poor and very poor; and then the next defrosting time is adjusted according to the evaluation result. According to the scheme, due to the fact that the next defrosting time is correspondingly adjusted according to the defrosting result, the defrosting time is continuously optimized, the problem that defrosting is not thorough or the defrosting time is too long is avoided, the operation efficiency of the air conditioner is improved, and thenthe comfort of a user is improved.

The present invention discloses a night vehicle detection method of adaptive feature selection. The method comprises the following steps of: S1: collecting video images of a night vehicle, and performing preprocessing of the images; S2: extracting color features, texture features and shape features of the vehicle, fusing the color features, the texture features and the shape features, and obtaining feature vectors of the night moving vehicle; S3: performing training of input samples by a Boosting fuzzy classifier designed through adoption of the Boosting algorithm and the genetic algorithm, each training obtaining one weight corresponding to one fuzzy rule in an iteration mode by employing the genetic algorithm, and performing adaptive selection of features; and S4: employing the Boostingmethod to perform adaptive selection of features being the best for detection in a weighted voting mode to achieve detection of the night moving vehicle.

A handwriting recognition device using fuzzy logic and cellular neural network for unconstrained handwritten numeral classification is provided. The current mode VLSI classifier has a I / O circuit for inputting and outputting a plurality of membership functions. An extraction unit comprising a CCD extractor with a CNN structure and a compression unit receives a to-be-recognized character having a plurality of input features for generating a plurality of features values after compression. A membership function generator stores the plurality of membership functions and receives the plurality of features values to generate a plurality of current-type membership degrees. A plurality of switched-current integrators receives the plurality of current-type membership degrees for generating a plurality of synthesis membership degrees. A k-WTA circuit is provided for comparing the plurality of synthesis membership degrees and output the plurality of synthesis membership degrees as well as the corresponding characters in an order of magnitude.

A reflectivity correction method using a double polarization variable-based bright band detection result includes a preprocessing operation for correcting a double polarization variable observation error and calculating a depolarization ratio; a fuzzy classifier generation operation for calculating a weighting and a membership function of each characteristic variable using a bright band height extracted from a quasi-vertical profile generated from specific elevation angle data, a bright band detection operation for detecting a bright band using a depolarization ratio and a fuzzy classifier for each elevation angle, and a reflectivity correction operation for correcting reflectivity over-observation for a detected bright band region on the basis of a correction factor calculated using an apparent profile of reflectivity generated by averaging reflectivity data for the bright band region for each elevation angle. Thus, it is possible to improve the accuracy of precipitation estimation by using the corrected reflectivity.

The invention relates to the field of image classification, and provides a training method of an image classification model, and an image classification method and device. According to the invention, a plurality of TSK fuzzy classifiers are respectively used as a layer of subunits, the plurality of layers of subunits are constructed into the image classification model in a layer-by-layer cascade mode, and the input of each layer of subunits is updated layer by layer by using the TSK fuzzy rules which have great influence on the performance of the model and the importance fuzzy rules which are determined based on the incidence relation among the TSK fuzzy rules. The association between the TSK fuzzy rules is considered, the TSK fuzzy rule which greatly influences the classification performance has a larger weight, and the input of the subunits is updated layer by layer by using the importance fuzzy rule, so that the distinguishability of different types of samples can be improved, and therefore, the image classification model can effectively improve the classification performance layer by layer and has excellent generalization ability; when the problem of large-scale image classification is solved, higher-level performance can be obtained, and the large-scale image classification precision is improved.

The invention discloses a TSK fuzzy classifier fusing decoupling knowledge distillation and curriculum learning, samples are subjected to difficulty ranking through curriculum learning, a training strategy is formulated, a model is made to start learning from easy samples and gradually go to complex samples and knowledge, the TSK fuzzy classifier has higher convergence efficiency, and the classification efficiency is improved. And then knowledge contained in the teacher model is extracted into the TSK fuzzy classifier through decoupling knowledge distillation, and target category loss and non-target category loss can be better coupled through decoupling knowledge distillation, so that the precision and performance of the TSK fuzzy classifier are improved.

The present invention proposes a deep TSK fuzzy classifier based on multi-level feature fusion, including a feature learning module and a knowledge reasoning module; the feature learning module is a feature learning module based on a convolutional neural network, which takes the original data as the input of the deep TSK fuzzy classifier based on multi-level feature fusion, and the feature learning module extracts local information in a local connection manner (i.e., convolution kernel), and obtains the deep features of the hidden layer through layer-by-layer neural expression to automatically extract deep features from the original data; the knowledge reasoning module is a knowledge reasoning module based on the TSK fuzzy classifier, which takes the deep features as the training parameters of the fuzzy rules, adopts the FCM clustering algorithm to generate the antecedent parameters of the fuzzy rules, and uses the original data to train the consequent parameters of the fuzzy rules. By fusing the feature extraction capability of the convolutional neural network and the uncertainty processing capability of the fuzzy representation, a deep TSK fuzzy classifier that is easier to understand is formed.

