A method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition

Abstract

The present disclosure relates to a method and device for extracting and detecting time-frequency joint features of epilepsy based on covariance decomposition, an electronic device and a storage medium. The method comprises: performing data preprocessing on collected electroencephalogram (EEG) data to generate EEG data; calculating a covariance matrix, eigenvalue decomposition, feature extraction and stretching processing on the EEG data after decentralization to generate a time-domain feature vector; performing spectral feature extraction and frequency-domain covariance feature extraction respectively to generate a spectral feature vector and a frequency-domain covariance feature vector; generating time-frequency joint features based on a preset feature fusion strategy; and classifying and identifying the time-frequency joint features based on a preset binary classification method to complete epilepsy prediction. The present disclosure extracts and fuses time-frequency multi-scale feature information, which can effectively shorten the manual data labeling time of medical workers, improve the labeling efficiency of epilepsy seizure events, and provide a new approach and method for clinical application of epilepsy detection.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of electroencephalogram signal processing and application, and in particular, to a method and device for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition, an electronic device, and a computer readable storage medium. BACKGROUND

[0002] Epilepsy is a chronic non-communicable disease caused by abnormal paroxysmal hyper-synchronous electrical activity of neurons in the brain, and is one of the most common neurological diseases in the world. Accurate detection of the status of seizures is a prerequisite for providing diagnosis and treatment for patients. The gold standard for traditional epilepsy detection is manual annotation of long-term electroencephalogram data collected from patients with epilepsy by medical workers based on their experience. However, due to the randomness of seizure time and the uncertainty of seizure duration, the traditional detection method requires a large amount of time and effort from medical workers. In order to overcome the limitations of traditional diagnosis methods and improve the efficiency of diagnosis, it has become a research hotspot to study epilepsy detection methods using machine learning or deep learning methods.

[0003] Epilepsy detection includes seizure onset detection (SOD) and seizure event detection (SED). SOD refers to identifying the start time of a seizure as soon as possible with the shortest delay during continuous electroencephalogram monitoring when a seizure occurs, and describing the duration, which can provide a clinical reference for locating the epileptogenic focus. SED refers to accurately identifying seizure signals from multiple electroencephalogram signals, which can be used in clinical scenarios such as electroencephalogram data annotation, and can reduce the burden on medical workers, and has important theoretical significance and practical application value. For SED, researchers extract epilepsy features from four aspects: time domain, frequency domain, time-frequency domain, and nonlinear methods. Many feature extraction methods have been developed, such as amplitude, zero-crossing rate, sample entropy, etc., which have expanded the means of epilepsy detection. However, existing research methods usually extract single-dimensional (such as time domain) feature information, and do not extract and fuse feature information contained in different scale spaces, which limits the further improvement of epilepsy detection performance.

[0004] Therefore, one or more methods are needed to solve the above problems.

[0005] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The purpose of the present disclosure is to provide a covariance decomposition-based epilepsy time-frequency joint feature extraction and detection method, device, electronic equipment and computer readable storage medium, thereby at least to some extent overcoming one or more problems caused by the limitations and defects of the related art.

[0007] According to one aspect of the present disclosure, a covariance decomposition-based epilepsy time-frequency joint feature extraction and detection method is provided, comprising:

[0008] A data preprocessing step is performed on the collected electroencephalogram data to generate multiple data segments of electroencephalogram data;

[0009] A time-domain feature extraction step is performed on the multiple data segments of electroencephalogram data after data preprocessing to calculate the covariance matrix, eigenvalue decomposition, feature extraction and stretching processing, and generate a time-domain feature vector;

[0010] A frequency-domain feature extraction step is performed on the multiple data segments of electroencephalogram data after data preprocessing to extract spectral features and frequency-domain covariance features, respectively, and generate a spectral feature vector and a frequency-domain covariance feature vector;

[0011] A feature fusion step is performed based on a pre-set feature fusion strategy to fuse the time-domain feature vector, spectral feature vector and frequency-domain covariance feature vector to generate a time-frequency joint feature;

[0012] A classification and recognition step is performed based on a pre-set binary classification and recognition method to classify and recognize the time-frequency joint feature and complete epilepsy prediction.

[0013] In an exemplary embodiment of the present disclosure, the data preprocessing step of the method further comprises:

[0014] The collected electroencephalogram data is filtered, down-sampled, segmented, baseline corrected and pseudo-trace removed for data preprocessing to generate multiple data segments of electroencephalogram data.

