97results about How to "Reduce the number of features" patented technology

The invention discloses a hyperspectral image super-resolution restoration method based on a 3D convolutional neural network. According to the technical scheme, the 3D residual dense network is characterized by comprising 3D convolution kernel to convolve the hyperspectral image spectral dimension and 3D sub-pixel recombination to enlarge the image and reconstruct the high-resolution image part, and unifying the two parts in the deep convolutional neural network framework 3D-RDN; hierarchical characteristics of the convolutional layer are fully utilized through structures such as residual dense blocks, and super-resolution restoration of the hyperspectral image is achieved. At present, when an existing method based on deep learning is applied to a hyperspectral image, the characteristics of the hyperspectral image are not fully considered, and therefore it is difficult to effectively utilize rich spectral dimension information of the hyperspectral image to reconstruct a high-resolutionimage. According to the method, all spatial spectrum information of the hyperspectral image is fully utilized, efficient super-resolution restoration is achieved, and the PSNR value is superior to that of an existing method.

The invention belongs to the technical field of mode recognition and information processing and provides a mass image retrieval system based on cluster compactness. Steps include 1, calculating local features of images in a sample image library and a test image library; 2, calculating cluster compactness of each image, namely clustering the local features to acquire each type of cluster centers, counting a local feature distribution histogram and spatial statistical information of each cluster, and generating cluster compactness; 3, randomly sampling cluster compactness of the sample image library, clustering components of the cluster centers in the sampled cluster compactness to generate a vocabulary tree, and quantizing the cluster compactness of the images in the test image library to the vocabulary tree to generate corresponding inverted files; 4, retrieving by a modified retrieval algorithm based on the vocabulary tree, namely retrieving, by retrieving the inverted files in the vocabulary tree and calculating the weight of similarity between retrieval images and the image library image cluster compactness.

The invention discloses a topic model-fused scene image classification method, and relates to the field of deep learning and image classification. The method comprises the following steps of: preprocessing data sets, and expanding the quantity of obtained data sets so as to obtain an image data format according with deep learning model processing; constructing a convolutional neural network modelaccording with scene image classification, and pre-training the processed image data sets by using a convolutional neural network; and carrying out end-to-end iterative training on the constructed convolutional neural network by using training sets, adjusting parameters in the network, verifying the trained model by using verification sets, modeling extracted scene image features with discrimination ability, extracting hidden topic variables between the features and images so as to obtain image topic distribution represented by a k-dimensional vector, wherein k represents a topic quantity, each image can be considered as a probability distribution diagram formed by a plurality of topics, and scene image classification is realized by utilizing a classifier.

The invention provides a fault feature screening method based on weighted multi-feature fusion and SVM classification. The method comprises the following steps of 1, obtaining the time series data; 2,extracting a time domain (T), a frequency domain (P), the energy (E) and an entropy feature (S), and forming a high-dimensional feature set (Q); 3, screening out the features (Q1) with the diagnosisrate greater than 50%, carrying out the correlation analysis, and removing the features (Q2) with the similarity greater than 85%; 4, selecting a feature with the highest score through PCA and a loadscoring method to form a new sub-feature set (T3, P3, E3, S3); 5, carrying out SVM diagnosis on the T3, the P3, the E3 and the S3, and obtaining a weight Wi according to the diagnosis rate Ri; 6, performing the weighted fusion of the features; and 7, inputting the fused features into a classifier for diagnosis. Through the above steps, a group of optimal features capable of maintaining the fault intrinsic information is obtained, the original failure information represented by the features is ensured, the fault diagnosis accuracy is improved, and the method is of great significance to the efficient mechanical fault diagnosis.

The application provides a medical image data processing method. The method comprises the steps as follows: acquiring a first training image which has first contrast information; acquiring second contrast information of a second training image, wherein the second training image is generated by adjusting window width and / or window level of the first training image; training a first neural network model based on the first training image and the second contrast information, and configuring the trained first neural network model to be able to convert contrast information of an image to be processed into the contrast information of a target image.