The invention belongs to the technical field of remote sensing applications, and relates to a region-adaptive cultivated land extraction method based on high-resolution images, including: preprocessing data; performing multi-scale segmentation on high-resolution images, and determining the optimal segmentation scale of cultivated land , to obtain the optimal segmentation layer of cultivated land classification; construct multiple cultivated land sample sets in the optimal segmentation layer to form the attribute feature space of each cultivated land sample set; select the optimal image feature space; construct the cultivated land according to the optimal image feature space The extracted fuzzy classifier; based on the cultivated land sample set and the fuzzy classifier, the cultivated land is extracted. The invention realizes the automatic recognition of the cultivated land object through the combination of the cultivated land sample set and the fuzzy classifier. By selecting the optimal segmentation scale and the attribute feature space of different farmland sample sets, the homogeneity of the segmentation objects of high-resolution images and the extraction accuracy of the fuzzy classifier are improved; Requirements for phase and quantity restrictions.

The invention discloses an online learning method and device based on data fuzziness and a medium, and aims to solve the technical problems that too much input data is input into an online learning model, and typicality data in the input data is relatively sparse. The method comprises the steps that a server completes training of fuzzy classifiers based on training sample data, the number of the fuzzy classifiers is 2N+1, and N is a positive integer; a data stream from the terminal is input into the fuzzy classifier; the ambiguity of each input data is determined in the data stream; according to a preset selection mode and the ambiguity of the input data, input data with typicality in a data stream from the terminal is determined, the input data with the typicality is sent to the terminal, wherein the input data with the typicality is used for enabling the terminal to update an online learning model of the random weight neural network. According to the method, the quantity of the input data input into the online learning model is reduced, and meanwhile, the typicality of the input data is ensured.

The invention provides an air conditioner control method and device and electronic equipment. The air conditioner control method is used for controlling an air conditioner with a preset working area, and comprises the following steps that an outdoor temperature and a working area temperature are acquired; the outdoor temperature and the working area temperature are input into a fuzzy classifier to obtain an air conditioner set temperature; and the air outlet temperature of the air conditioner is controlled based on the air conditioner set temperature. The fuzzy classifier is adopted to calculate the set temperature, the actual outdoor temperature and the working area temperature are comprehensively considered, automatic adjustment can be carried out according to the outdoor environment change and the working area temperature change, the comfort of a user during cooking is guaranteed, the operation frequency of the user is further reduced, and therefore the user experience is improved.

The present application provides a method and apparatus for medical data classification based on mutual correction and information fusion, the method includes: obtaining a CT image and / or an MRI image of a target object; and obtaining a medical data classification result output from a fuzzy classification model, and the training process of the fuzzy classification model includes: training and internal correction of the decision information of the CT view and the MRI view; and information fusion of the decision information of the two views to obtain a medical data classification result. The embodiment of the present invention uses a zero-order TS fuzzy classifier as a basic training unit to ensure the interpretability of model training, reduces the feature dimension by using KL divergence to reduce the pressure of the decision process, corrects the decision information required for classification by multiple sub-views in the two views, and performs information fusion to capture the diversity of decision information, and utilizes the advantages of each view in the decision aspect to realize further optimization of the decision information, thereby improving the accuracy of classification.

The invention discloses a fuzzy integrated classifier with high interpretability based on parallel learning. The fuzzy integrated classifier sequentially comprises the following steps of constructing a zero-order TSK fuzzy classifier fastly; constructing an enhanced verification data set, calculating the value of each corresponding output function for each sample of the verification data set, and taking the value of each corresponding output function as an enhanced feature of the original sample so as to form the enhanced verification data set; respectively calling an FCM algorithm for each class on the enhanced verification data set to generate a series of representative center points and corresponding tags thereof, and removing the enhanced features of the representative center points; for any test sample, finding k representative center points closest to the test sample by using the KNN to determine the class label of the test sample. According to the present invention, a parallel learning method is adopted, the operation complexity is low, the structure is simple and easy to understand, the fast learning can be carried out, and the good performance is achieved on the classification problem processing.