[0015] In an exemplary embodiment of the present disclosure, the time-domain feature extraction step of the method further comprises:

[0016] The multiple data segments of electroencephalogram data after data preprocessing are respectively decentered to calculate the covariance matrix;

[0017] The eigenvalue decomposition is performed on the covariance matrix to obtain a feature vector;

[0018] The feature vector is spliced to generate a feature splicing matrix;

[0019] The feature splicing matrix is one-dimensionally stretched to obtain a time-domain feature vector.

[0020] In an example embodiment of the present disclosure, the method further comprises:

[0021] sequentially sorting the feature vectors in descending order based on the magnitudes of the feature values corresponding to the feature vectors;

[0022] multiplying the feature vectors by the feature values corresponding to the feature vectors;

[0023] performing feature concatenation on the feature vectors to be multiplied to generate a feature concatenation matrix.

[0024] In an example embodiment of the present disclosure, the frequency domain feature extraction step of the method further comprises:

[0025] performing fast Fourier transform on each lead of each segment of the electroencephalogram data of the multiple data segments after data preprocessing to generate a spectral response amplitude;

[0026] performing logarithmic operation on the spectral response amplitude;

[0027] performing feature stretching on the spectral response amplitude to generate a spectral feature vector.

[0028] In an example embodiment of the present disclosure, the method further comprises:

[0029] performing covariance calculation, covariance decomposition, feature concatenation and feature stretching processing on the spectral response amplitude to generate a frequency domain covariance feature vector.

[0030] In an example embodiment of the present disclosure, the classification and recognition step of the method further comprises:

[0031] performing classification and recognition on the time-frequency joint feature based on a preset threshold according to a preset binary classification and recognition method to complete epilepsy prediction.

[0032] In an aspect of the present disclosure, a device for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition is provided, comprising:

[0033] a data preprocessing module configured to perform data preprocessing on collected electroencephalogram data to generate multiple data segments of electroencephalogram data;

[0034] a time domain feature extraction module configured to perform covariance matrix calculation, eigenvalue decomposition, feature extraction and stretching processing on the multiple data segments of electroencephalogram data after data preprocessing to generate time domain feature vectors;

[0035] a frequency domain feature extraction module configured to perform spectral feature extraction and frequency domain covariance feature extraction on the multiple data segments of electroencephalogram data after data preprocessing to generate spectral feature vectors and frequency domain covariance feature vectors, respectively;

[0036] The feature fusion module is used to fuse the time-domain feature vector, the spectral feature vector, and the frequency-domain covariance feature vector based on a preset feature fusion strategy to generate time-frequency joint features.

[0037] The classification and recognition module classifies and recognizes the time-frequency joint features based on a preset binary classification and recognition method to complete epilepsy prediction.

[0038] In one aspect of this disclosure, an electronic device is provided, comprising:

[0039] Processor; and

[0040] A memory storing computer-readable instructions that, when executed by the processor, implement the method according to any one of the preceding claims.

[0041] In one aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the method according to any one of the preceding claims.

[0042] An exemplary embodiment of this disclosure discloses a method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition. The method includes: preprocessing acquired EEG data to generate multi-segment EEG data; decentralizing the preprocessed multi-segment EEG data, calculating the covariance matrix, performing eigenvalue decomposition, feature extraction, and stretching to generate time-domain feature vectors; extracting spectral features and frequency-domain covariance features from the preprocessed multi-segment EEG data to generate spectral feature vectors and frequency-domain covariance feature vectors, respectively; fusing the time-domain feature vectors, spectral feature vectors, and frequency-domain covariance feature vectors based on a preset feature fusion strategy to generate time-frequency joint features; and classifying and recognizing the time-frequency joint features based on a preset binary classification method to complete epilepsy prediction. This disclosure, by extracting and fusing time-frequency multi-scale feature information, can effectively shorten the manual data annotation time for medical workers and improve the annotation efficiency of epileptic seizure events, providing a new approach and method for the clinical application of epilepsy detection.

[0043] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0044] The above and other features and advantages of this disclosure will become more apparent from the detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0045] Figure 1A flow chart of a method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition is shown according to an example embodiment of the present disclosure.

[0046] Figure 2 A block diagram of a method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition is shown according to an example embodiment of the present disclosure.

[0047] Figure 3 A structural block diagram of an apparatus for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition is shown according to an example embodiment of the present disclosure.

[0048] Figure 4 A block diagram of an electronic device is schematically shown according to an example embodiment of the present disclosure.

[0049] Figure 5 A schematic diagram of a computer readable storage medium is schematically shown according to an example embodiment of the present disclosure. DETAILED DESCRIPTION

[0050] Example embodiments now will be described more fully hereinafter with reference to the accompanying drawings. Example embodiments, however, can be implemented in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of example embodiments to those skilled in the art. Like reference numerals refer to like elements throughout the several views.

