CN115299963A - High-frequency oscillation signal automatic detection algorithm and system based on waveform characteristic template - Google Patents
High-frequency oscillation signal automatic detection algorithm and system based on waveform characteristic template Download PDFInfo
- Publication number
- CN115299963A CN115299963A CN202210907931.4A CN202210907931A CN115299963A CN 115299963 A CN115299963 A CN 115299963A CN 202210907931 A CN202210907931 A CN 202210907931A CN 115299963 A CN115299963 A CN 115299963A
- Authority
- CN
- China
- Prior art keywords
- frequency oscillation
- oscillation signal
- patient
- electroencephalogram
- waveform
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/369—Electroencephalography [EEG]
- A61B5/372—Analysis of electroencephalograms
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/40—Detecting, measuring or recording for evaluating the nervous system
- A61B5/4076—Diagnosing or monitoring particular conditions of the nervous system
- A61B5/4094—Diagnosing or monitoring seizure diseases, e.g. epilepsy
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7246—Details of waveform analysis using correlation, e.g. template matching or determination of similarity
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Signal Processing (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Veterinary Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Neurology (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Neurosurgery (AREA)
- Power Engineering (AREA)
- Psychology (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses a high-frequency oscillation signal automatic detection algorithm and a system based on a waveform characteristic template, wherein the method comprises the following steps: collecting electroencephalogram signals of a patient and setting system parameters; carrying out data preprocessing on the electroencephalogram signals of the patient according to set system parameters; detecting the preprocessed patient electroencephalogram signals according to set system parameters to obtain suspected high-frequency oscillation signal fragments, and constructing a high-frequency oscillation signal waveform template library; detecting the preprocessed patient electroencephalogram signal according to a high-frequency oscillation signal waveform template library, and outputting a high-frequency oscillation signal segment; and evaluating the effectiveness of the high-frequency oscillation signal segment. By using the method and the device, the detection efficiency of the electroencephalogram high-frequency oscillation signal can be improved, and the accuracy of the detection result of the electroencephalogram high-frequency oscillation signal can be improved. The invention is used as a high-frequency oscillation signal automatic detection algorithm and system based on the waveform characteristic template, and can be widely applied to the technical field of medical signal processing.
Description
Technical Field
The invention relates to the technical field of medical signal processing, in particular to a high-frequency oscillation signal automatic detection algorithm and system based on a waveform characteristic template.
Background
The researches show that High-Frequency oscillation signals (High Frequency oscillation signals in epileptic electroencephalograms) in epileptic electroencephalograms can be used as important biomarkers for positioning epileptic areas and also become 'gold standards' for positioning epileptic Onset zones (SOZ), the currently accepted High-Frequency oscillation signals are signals with frequencies between 80Hz and 500Hz, experts divide the signals into Rs (clips, 80 Hz) waves and FRs (fast clips, 250 Hz) waves, the traditional detection method of the High-Frequency oscillation signals is characterized in that the signals are marked and judged and identified through professional doctors 'artificial vision, the method is time-consuming and has subjective duration oscillation interference of doctors' individuals, so that false positive High-Frequency signals are difficult to avoid, most automatic High-Frequency oscillation detection algorithms compare the High-Frequency oscillation signals in Frequency bands of interest with the conventional High-Frequency oscillation signal threshold, the detection method is easy to perform High-Frequency oscillation signal detection, and the detection method is easy to perform High-Frequency oscillation signal detection and High-Frequency noise detection is easy to perform.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide a waveform feature template-based high-frequency oscillation signal automatic detection algorithm and system, which can improve the detection efficiency of the electroencephalogram high-frequency oscillation signal and improve the accuracy of the detection result of the electroencephalogram high-frequency oscillation signal.
The first technical scheme adopted by the invention is as follows: the high-frequency oscillation signal automatic detection algorithm based on the waveform characteristic template comprises the following steps of:
s1, acquiring electroencephalograms of a patient and setting system parameters;
s2, preprocessing data of the electroencephalogram signals of the patient according to set system parameters;
s3, detecting the preprocessed electroencephalogram signal of the patient according to set system parameters to obtain a suspected high-frequency oscillation signal fragment;
s4, performing feature positioning processing and storage on the suspected high-frequency oscillation signal segments based on the waveform template to form a high-frequency oscillation signal waveform template library;
and S5, detecting the preprocessed patient electroencephalogram signal according to the high-frequency oscillation signal waveform template library, and outputting a high-frequency oscillation signal segment.