The invention relates to a method for extracting the correlation coefficient between special loads and day loads of a power grid. The method includes the steps of collecting historical day load data of the power grid, the special loads and influence factors of the special loads on the day loads; refining the influence factors to obtain an influence factor sequence; conducting analysis to extract a corresponding correlation coefficient sequence M of the influence factors and the historical day load data; building a day load influence factor fuzzy classifier including all the influence factors, conducting fuzzy assignment classification on the influence factors, and obtaining fuzzy classifier vectors; conducting analysis to obtain a proportion weighting relation sequence R between the special loads and the influence factors in the fuzzy classifier vectors; conducting analysis to obtain a correlation coefficient sequence N between the special loads and the day loads in cooperation with the correlation coefficient sequence M and the proportion weighting relation sequence R in a weighting mode, and completing analysis of the correlation of the special loads to the day loads. By means of the method, the influences of the multiple special loads on the day loads can be systematically analyzed, and good and effective technical bases can be provided for forecast work for giving consideration to the influences of the special loads on the day loads.

The invention relates to an active forewarning method for an express way maintenance construction area, the method specifically comprises the following steps: S1, collecting various basic data of a express way maintenance construction section; S2, selecting an original sample by using a K-means clustering algorithm; S3, preprocessing the original sample and calculating an adjoint vector; S4, preprocessing input data to obtain an initial input vector and a traffic state judgment result; S5, judging a fuzzy classifier by combining with weather and static data design forewarning based on the traffic state judgment result obtained by synergetic neural network; and S6, judging data collected in a new period in real time, combing the data into historical data, and performing real time adjustment on parameters of a traffic state judgment model continually by using the synergetic neural network. According to the invention, factors influencing safety around the express way maintenance construction area is considered synthetically, real time judgment and forewarning are performed on the security state of the express way maintenance construction area, and the method has very strong practicability and wide application prospect.

The invention relates to a cotton foreign fiber classification system and method thereof. Multi-featured vector is formed by selecting color feature, shape feature and texture feature, and corresponding vague classifier will be structured, therefore the class recognition of foreign fiber is achieved. The invention detects the content of foreign fiber in the cotton with convenience, rapidity and precision, and it also classifies the cotton. The invention make the cotton seller reduce the content of foreign fiber in cotton.

The invention discloses an integrated TSK fuzzy classifier based on IFCM, KNN and a data dictionary. A method sequentially comprises the following steps: randomly sampling a training data subset fromDtr for each sub-classifier; training l zero-order TSK fuzzy sub-classifiers in parallel; for each sample of the verification data set, calculating the value of each corresponding output function, taking the value as an enhancement feature, and combining the original feature with the enhancement feature to form an enhancement verification data set; after an IFCM algorithm is used for up-regulationof the enhanced verification data set, generating a representative center point and a corresponding tag thereof, and removing enhanced characteristics to obtain a data dictionary; and for any test sample, finding k nearest points on the data dictionary by using a KNN method, and determining the class label of the test sample based on a voting strategy. The classification method has high classification performance, high robustness and high interpretability.

This invention relates to the field of TSK fuzzy classifier technology, and particularly to an AD-TSK-based data processing method and system. The method includes training a training set containing active adversarial examples using a TSK fuzzy classifier to generate an ATSK fuzzy classifier; training ATSK layer by layer to obtain latent class information; generating an output error gradient matrix based on attack labels using dropout, and adding it to the corresponding sample features to update the selected input sample features, which are then used as input for the next layer of ATSK; repeating this process until the maximum number of layers is reached, resulting in an AD-TSK model. This invention addresses the problem that embedding gradient information in all sample features in existing DSA-FC methods may lead to a decrease in model generalization performance, and that the inversion operations involved cause significant computational overhead during training.

The application discloses a fall risk classification method, medium and equipment based on daily walking foot pressure data, and belongs to the technical field of data classification and prediction. The method comprises the following steps: collecting foot pressure data in a target daily walking process; pre-processing and extracting features of the foot pressure data to obtain foot pressure feature data; taking the foot pressure feature data as a sample to be classified, and using a trained fuzzy classifier to obtain a fall risk classification result; and calculating the membership degree between samples to obtain the fall risk classification result for the foot pressure feature data. The application has low complexity, high execution efficiency, small required computing power, can be processed in real time on an edge terminal, and is more widely applied.

The invention relates to an interframe error code masking method based on fuzzy classification, which belongs to the technical field of video communication based on macroblock coding. The method mainly comprises the following three steps of: 1. establishing a cache region of a decoded frame, carrying out fuzzy characteristic value extraction on the decoded frame and classifying a trained fuzzy classifier; 2. carrying out interframe error code masking on a lost macroblock of the current decoded frame by utilizing all available decoded frames judged by the fuzzy classifier in the cache region of the decoded frame; and 3. fitting data obtained after all available rebuilding frames carry out error code masking on the current error code frame to obtain a final error code masking result.