[0051] Also, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the disclosure. One skilled in the relevant art will recognize, however, that the embodiments of the disclosure can be practiced without one or more of the specific details, or

[0052] The block diagrams in the drawings show functions and functionality as they can be implemented in software rather than circuitry. Described functions can be implemented using digital circuitry, as one skilled in the art will appreciate. Furthermore, the

[0053] In the present example embodiment, a method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition is first provided. Referring to Figure 1As shown in the method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition, the method can include the following steps:

[0054] In the data preprocessing step S110, the collected electroencephalogram data is preprocessed to generate electroencephalogram data in multiple data segments.

[0055] In the time domain feature extraction step S120, the electroencephalogram data in multiple data segments after data preprocessing is respectively decentered to calculate a covariance matrix, eigenvalue decomposition, feature extraction, and stretching processing to generate a time domain feature vector.

[0056] In the frequency domain feature extraction step S130, the electroencephalogram data in multiple data segments after data preprocessing is respectively subjected to spectral feature extraction and frequency domain covariance feature extraction to respectively generate a spectral feature vector and a frequency domain covariance feature vector.

[0057] In the feature fusion step S140, the time domain feature vector, the spectral feature vector, and the frequency domain covariance feature vector are fused based on a preset feature fusion strategy to generate a time-frequency joint feature.

[0058] In the classification and recognition step S150, the time-frequency joint feature is classified and recognized based on a preset binary classification and recognition method to complete epilepsy prediction.

[0059] In an exemplary embodiment of the present disclosure, a method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition is provided, wherein the method includes: preprocessing collected electroencephalogram data to generate electroencephalogram data in multiple data segments; decentering the electroencephalogram data in multiple data segments after data preprocessing to calculate a covariance matrix, eigenvalue decomposition, feature extraction, and stretching processing to generate a time domain feature vector; performing spectral feature extraction and frequency domain covariance feature extraction on the electroencephalogram data in multiple data segments after data preprocessing to respectively generate a spectral feature vector and a frequency domain covariance feature vector; fusing the time domain feature vector, the spectral feature vector, and the frequency domain covariance feature vector based on a preset feature fusion strategy to generate a time-frequency joint feature; and classifying and recognizing the time-frequency joint feature based on a preset binary classification and recognition method to complete epilepsy prediction. The present disclosure extracts and fuses time-frequency multi-scale feature information, which can effectively shorten the manual data labeling time of medical workers, improve the labeling efficiency of epilepsy seizure events, and provide a new approach and method for clinical application of epilepsy detection.

[0060] In the following, a method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition in the present exemplary embodiment will be further described.

[0061] Embodiment One:

[0062] In the present exemplary embodiment,

[0063] In the data preprocessing step S110, the collected electroencephalogram data can be preprocessed to generate electroencephalogram data of multiple data segments.

[0064] In the embodiment of the present example, the data preprocessing step of the method further comprises:

[0065] The collected electroencephalogram data is filtered, down-sampled, segmented, baseline corrected, and preprocessed to remove artifacts to generate electroencephalogram data of multiple data segments.

[0066] In the time domain feature extraction step S120, the data preprocessed electroencephalogram data of multiple data segments can be respectively decentered to calculate a covariance matrix, eigenvalue decomposition, feature extraction, and stretching processing to generate a time domain feature vector.

[0067] In the embodiment of the present example, the time domain feature extraction step of the method further comprises:

[0068] The data preprocessed electroencephalogram data of multiple data segments are respectively decentered to calculate a covariance matrix;

[0069] Eigenvalue decomposition is performed on the covariance matrix to obtain an eigenvector;

[0070] The eigenvector is spliced to generate a feature splicing matrix;

[0071] The feature splicing matrix is one-dimensionally stretched to obtain a time domain feature vector.

[0072] In the embodiment of the present example, the method further comprises:

[0073] The eigenvector is sorted in descending order based on the size of the eigenvalue corresponding to the eigenvector;

[0074] The eigenvector is multiplied by the eigenvalue corresponding to the eigenvector;

[0075] The multiplied eigenvector is spliced to generate a feature splicing matrix.

[0076] In the frequency domain feature extraction step S130, the data preprocessed electroencephalogram data of multiple data segments can be respectively subjected to spectral feature extraction and frequency domain covariance feature extraction to generate a spectral feature vector and a frequency domain covariance feature vector.