Further, the setting of the system parameters in step S1 specifically includes setting a filtering frequency range, a type of the filter, and adjusting a display range of a horizontal axis and a vertical axis of the waveform display interface.
Further, the step S2 of performing data preprocessing on the patient electroencephalogram signal specifically includes the following steps:
s21, performing eye electrical noise artifact processing on the electroencephalogram signals of the patient, and separating artifact noise in the electroencephalogram signals of the patient;
s22, eliminating artifact noise in the electroencephalogram signals of the patient to obtain patient electroencephalogram signals with noise eliminated;
s23, carrying out data normalization processing on the patient electroencephalogram signals with the noise removed to obtain the patient electroencephalogram signals with a certain frequency range;
s24, performing band-pass filtering processing on the patient electroencephalogram signals with a certain frequency range, and selecting the patient electroencephalogram signals which accord with a preset frequency range;
and S25, dressing and filtering the electroencephalogram signals of the patient according with the preset frequency range to obtain the preprocessed electroencephalogram signals of the patient.
Further, the step S3 of detecting the preprocessed patient electroencephalogram signal specifically includes the following steps:
s31, carrying out feature extraction processing on the preprocessed patient electroencephalogram signal to obtain the features of the patient electroencephalogram signal;
s31, setting a characteristic threshold range, screening out the electroencephalogram signal characteristics of the patient belonging to the characteristic threshold range, and outputting a suspected high-frequency oscillation signal segment.
Further, the method comprises the step of perfecting the constructed high-frequency oscillation signal waveform template library, and the specific steps are as follows:
detecting a suspected high-frequency oscillation signal segment by a threshold method, observing whether the time-frequency domain bright spots of the signal have the island phenomenon or not by a wavelet time-frequency graph, and supplementing the detection result of the island phenomenon into a high-frequency oscillation signal waveform template library;
analyzing the high-frequency oscillation signal based on an empirical model, extracting waveform template data from an analysis result, and supplementing the waveform template data into a high-frequency oscillation signal waveform template library;
and supplementing the high-frequency oscillation signal detected by the waveform characteristic template matching module to a template library, and increasing a data set of the high-frequency oscillation signal in the high-frequency oscillation signal waveform template library.
Further, the step S5 of detecting the preprocessed patient electroencephalogram signal according to the high-frequency oscillation signal waveform template library specifically includes the following steps:
s51, setting a sliding window, a step length and a correlation coefficient threshold;
s52, circularly calculating the correlation coefficient between the waveforms of the preprocessed patient electroencephalogram signal of the sliding window and the suspected high-frequency oscillation signal segment in the high-frequency oscillation signal waveform template library through a Pearson correlation coefficient calculation formula;
and S53, taking an absolute value of the correlation coefficient, processing and judging, judging that the absolute value of the correlation coefficient is greater than a correlation coefficient threshold value, and outputting a high-frequency oscillation signal segment.
Further, the method also comprises the step of evaluating the effectiveness of the high-frequency oscillation signal segment.
Further, the evaluation of the effectiveness of the high-frequency oscillation signal segment specifically includes:
calculating the sensitivity, specificity, accuracy, real-time property and generalization of the high-frequency oscillation signal segments by corresponding index algorithms;
and (4) evaluating the effectiveness of the high-frequency oscillation signal segment by combining different calculation results.
The second technical scheme adopted by the invention is as follows: high frequency oscillation signal automatic check out system based on wave form characteristic template includes:
the user module is used for acquiring the electroencephalogram signals of the patient and setting system parameters;
the electroencephalogram signal preprocessing module is used for preprocessing data of the electroencephalogram signals of the patient according to set system parameters;
the high-frequency oscillation signal waveform template base establishing module is used for detecting the preprocessed patient electroencephalogram signal according to the set system parameters to obtain suspected high-frequency oscillation signal fragments and establishing a high-frequency oscillation signal waveform template base;
the waveform characteristic template matching module is used for detecting the preprocessed patient electroencephalogram signal according to the high-frequency oscillation signal waveform template library and outputting a high-frequency oscillation signal segment;
and the comprehensive evaluation index module is used for evaluating the effectiveness of the high-frequency oscillation signal segments.