The invention discloses an intelligent rotary machine fault depth network feature identification method. A vibration sensor is arranged at a to-be-detected rotating mechanical part of a train rollingbearing; collecting an original vibration sequence when the rolling bearing works; decomposing and reconstructing the original vibration sequence through a singular spectrum analysis method; extracting a root-mean-square value of the reconstructed vibration sequence; standard deviation, skewness index and peak value; a fault position is judged by using a rotary machine fault position diagnosis model obtained by training of a support vector machine; and then, carrying out ensemble empirical mode decomposition on the reconstructed vibration sequence, calculating the permutation entropy values ofa group of decomposed intrinsic mode components, taking the permutation combination of the permutation entropy values as a detection characteristic, and judging the fault type by using a rotary machine fault type diagnosis model obtained by training of a support vector machine. The fault position and the fault type of the rotary machine can be detected more timely, and the fault diagnosis accuracy and reliability are improved.

The invention discloses a face recognition model construction method and device, a face recognition method and device and a storage medium. The method comprises steps of preprocessing face pictures, extracting four local areas, namely a left eye area, a right eye area, a nose area and a mouth area, of the preprocessed face image to perform face partitioning; amplifying the data on the basis of a partitioning result, respectively amplifying the data by considering the shielding of 1 / 2 and 1 / 4 conditions, and finally training a deep learning neural network by using the amplified data to construct a training feature data set, thereby realizing the face recognition based on feature matching. The method is good in shielded face recognition effect, does not need a large number of shielded samples, and is small in occupied memory.

The invention discloses an electroencephalogram signal processing method and an epilepsy detection system. The electroencephalogram signal processing method comprises the steps that data is preprocessed, wherein five electroencephalogram frequency bands are decomposed by performing discrete wavelet transformation on an electroencephalogram signal data set; a wavelet coefficient is solved; one electroencephalogram frequency band with the highest frequency is eliminated, the four remaining low-frequency frequency bands are reconstructed through reverse discrete wavelet transformation, a signal after high-frequency components are removed is obtained; time-domain features and entropy-based features are extracted from the signal obtained through reconstruction, and an alternative feature set isconstructed; the optimal feature subset is selected from the alternative feature set by using a feature selection method based on feature correlation. The electroencephalogram signal processing method has the effect of selecting the optimal feature subset by utilizing the improved feature selection method based on the feature correlation for the extracted features.

The invention relates to an image recognition method and device. The method comprises the following steps that: the features of a target image are extracted; dimensionality reduction is performed on the features of the target image according to a principal component analysis method; and the target image which has been subjected to dimensionality reduction is recognized according to a pre-trained image model. With the image recognition method adopted, a terminal can perform dimensionality reduction on the features of the target image according to the principal component analysis method, and therefore, the number of the features of the target image can be decreased, and the key features of the target image can be reserved; the terminal can recognize the target image which has been subjected to dimensionality reduction according to the pre-trained image model; and therefore, the efficiency of image recognition can be improved, and the accuracy of image recognition will not be reduced, and the use experience of a user can be improved.

The invention discloses a playback speech detection method. At a training stage, a first variation coefficient vector, a normalized first cepstrum feature matrix, a second variation coefficient vectorand a normalized second cepstrum feature matrix of each speech sample in a speech database are firstly acquired as four kinds of features; the four kinds of features of all positive samples are theninputted to a GMM model for training, four positive sample feature models are obtained, and besides, four negative sample feature models are obtained; at a testing stage, four kinds of features for to-be-detected speech are acquired in the same way, the kinds of features are inputted to the corresponding positive sample feature models and the negative sample feature models respectively, and four likelihood ratio scores are obtained; and according to the four likelihood ratio scores, a final score is obtained, and through compairing the final score and a judgment threshold, whether to be a playback speech is judged. The playback speech detection method is not only limited to a text-related voiceprint authentication system, and has the advantages of low equal error detection possibility, strong robustness and relatively low calculation complexity.

The invention provides a disk fault prediction method and system, a terminal and a storage medium. The disk fault prediction method comprises the steps of screening SMART feature items related to diskfault prediction by utilizing a recursive feature elimination algorithm; generating a disk health parameter according to the fluctuation condition of the related feature item of the disk in the preset time period; training a neural network model by using the historical disk SMART feature data; and inputting the disk health parameters into a neural network model to obtain a disk fault prediction result. The SMART features are screened by using the RFE algorithm, the number of the features is reduced, the calculated amount is reduced, overfitting is inhibited, the new feature value is generatedaccording to the change times of the SMART feature value within seven days, and the proportion, the balance accuracy and the recall rate of two types of errors in the loss function are trained and adjusted more easily.