[0077] In the embodiment of the present example, the frequency domain feature extraction step of the method further comprises:

[0078] Fast Fourier transform of each segment of each lead of the data preprocessed electroencephalogram data of multiple data segments is performed to generate a spectral response amplitude;

[0079] logarithm operation is performed on the spectral response amplitude;

[0080] The spectral feature vector is generated by feature stretching on the spectral response amplitude subjected to the logarithm operation.

[0081] In the embodiment of the present example, the method further comprises:

[0082] The spectral response amplitude subjected to the logarithm operation is subjected to covariance calculation, covariance decomposition, feature splicing and feature stretching processing respectively to generate the covariance feature vector in the frequency domain.

[0083] In the feature fusion step S140, the time-domain feature vector, the spectral feature vector and the covariance feature vector in the frequency domain can be fused based on a preset feature fusion strategy to generate a time-frequency joint feature.

[0084] In the classification and recognition step S150, the time-frequency joint feature can be classified and recognized based on a preset binary classification and recognition method to complete the epilepsy prediction.

[0085] In the embodiment of the present example, the classification and recognition step of the method further comprises:

[0086] According to the preset binary classification and recognition method, the time-frequency joint feature is classified and recognized based on a preset threshold to complete the epilepsy prediction.

[0087] Embodiment two:

[0088] In the embodiment of the present example, as shown in Figure 2 The covariance matrix decomposition-based epilepsy time-frequency joint feature extraction and detection method of the present disclosure includes five steps of data preprocessing, time-domain feature extraction, frequency-domain feature extraction, feature fusion and classification and recognition. The technical process is to first perform preprocessing operations such as downsampling and filtering on the collected original electroencephalogram data (both implantable and non-implantable), then calculate the time-domain covariance matrix and the frequency-domain covariance matrix according to the correlation between the multi-channel epilepsy data, then perform matrix decomposition to extract the corresponding time-domain features and frequency-domain features, then perform time-frequency multi-scale feature fusion at the feature level, and finally use a classification and recognition method to realize the classification and recognition of the seizure event. The present invention designs a new epilepsy feature extraction method, extracts and fuses time-frequency multi-scale feature information, can effectively shorten the manual data labeling time of medical workers, improve the labeling efficiency of the seizure event, and provides a new way and a new method for the clinical application of epilepsy detection.

[0089] In the embodiment of the present example, the data preprocessing step comprises:

[0090] The main purpose of preprocessing is to reduce the influence of noise and improve the quality of EEG signal, including filtering, downsampling, data segmentation, baseline correction, artifact removal and other steps. Filtering is to remove some frequency range data by taking a filter that meets certain requirements, and to improve the signal-to-noise ratio of the data in the frequency band of interest. According to the filtered and retained frequency, it can be divided into low-pass, high-pass, band-pass and notch filter. Downsampling is to reduce the data sampling rate, which has the effect of less calculation and improving the operation speed. Generally, before data downsampling, the data will be input into a low-pass filter (at this time it is also called an anti-aliasing filter), which suppresses the high-frequency components in the signal to prevent aliasing. Data segmentation refers to the continuous EEG data being cut off according to the time of the stimulus event, and then divided into several equal-length data segments for subsequent feature analysis. Baseline correction is to subtract the baseline value of the data segment from each data point in the segmented data. Generally, the average value of the data in a certain period of time (such as 200 milliseconds (ms)) before the start of the stimulus is taken as the baseline value. Artifact removal is to remove specific artifacts using appropriate methods. For other types of physiological artifacts, they can be distinguished and removed according to other modalities of biological data. For example, eye movement-related artifacts can be identified according to electrooculogram (EOG) data; artifacts caused by head movement can be identified according to accelerometer data. In addition to this, Independent component analysis (ICA) or regression analysis can also be used to remove eye movement artifacts. Non-physiological artifacts are usually caused by external environmental interference, so the first choice is to eliminate them by changing the environmental settings, such as ensuring good contact when wearing electrodes and avoiding excessive body movement during the experiment. Secondly, appropriate data processing methods can be used, such as setting a 50 Hz notch filter to filter out power frequency interference, etc. In addition, there are methods such as re-referencing, removing bad leads or bad segments, etc.

[0091] According to the research purpose, the quality of the collected epilepsy data, and the subsequent analysis to be carried out, the specific preprocessing method can be selected. Taking the typical epilepsy electroencephalogram (EEG) database CHB-MIT as an example, this database collects scalp EEG data with a sampling rate of 256 Hertz (Hz). According to the characteristics of epilepsy and the database, the main preprocessing method adopted is data filtering, without adopting other preprocessing methods such as downsampling or upsampling, artifact removal, etc. The main process is as follows: a band-pass filter with a passband of [1 Hz, 47 Hz] is designed to filter the original data, filtering out high-frequency components and low-frequency slow wave components in the data. Then, the filtered data is divided into several data segments, which completes the preprocessing operation. For example, for a 30-second (s) multi-lead data N c is the number of leads, N p is the number of data points, and the length of the data segment to be obtained is 0.5 s, which can be divided into 60 data segments.