The method and the system have the beneficial effects that: according to the invention, the electroencephalogram signal of the patient is filtered for a plurality of times, so that the interference of high-frequency noise and peaks in the electroencephalogram signal can be reduced, then a high-frequency oscillation signal waveform template base is constructed based on the preprocessed electroencephalogram signal of the patient, the preprocessed electroencephalogram signal of the patient is detected, the sensitivity of a detection algorithm can be improved, the calculation efficiency of the detection algorithm is improved, and finally, the high-frequency oscillation signal segment is evaluated through an index algorithm, so that the specificity and the accuracy of a detection result can be effectively improved.
Drawings
FIG. 1 is a flow chart of the steps of the high frequency oscillation signal automatic detection algorithm based on the waveform feature template of the present invention;
FIG. 2 is a block diagram of the structure of the electroencephalogram high-frequency oscillation signal automatic detection system based on waveform characteristics.
Detailed Description
The invention is described in further detail below with reference to the figures and the specific embodiments. The step numbers in the following embodiments are provided only for convenience of illustration, the order between the steps is not limited at all, and the execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.
Referring to fig. 1, the invention provides a high-frequency oscillation signal automatic detection algorithm based on a waveform feature template, and the method comprises the following steps:
s1, acquiring electroencephalograms of a patient and setting system parameters;
specifically, a user leads an electroencephalogram signal to be analyzed into the system through a user selection module; the filter band range (f) can also be set 1 ～f 2 ) The type of the filter and the range of the horizontal axis and the vertical axis of the waveform display interface are adjusted, the setting result of the user parameter is stored after the setting is finished and is sent to the data preprocessing module for use, and the storage module stores and manages the data.
S2, preprocessing data of the electroencephalogram signals of the patient according to set system parameters;
specifically, the method comprises an electro-ocular noise and artifact removing unit, a data normalization unit, a band-pass filtering unit and a dressing filtering unit;
s21, an electro-ocular noise and artifact removing unit separates an electroencephalogram signal and an electro-ocular signal to remove according to the idea of signal decomposition;
specifically, the ocular artifacts are the main interference noise in the electroencephalogram signal and are removed by Independent Component Analysis (ICA). The independent component analysis was developed to solve the blind source separation problem, and was first used to solve the "cocktail party" problem, which can be abstracted as follows:
x(t)＝As(t)
in the above formula, s (t) represents an independent source signal vector, x (t) represents an observed signal vector, and a represents a mixing matrix;
the objective of the independent component analysis is to solve a linear transformation ω to separate the source signal from the observed signal, which is expressed as follows:
y(t)＝ωx(t)＝ωAs(t)
in the above formula, y (t) is an estimated vector of s (t);
when the linear transformation omega is an inverse matrix of A, the source signal s (t) can be accurately calculated, namely the real electroencephalogram signal can be accurately separated;
s22, the data normalization unit is mainly used for limiting the electroencephalogram signals to be processed within a certain range after the electroencephalogram signals are processed, so that the operation speed is increased, and the subsequent processing is facilitated;
specifically, the data normalization method adopts Min-Max standardization to map the EEG signal [ l [ ] 1 ，l 2 ]In the above, the transfer function is as follows:
in the above formula, x min Represents the minimum value, x, of the lead in which it is located max Represents the maximum value of the lead in, y ij Representing the normalized brain electrical signal,/ 1 Denotes the lower range (or minimum) of the normalization, l 2 Denotes the normalized upper range (which may also be the maximum value), x ij Representing an original brain electrical signal matrix, i representing the sampling number of the original brain electrical signal, and j representing the lead number of the original brain electrical signal;
the above formula is to normalize the single lead individually, and the original EEG signal matrix is X = (X) ij ) mn (m rows and n columns, where i is less than or equal to m and j is less than or equal to n), mapping the EEG signals to [ -1,1]Performing the following steps;
s23, the band-pass filtering unit performs band-pass filtering on the normalized electroencephalogram signal according to the interested high-frequency oscillation signal frequency band, the signal frequency band range is filtered to be within the range of [80,500], the band-pass filter selects an IIR filter for filtering, and the IIR filter can further select a Butterworth filter or a Chebyshev filter;
specifically, band-pass filtering preprocessing is performed by adopting a Chebyshev II type IIR band-pass filter, the IIR filter is designed by adopting a bilinear transformation method, and a corresponding transfer function of the analog filter is designed according to the performance index of the IIR filter and expressed as follows:
in the above formula, G(s) represents the transfer function of the analog filter, N represents the total number, A k Denotes the kth gain, s denotes the s-domain, s k Represents the kth pole;
then, G(s) obtains the system function H (z) of the required digital filter by a bilinear transformation method, which is specifically as follows:
H(z)＝G(s)
in the above formula, z represents a z field, and T represents a sampling interval time;
removing the electroencephalogram signals in certain frequency ranges by using a Chebyshev II type band-pass filter, reserving the electroencephalogram signals in the interested frequency band range of 80-500 Hz, and improving the signal-to-noise ratio of the electroencephalogram signals;
and S24, removing power frequency interference influence by using a 50Hz toilet filter through the toilet filtering unit, and sending the preprocessed data to the waveform characteristic template matching module for high-frequency oscillation signal detection after the processing of the unit.