The invention relates to a power system reliability evaluation technology, in particular to a power system risk evaluation method based on cross entropy theory optimization support vector machine, wherein the cross entropy method CEM is applied to a support vector machine SVM to carry out feature selection and SVM parameter optimization; the n-dimensional feature quantity is used as the input of SVM, and the training model is generated by off-line training. The risk level of the power system under the current state is predicted by learning the historical samples. The method includes generatingoff-line data, training the risk sample data by using cross-entropy optimization SVM method, and obtaining the optimized characteristics and parameters of power network risk assessment model; the on-line real-time power system risk assessment is carried out. Using cross-entropy algorithm to optimize support vector machine for risk prediction can effectively remove redundant and irrelevant features, reduce the number of features, combine the optimized parameters can have good convergence, and reduce the computational complexity and time-consuming. It has the characteristics of low error rate and short test time.

The invention discloses a method for realizing image registration of a synthetic aperture radar (SAR) by using three components of monogenic signals, belonging to the technical field of image processing and registration of SARs. The conventional gradient information-based detection method and cross correlation matching method have the defects of excessive detected characteristic points, overlong detection time, low registration accuracies and the like when applied to image registration of an SAR. A method for matching a detection algorithm of monogenic signal phase congruency with monogenic signal phase correlation is given, and three orthogonal monogenic signal component signals are generated by designing a frequency domain Log-Gabor filter. One path of the monogenic signal component signals is transmitted to a characteristic detector, i.e., a local amplitude and a local phase are resolved by using three components to construct a monogenic signal phase congruency function for detecting the phase congruency characteristic. Another path of the monogenic signal component signals is transmitted to a matcher, and a characteristic description vector is constructed by using three components of the monogenic signals; a characteristic vector correlation matrix is obtained by calculating the characteristic vector correlation of a reference image and characteristic points in an image to be registered; and largest line elements and column elements in the characteristic vector correlation matrix are searched and are indexed to a coarsely-matched characteristic point pair, so that coarse characteristic matching is realized. An affine basic matrix is fitted by using an RANSAC (Random Sample Consensus Algorithm), so that accurate matching of characteristic points is completed. An affine conversion model is used for realizing image registration of the SAR. The algorithm disclosed by the invention has the advantages of realization of automatic registration of SAR images, high registration speed, small influence by speckle noise, high registration accuracy and popularization and application values.

The invention relates to the technical field of data processing and analysis. The invention discloses a big data exploration and cognition method, device and equipment and a computer storage medium. Therefore, in the aspects of big data application engineering, ETL development, data processing, cleaning, integration, analysis modeling and the like, and massive data of various different data sources, the total collection information containing each data table and even each field can be explored and obtained, and the total collection information is automatically displayed in a visual view and table, so that a user obtains comprehensive cognition on the data. At present, many data warehouses, ETL, BI, data mining, machine learning and big data analysis projects are often informed of tiger heads and snakes or failures. The important reason is that the data universal set is inadequate in cognition, has deviation or is generally complete at the beginning, so that the user can be better helped and guided to plan and develop the projects in a targeted and symptomatic way, and the projects are prevented from being ended by tiger snakes or failures as much as possible.

The invention discloses a sleep staging method based on EEG time domain multi-dimensional features and M-WSVM, and a wearable device. The method comprises the following steps: acquiring EEG continuoustime signals, and extracting time domain multi-dimensional features of the EEG continuous time signals by amplitude-time mapping; selecting the extracted signal features to obtain optimal signal features; and analyzing and processing the sleep stage at different classification levels by using an M-WSVM algorithm, and monitoring the sleep staging in real time. The device comprises a signal collecting module, a signal processing module and a signal transmission module, and can communicate with a user end of an intelligent device in real time, model learning of the EEG training data is carried out on a PC end, and a learning algorithm model is transplanted on the intelligent device to monitor the sleep staging in real time. The method for extracting and classifying the features of the EEG signals is used to simplify the complexity of the sleep staging, and a physiological signal measurement circuit is used to develop the wearable sleep staging device in order to obtain thee real-time andhigh-precision automatic sleep staging effect.

The invention discloses a method and device for selecting characteristics based on neural network sensitivity, and belongs to the field of machine learning in intelligent science and technologies. The method and device are based on the sensitivity, the characteristics high in sensitivity are selected to search for the characteristics with output greatly changed when disturbance happens to the characteristics, and the characteristics are important to a training classifier. The method and device can effectively select the sample characteristics important to the training classifier, and therefore the performance of the classifier is improved.