[0092] In the embodiment of the present example, the time domain feature extraction method step includes:

[0093] The time domain feature is the most basic feature of the epilepsy electroencephalogram data, and contains important feature information of epilepsy. The time domain feature extraction method of the present application is as follows: assuming that the multi-lead electroencephalogram data is obtained after the above-mentioned preprocessing, the electroencephalogram data of multiple data segments is obtained Where N t is the number of data segments (also known as the number of trials). For the i-th data segment The following four steps are performed:

[0094] ① The decentering processing is performed according to the lead dimension to obtain The covariance matrix is calculated as follows:

[0095]

[0096] ② The obtained covariance matrix is subjected to eigenvalue decomposition as follows:

[0097]

[0098] Where λ j is the eigenvalue of the covariance matrix, ξ j is the corresponding eigenvector, and j has a value range of [1, N c ].

[0099] ③ The obtained eigenvector is sorted in descending order according to its eigenvalue, and the eigenvalue is multiplied by the corresponding eigenvector. Finally, the obtained vector is subjected to feature splicing to obtain a matrix, as shown in the following formula:

[0100]

[0101] ④ Finally, the spliced matrix is subjected to feature stretching operation according to the last dimension to obtain a vector f1, that is, the time domain feature extraction of epilepsy is completed. It should be noted that in formula (3), the corresponding number of eigenvalues and eigenvectors can be selected according to the actual application effect and calculation cost demand, such as selecting the first three largest eigenvalues, without selecting all eigenvalues for operation.

[0102] In the embodiment of the present example, the frequency domain feature extraction method step includes:

[0103] The frequency domain feature extraction mainly includes two aspects, which are frequency spectrum feature extraction and frequency domain covariance feature extraction.

[0104] For frequency spectrum feature extraction, the i-th data segment X epoch of the electroencephalogram data X iCalculate the Fast Fourier Transform (FFT) for each lead, and then calculate the amplitude of the transformed spectral response and perform a logarithmic operation to obtain the result. Where N fp This represents the number of corresponding frequency points. Finally, P... i Perform feature stretching to obtain the spectral feature vector f2, which is then used as the spectral feature.

[0105] For frequency domain covariance feature extraction, take the P obtained during the spectral feature extraction process. i Following the "temporal feature extraction method" described above, P is sequentially processed... i The frequency domain covariance eigenvector f3 can be obtained by performing four steps: covariance calculation, covariance decomposition, feature concatenation, and feature stretching.

[0106] In this example embodiment, the feature fusion strategy steps include:

[0107] The time-domain and frequency-domain features calculated in the above steps are fused at the feature layer using a feature concatenation method, as shown in the following formula:

[0108] f all =[f1,f2,f3] (4)

[0109] This completes the extraction, processing, and fusion of time-frequency joint features.

[0110] In this example embodiment, the classification and identification method steps include:

[0111] Regarding classification and recognition methods, appropriate binary classification methods can be used depending on the classification task, such as Support Vector Machine (SVM) and Linear Discriminant Analysis (LDA). The following section will use LDA as an example to introduce the classification method and explain how to use it.

[0112] LDA is a classic supervised dimensionality reduction algorithm. The main idea is to find an optimal projection direction that makes data within the same class as close as possible after projection, while the class centers of data from different classes are as far apart as possible, i.e., minimizing intra-class variance and maximizing inter-class variance. The implementation process is as follows: Assume D = {(x1,y1),(x2,y2),…,(x...} m ,y m Let} represent a dataset containing m samples, where any sample x i It is an n-dimensional vector, meaning the feature vector has a dimension of n. It is a label for the sample. is the set of all labeled data. Define N c (c = 0, 1) is the number of samples in class c, X c (c = 0, 1) is the set of samples in class c, then the mean vector μ c ∈ R n×1 (c = 0, 1) is the set of samples in class c, then the mean vector μ

[0113]

[0114] The covariance matrix S c ∈ r n×n of class c samples is:

[0115]

[0116] The operation of sample minus the mean vector of the class it belongs to in the above formula is called decentralization. Next, define the within-class scatter matrix (Between-class scatter matrix) and the between-class scatter matrix (Between-class scatter matrix), denoted as S w ∈ R n×n and S b ∈ R n×n :