Specifically, a comb filter is used to remove commercial 50Hz power frequency multiplication interference so as to block the passing of 50Hz and its frequency multiplication signals, and the general form of the transfer function of the comb filter is as follows:
in the above formula, N represents the order of the comb filter.
S3, detecting the preprocessed patient electroencephalogram signals according to set system parameters to obtain suspected high-frequency oscillation signal fragments, and constructing a high-frequency oscillation signal waveform template library;
specifically, the method comprises the following steps of extracting the characteristics of a high-frequency oscillation signal from a preprocessed electroencephalogram signal, setting a characteristic threshold, wherein the characteristic threshold is set as the sum of the average value of the amplitude of electroencephalogram data to be analyzed after preprocessing and N times of standard deviation, the value of N is generally 3-5, selecting 3, and outputting a signal with the screening condition larger than the set characteristic threshold as a suspected high-frequency oscillation signal segment, so that the suspected high-frequency oscillation signal is positioned, and the method comprises the following steps of supplementing, optimizing and perfecting a high-frequency oscillation signal waveform template library:
the high-frequency oscillation signals can be preliminarily detected by using a threshold method, then whether the time-frequency domain bright spots have the phenomenon of 'island' or not is observed by using a wavelet time-frequency graph, if the phenomenon of 'island' exists, the phenomenon is supplemented to a template library, because false positives exist in the suspected high-frequency oscillation signals (the false positives are not the high-frequency oscillation signals actually, but the method identifies the suspected high-frequency oscillation signals as the high-frequency oscillation signals), the suspected high-frequency oscillation signals are subjected to subsequent processing, the false positives are removed, and the detection accuracy of the algorithm is improved;
directly using the positioning result of an experienced clinician, extracting waveform template data from the result lead, and supplementing the waveform template data into a template library;
and supplementing the high-frequency oscillation signal detected by the waveform characteristic template matching module to a template library, and increasing a data set of the high-frequency oscillation signal in the template library.
S4, detecting the preprocessed patient electroencephalogram signal according to a high-frequency oscillation signal waveform template library, and outputting a high-frequency oscillation signal segment;
specifically, data passing through an electroencephalogram preprocessing module is subjected to high-frequency oscillation signal template library, a sliding window is set to be 100ms, the time length of the high-frequency oscillation signal is about 100ms, the time length is set to be 100ms, the step length is 10ms, a correlation coefficient threshold Th is 0.9, the correlation is stronger when the absolute value of the correlation coefficient is larger, the absolute value of the correlation coefficient is larger than 0.9, the correlation is called as strong correlation, the correlation coefficient between the waveform data to be detected of the sliding window and the waveform of the two waveforms in the template library is calculated in a circulating mode by using a Pearson correlation coefficient, when the absolute value of the correlation coefficient between the waveform data to be detected of the sliding window and the waveform of the two waveforms in the template library is larger than the threshold, the waveform of the sliding window is a high-frequency oscillation signal, otherwise, the waveform of the non-high-frequency oscillation signal is obtained; when the loop exceeds the length of the data, the loop ends, and the pearson correlation coefficient, defined as follows:
in the above formula, ρ x，y Pearson correlation coefficient, X, representing the template waveform and the waveform to be detected i Representing template waveform data, Y i Representing the waveform data to be detected,
respectively representing the mean values, p, of the template waveform and the waveform to be detected x，y The value is between-1 and 1, the larger the absolute value is, the stronger the correlation is, and the higher the waveform similarity is;
and S5, evaluating the effectiveness of the high-frequency oscillation signal segment.