The invention discloses an air conditioner fault diagnosis method based on a Bayesian optimization PCA-limit random tree. The air conditioner fault diagnosis method comprises the following steps: 1) acquiring operation data of an air conditioner under normal and different faults and normalizing the operation data; 2) carrying out dimensionality reduction on the normalized data through a PCA algorithm, and taking the normalized data as the input of an ExtraTree model; 3) establishing a limit random tree classification model, training and testing a classifier, and obtaining a PCA-limit random tree fault diagnosis model for an air conditioner; 4) utilizing a Bayesian optimization algorithm to optimize the feature number and the CART decision tree number of a PCA-extreme random tree fault diagnosis model after the PCA dimension reduction to obtain the optimal feature number and the optimal CART decision tree number after the dimension reduction; and 5) then, taking the calculated optimal PCA dimension-reduced feature quantity value and CART decision tree quantity value as parameters of a PCA-limit random tree model, training a sample to obtain a PCA-limit random tree fault diagnosis model, and then using the diagnosis model to diagnose real-time data.

The invention discloses a driver eye state monitoring method based on invariant moment. After an facial image is preprocessed, an eye area is extracted out according to an image threshold segmentation method based on the two-dimension digital image fractional order integral and the Legendre moment, first three orders of central moments and four feature quantities of the eye area are combined to serve as a feature vector for matching recognition of eye states, Euclidean distances between feature vectors of an area to be detected and feature vectors of a template image are calculated one by one, and thus the eye state of a driver is judged out. The respective advantages of the central moments and the feature quantities of the eye area are combined, the number of features for matching recognition is decreased, and space dimensionality of the features is reduced; by introducing the Euclidean distances between the feature vectors of the candidate eye area and the feature vectors of the eye template image, algorithm complexity is further lowered, and the system recognition speed is increased.

The invention discloses a text classification method based on an improved firefly algorithm and K neighbors. A text feature selection model is constructed by combining information gain and the fireflyalgorithm. The method comprises the following steps: all features are sorted by using information gain, and then a more representative feature subset on a feature set sorted in the front is found out by using the relatively strong optimization capability of an improved firefly algorithm. The step length factor alpha in the firefly algorithm is adjusted, so that the global search capability of the algorithm is ensured, and the local search capability is also ensured. A new fitness function is introduced, so that the dimensionality of the features is properly reduced on the basis of improvingthe precision of the feature subsets. And finally, the model is used for text feature selection, and the obtained feature subset is used for KNN text classification. According to the method, the defects that a firefly algorithm is prone to early maturing and falling into local optimum, the convergence speed is low and the like in the process of searching for the optimal text feature subset can bewell overcome, so that a more accurate subset is obtained, and the text classification accuracy is improved.

In the present invention, there is provided a data processing device, including: a first memory configured to store identification information for identification of video data in association with a feature included in the video data; a second memory configured to store the feature included in the video data in association with the identification information for identification of the video data; a first reader configured to read out the identification information stored in the first memory based on a feature included in input video data; a second reader configured to read out the feature stored in the second memory based on the identification information read out by the first reader; and a checker configured to compare the feature included in the input video data with the feature read out by the second reader to determine whether or not the input video data matches video data whose feature is stored in the second memory.

The invention discloses a network intrusion data detection method. The method comprises the following steps: acquiring real data in a network as target data; according to the category of the target data, extracting all features of the target data and performing preprocessing to generate a first feature set; judging whether the relevance between each feature in the first feature set and the intrusion data is irrelevant or not; if not, removing the features to generate a second set of features; determining the importance degree of the target feature in the second feature set; judging whether thetarget feature is a redundant feature or not according to the importance degree; if yes, removing the target features to generate a target feature set; and detecting intrusion data in future networkdata according to the target feature set. According to the invention, the detection efficiency and accuracy are improved, and the security of network data is better ensured. In addition, the inventionalso provides a network intrusion data detection device, equipment and a medium corresponding to the method.

Embodiments of the invention provide a target characteristic data mining method and apparatus. The method comprises the steps of performing characteristic frequency statistics on first characteristic data; filtering low-frequency characteristic data from the first characteristic data according to a characteristic frequency, thereby obtaining second characteristic data; and filtering at least part of medium-frequency characteristic data from the second characteristic data according to the characteristic frequency, thereby obtaining target characteristic data. According to the method and the apparatus, the performance of a model is basically not influenced; and while the machine learning effect is ensured, the characteristic quantity is greatly reduced, so that the required machine quantity and resource quantity are greatly reduced, the training time is greatly shortened, the training speed is increased, and the training cost is greatly reduced.