[0117]

[0118] S b = (μ0- μ1)(μ0- μ1) T (8)

[0119] Suppose the projection vector to be found is w, then any sample x i After projection in this direction, w T x i , the corresponding two class centers are converted to: w T μ0and w T μ1. For the optimization objective to make the distance between the class centers of different categories data as large as possible, it is converted to where ||a|| F represents the Frobenius norm; at the same time, the sample projection points within the same category are as close as possible, that is, the covariance of the same class sample projection points w T S0wand w T S1wshould be as small as possible, which is converted to minimizing w T S0w+w T S1w. In summary, the optimization objective function is converted to:

[0120]

[0121] From the above equation, the solution of J(w) is related to the direction of w, but not its length. Let w T S w w=1, the objective function is equivalent to:

[0122]

[0123] s.t. w T S w w=1 (11)

[0124] The equation is constructed by Lagrange multiplier method:

[0125] c(w,λ)=-w T S b w+λ(w T S w w-1) (12)

[0126] Where λ is the Lagrange multiplier, and the above equation is the partial derivative of the vector w:

[0127]

[0128] The above equation is equivalent to:

[0129] S b w=λ(μ0-μ1) (14)

[0130] Therefore, the analytical solution of the objective function is In addition, the decision threshold y t is calculated for the prediction of subsequent test samples:

[0131]

[0132] For test sample x test ∈r n×1 , the projection obtains the decision value dv=w T x test , and finally the category prediction is made according to the following equation:

[0133]

[0134] In the embodiment of the present example, the existing epilepsy detection method usually extracts epilepsy features at a single scale (such as time domain or nonlinearity), without extracting and fusing feature information contained in different scale spaces, which limits the improvement of epilepsy detection performance. The time-frequency joint feature extraction and detection method based on covariance matrix decomposition in the present disclosure extracts feature information covering multi-scale spaces in time domain and frequency domain according to the correlation of multi-channel epilepsy data, designs a feature fusion strategy to fuse multi-scale features, realizes accurate detection of seizure events, shortens the manual data labeling time of medical workers, and provides a new method for clinical application of epilepsy detection.

[0135] In the embodiment of the present example, the novel epilepsy time-frequency joint feature extraction and detection method of the present disclosure can effectively extract epilepsy feature information covering multiple scales such as time and frequency, and realize accurate identification of seizure events. Compared with the traditional method of manually labeling seizures by doctors, this method can automatically extract features representing seizures and identify them, which can shorten the data labeling time and save the time and effort of doctors, and has a wide application prospect in epilepsy detection.

[0136] It should be noted that although the steps of the method in the present disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. In addition or alternatively, some steps can be omitted, multiple steps can be combined into one step, and / or one step can be divided into multiple steps.

[0137] In addition, in the embodiment of the present example, a covariance decomposition-based epilepsy time-frequency joint feature extraction and detection device is also provided. Referring to Figure 3 As shown, the covariance decomposition-based epilepsy time-frequency joint feature extraction and detection device 300 can include a data preprocessing module 310, a time domain feature extraction module 320, a frequency domain feature extraction module 330, a feature fusion module 340, and a classification and recognition module 350. Among them:

[0138] The data preprocessing module 310 is configured to perform data preprocessing on the collected electroencephalogram data to generate electroencephalogram data of multiple data segments;

[0139] The time domain feature extraction module 320 is configured to calculate the covariance matrix, eigenvalue decomposition, feature extraction and stretching processing on the decentered multiple data segments of the electroencephalogram data after data preprocessing, respectively, to generate a time domain feature vector;

[0140] The frequency domain feature extraction module 330 is configured to perform frequency spectrum feature extraction and frequency domain covariance feature extraction on the EEG data of the plurality of data segments after data preprocessing, respectively, to generate a frequency spectrum feature vector and a frequency domain covariance feature vector.

[0141] The feature fusion module 340 is configured to fuse the time domain feature vector, the frequency spectrum feature vector, and the frequency domain covariance feature vector based on a preset feature fusion strategy to generate a time-frequency joint feature.

[0142] The classification and recognition module 350 is configured to perform classification and recognition on the time-frequency joint feature based on a preset binary classification and recognition method to complete epilepsy prediction.

[0143] The specific details of each of the above-described modules of the epilepsy time-frequency joint feature extraction and detection device based on covariance decomposition have been described in detail in the corresponding epilepsy time-frequency joint feature extraction and detection method based on covariance decomposition, and thus will not be described here again.