The invention discloses a botnet detection method and system and a storage medium, and the method comprises the steps: obtaining the original network flow data in a monitored network, carrying out thepreprocessing of the original network flow data, and obtaining the preprocessed network flow data; constructing a terminal access relation graph based on the preprocessed network traffic data; and mining a plurality of terminal identification lists accessing the same domain name from the terminal access relation graph to obtain a candidate node combination, and screening the candidate node combination based on a preset screening rule to obtain a botnet node detection result. According to the scheme, the types of the applied data are few, the number of the features extracted from the data flowis small, the calculation overhead is small, the detection efficiency can be effectively improved, detection does not need to be carried out based on the known botnet behavior features, and the scheme can be better applied to detection of unknown botnet threats.

The invention discloses a target tracking method based on fast tensor singular value decomposition feature dimensionality reduction. The target tracking method comprises: extracting multiple featuresfrom each frame of video data, and constructing a tensor structure; performing singular value decomposition on the constructed tensor; and training related filters by using the features after dimension reduction, and tracking the target. According to the method, the number of features can be effectively reduced, the tracking speed is increased, and compared with a traditional vector-based principal component analysis feature dimensionality reduction mode and other modes, the structure information of the features is better reserved; the tensor singular value decomposition has invariance to therotation of the feature to enhance the robustness of the tracker to the target rotation.

The invention discloses an image segmentation method based on background detection. The method comprises the following steps: acquiring N silicon wafer images shot by X-ray; respectively carrying outcolor image segmentation on the N silicon wafer images, screening out a single-channel image with the maximum contrast ratio from the N silicon wafer images, and performing further graying to obtain Nfirst target images; carrying out global threshold processing on the N first target images, carrying out hole filling until no gap exists in pixels, and obtaining N second target images; constructinga connected domain of the N second target images according to a set pixel adjacency relation, and further screening out a silicon wafer region by utilizing the area characteristics of the silicon wafer to obtain N third target images; and performing morphological opening operation on the N third target images, reducing the definition domain of the N third target images to a set definition domain,and performing matting processing to obtain N target images. According to the method and device, the original image is subjected to series processing, so that the feature number is reduced on the premise of ensuring the image quality, and the target image is quickly segmented from the original image.

The invention relates to a feature extraction method based on a wood color space. Firstly, a color space of wood colors in an RGB color space in a shooting environment is found out, and then effectivewood color features are extracted and can be used for training as features by taking a machine learning algorithm.

The invention discloses a malicious code obfuscation feature cleaning method, and belongs to the field of machine learning information safety. The method involves a feature selection method and an obfuscation feature cleaning method, and the effectiveness of a traditional malicious code feature extracting method is improved. Compared with the traditional malicious code feature extracting method, the malicious code obfuscation feature cleaning method can effectively prolong the effective time limit of a malicious code feature extracting algorithm, and improve the interference resistance of thefeature extracting algorithm. Firstly, a feature library is built through an n-gram feature extracting method. Since the feature extracting algorithm cannot solve the obfuscation operation problem ofmalicious codes, the feature library contains a large number of obfuscation feature values of the malicious codes. Through an obfuscation feature cleaning algorithm, the interference of abnormal datain a model identification rule can be removed. On this basis, from the aspect of the scale of a training dataset, a feature selection method is put forward. By means of the malicious code obfuscationfeature cleaning method, on the basis of guaranteeing that the model identification precision is not lowered, the number of features which are finally used in the model is effectively lowered.

The invention discloses a representation and classification method of high-order and high-dimensional image data, belonging to the field of pattern recognition, which solves the problems of noise influence in the process of image recognition, destruction of original image structure, high data feature dimension, incompatibility between expression algorithm and classification algorithm and the like.The technical scheme adopted by the invention is that the image data is projected from the original space into the low-rank subspace according to a pattern by using a projection matrix set to obtaina low-rank representation of the data; a sparse representation dictionary and a linear classifier are trained in the low rank subspace to classify the image data. The image data is projected from theoriginal space to the low rank subspace by using the projection matrix set, and the low rank representation of the image data is obtained. Sparse representation dictionaries and linear classifiers aretrained in low rank subspaces to classify image data.