[0144] It should be noted that, although several modules or units of the epilepsy time-frequency joint feature extraction and detection device 300 based on covariance decomposition are mentioned in the above detailed description, such division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into a plurality of modules or units.

[0145] In addition, in the exemplary embodiments of the present disclosure, an electronic device capable of implementing the above method is also provided.

[0146] Those skilled in the art can understand that each aspect of the present disclosure can be implemented as a system, a method or a program product. Therefore, each aspect of the present disclosure can be embodied as a complete hardware embodiment, a complete software embodiment (including firmware, microcode, etc.), or an embodiment combining software and hardware aspects, which can be collectively referred to as "circuitry", "module" or "system" herein.

[0147] The electronic device 400 according to this embodiment of the present disclosure will be described below with reference to Figure 4 Figure 4 The displayed electronic device 400 is merely an example and should not impose any limitation on the functions and use range of the embodiments of the present disclosure.

[0148] As Figure 4 ​As shown, the electronic device 400 is in the form of a general computing device. Components of the electronic device 400 can include, but are not limited to, the at least one processing unit 410 described above, the at least one storage unit 420 described above, a bus 430 that connects the various system components, including the storage unit 420 and the processing unit 410, a display unit 440.

[0149] The storage unit stores program code that can be executed by the processing unit 410, such that the processing unit 410 performs the steps described in the above "Exemplary Methods" section of this specification in accordance with various exemplary embodiments of the present application. For example, the processing unit 410 can perform the steps S110 to S150 as shown in FIG. 1. Figure 1

[0150] The storage unit 420 can include a readable medium in the form of volatile storage such as a random access memory (RAM) 4201 and / or cache memory 4202, and also can include a non-volatile storage such as a read-only memory (ROM) 4203.

[0151] The storage unit 420 can further include a program / utility 4204 having a set (at least one) of program modules 4205 that include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, and each or a combination thereof can include implementation of a network environment.

[0152] The bus 430 can be representative of one or more of several types of bus structures, including a storage bus or bus controller, a peripheral bus, a graphics bus (e.g., AGP), a processor or local bus using any of a variety of bus structures, and the like.

[0153] ​The electronic device 400 can also communicate with one or more external devices 470 such as a keyboard or pointing devices, a Bluetooth device, or a disk drive. These and other peripherals can be connected to the electronic device 400 by one or more peripheral interfaces 450, such as a USB port. The electronic device 400 can also include one or more buses 430, which can allow data to be exchanged between one or more of the above-described components. The above-described components (including but not limited to the memory and / or the processor) can each include one or more devices executing instructions that can be stored in one or more computer-readable media, which can be non-volatile media. For example, the computer-readable media can include one or more of a hard disk, a floppy disk, a magnetic tape, an optical disk, a compact disc (CD), a digital versatile disc (DVD), a Blu-ray disc, a flash memory, a volatile memory, a non-volatile memory, a storage medium, a memory device, and / or any other computer-readable media.

[0154] From the above description of the embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solutions according to the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a U disk, a mobile hard disk, etc.) or a network, and includes a plurality of instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to perform the methods according to the embodiments of the present disclosure.

[0155] In the example embodiments of the present disclosure, a computer-readable storage medium having a program product stored thereon capable of implementing the above-mentioned method of the present disclosure is also provided. In some possible embodiments, various aspects of the present disclosure can also be implemented in the form of a program product, which includes program codes for causing a terminal device to perform the steps described in the above-mentioned "example method" section according to various example embodiments of the present disclosure when the program product is run on the terminal device.

[0156] Reference Figure 5 As shown, a program product 500 for implementing the above-mentioned method according to the embodiments of the present disclosure is described, which can adopt a portable compact disc read-only memory (CD-ROM) and include program codes, and can be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited to this, and in this document, the readable storage medium can be any tangible medium containing or storing a program, which can be used or combined with an instruction execution system, device or apparatus.

[0157] The program product can employ any combination of one or more computer-readable media. The computer-readable media can be a computer-readable storage medium or a computer-readable signal medium. The computer-readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include the following: an electrical connection having one or more wires, a portable disc, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0158] The computer-readable signal medium can include a computer-readable storage medium that is propagated as a carrier wave. The computer-readable signal medium can further be any computer-readable medium that is not a storage medium. The computer-readable signal medium can be a computer-readable storage medium that is a propagated signal on a computer-readable storage medium.

[0159] The program code embodied on the computer-readable media can be transmitted using any appropriate medium, including but not limited to wireless, wired, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0160] The program code can be executed by one or more programmable processors, which can be implemented in one or more computer systems. In this context, a computer system generally includes a plurality of these programmable processors, which work in concert to perform a task. Additionally, the program code can be downloaded from an external source, including the internet, through a computer network, or through a broadcast medium.

[0161] Furthermore, the above-described diagrams are merely schematic illustrations of the processes included in the method according to the exemplary embodiments of the present application, and are not intended to be limiting. It is readily understood that the processes shown in the above-described diagrams do not indicate or limit the time sequence of the processes. In addition, it is readily understood that the processes can be executed synchronously or asynchronously, for example, in a plurality of modules.

[0162] Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the features disclosed herein. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

[0163] It should be understood that the present disclosure is not limited to the precise structures herein described and illustrated in the drawings, and that various modifications and changes can be made without departing from its scope. The scope of the present disclosure is limited only by the claims that follow.

Claims

1. A method for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition, characterized in that, The method comprises: a data preprocessing step of preprocessing the collected electroencephalogram data to generate electroencephalogram data of multiple data segments; a time domain feature extraction step of calculating a covariance matrix, eigenvalue decomposition, feature extraction and stretching processing on the data-preprocessed electroencephalogram data of multiple data segments to generate a time domain feature vector; the time domain feature extraction step further comprises: calculating a covariance matrix on the data-preprocessed electroencephalogram data of multiple data segments after decentralization; performing eigenvalue decomposition on the covariance matrix to obtain a feature vector; performing feature splicing on the feature vector to generate a feature splicing matrix; performing one-dimensional stretching on the feature splicing matrix to obtain a time domain feature vector; the feature splicing matrix generated by performing feature splicing on the feature vector comprises: sorting the feature vector in descending order based on the size of the eigenvalue corresponding to the feature vector; multiplying the feature vector by the eigenvalue corresponding to the feature vector; performing feature splicing on the multiplied feature vector to generate a feature splicing matrix; a frequency domain feature extraction step of performing spectral feature extraction and frequency domain covariance feature extraction on the data-preprocessed electroencephalogram data of multiple data segments to generate a spectral feature vector and a frequency domain covariance feature vector, respectively; the frequency domain feature extraction step further comprises: performing fast Fourier transform on each segment of each lead of the data-preprocessed electroencephalogram data of multiple data segments to generate a spectral response amplitude; performing logarithmic operation on the spectral response amplitude; performing feature stretching on the logarithmically operated spectral response amplitude to generate a spectral feature vector; performing covariance calculation, covariance decomposition, feature splicing and feature stretching processing on the logarithmically operated spectral response amplitude to generate a frequency domain covariance feature vector; a feature fusion step of fusing the time domain feature vector, the spectral feature vector and the frequency domain covariance feature vector based on a preset feature fusion strategy to generate a time-frequency joint feature; a classification and recognition step of classifying and recognizing the time-frequency joint feature based on a preset binary classification and recognition method to complete epilepsy prediction.

2. The method of claim 1, wherein, The data preprocessing step of the method further comprises: performing data preprocessing of filtering, downsampling, data segmentation, baseline correction and artifact removal on the collected electroencephalogram data to generate electroencephalogram data of multiple data segments.

3. The method of claim 1, wherein, The classification and recognition step of the method further comprises: classifying and recognizing the time-frequency joint feature based on a preset threshold according to a preset binary classification and recognition method to complete epilepsy prediction.

4. A device for extracting and detecting epilepsy time-frequency joint features based on covariance decomposition, characterized in that, Based on the method of any one of claims 1-3, the device comprises: a data preprocessing module for preprocessing the collected electroencephalogram data to generate electroencephalogram data of multiple data segments; a time domain feature extraction module for calculating a covariance matrix, eigenvalue decomposition, feature extraction and stretching processing on the data-preprocessed electroencephalogram data of multiple data segments to generate a time domain feature vector; a frequency domain feature extraction module for performing spectral feature extraction and frequency domain covariance feature extraction on the data-preprocessed electroencephalogram data of multiple data segments to generate a spectral feature vector and a frequency domain covariance feature vector, respectively; The feature fusion module is configured to fuse the time domain feature vector, the frequency spectrum feature vector and the covariance feature vector in the frequency domain based on a preset feature fusion strategy to generate time-frequency joint features. The classification and recognition module is configured to perform classification and recognition on the time-frequency joint features based on a preset binary classification and recognition method to complete the epilepsy prediction.

5. An electronic device, comprising: The method comprises a processor; and a memory having computer readable instructions stored thereon, the computer readable instructions being executable by the processor to implement the method according to any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, a computer program stored thereon, the computer program being executable by a processor to implement the method according to any one of claims 1 to 3.

Images