Specifically, the sensitivity, specificity, accuracy, instantaneity and generalization of the high-frequency oscillation signal segments are calculated through a corresponding index algorithm;
the sensitivity index algorithm is specifically as follows:
the specificity index algorithm is specifically as follows:
the accuracy index algorithm is specifically as follows:
in the above formula, TP represents true positive, TN represents true negative, FP represents false positive, and FN represents false negative;
furthermore, the sensitivity index, the specificity index and the accuracy index are all evaluation indexes reflecting the authenticity of detection. When the sensitivity index is higher, the false negative rate is lower, more high-frequency oscillation signal events in the real electroencephalogram signals of the epilepsy are screened out, and the omission ratio is lower. When the specificity is higher, the false positive rate is lower, namely, the high-frequency oscillation signal event in the actual non-epileptic electroencephalogram signal can be detected as the high-frequency oscillation signal event in the non-epileptic electroencephalogram signal, and the false detection rate is lower;
the TP represents the high-frequency oscillation signals in the algorithm detection result, and the expert visual markers are also the high-frequency oscillation signals in the epileptic electroencephalogram signals, namely the number of the high-frequency oscillation signals in the correctly detected epileptic electroencephalogram signals, the TN represents the high-frequency oscillation signals in the algorithm detection result, and the expert visual markers are also the high-frequency oscillation signals in the non-epileptic electroencephalogram signals, namely the number of the high-frequency oscillation signals in the correctly detected non-epileptic electroencephalogram signals, the FP represents the high-frequency oscillation signals in the algorithm detection result, and the expert visual markers are the high-frequency oscillation signals in the non-epileptic electroencephalogram signals, namely the number of the high-frequency oscillation signals in the incorrectly detected epileptic electroencephalogram signals, the FN represents the high-frequency oscillation signals in the non-epileptic electroencephalogram signals, and the expert visual markers are the high-frequency oscillation signals in the epileptic electroencephalogram signals, namely the number of the incorrectly detected non-epileptic electroencephalogram signals;
the real-time index algorithm is specifically as follows:
the time consumed by the algorithm in the specified time D30 is marked as T D30 The real-time index is time consumed by the algorithm in a specified time period, and can evaluate whether the algorithm can accurately detect the high-frequency oscillation signal in real time or not and the time consumed by detecting the high-frequency oscillation signal in the time period;
the generalization index algorithm specifically comprises multi-channel generalization intensity and single-channel generalization intensity;
the index function of the multichannel generalization strength is as follows:
GS nm ＝a*SEN m +b*SPE m +c*T D3
in the above formula, GS nm Index function, SE, representing the intensity of a multi-channel generalization m Sensitivity index value, SPE, representing the mth lead m A specific index value representing the mth lead, a, b and c representing influence factors of the sensitivity, specificity and real-time performance indexes of the algorithm, D30 representing the time consumed by the algorithm in a specified time 30s period, T D30 Representing a real-time performance index value;
the multichannel generalization strength, the algorithm is usually directed at the algorithm of single channel lead, and the multichannel generalization strength index evaluates the effectiveness after being expanded to other leads; assuming that a lead corresponding to the template characteristic is n, and a lead to be analyzed is m;
the index function of the single-channel generalization strength comprises the generalization strength of the algorithm in the inter-seizure data segment and the generalization strength of the algorithm in the epileptic seizure period;
the index function of the generalized intensity of the algorithm over the inter-episode data segment is as follows:
GS nn ＝a*SEN n +b*SPE n +c*T PD
in the above formula, GS nn Index function representing the generalization strength of the algorithm in the inter-episode data segment, SE n Sensitivity index value, SP, representing the nth lead n Indicating the value of the specificity index, T, of the nth lead PD Represents the time consumed by the algorithm in a specified time PD (period for short);
the indicator function of the generalized intensity of the algorithm during epileptic seizure is as follows:
GS nn ＝a*SE n +b*SPE n +c*T EP
in the above formula, GS nn Index function, T, representing the generalization strength of the algorithm during epileptic seizure EP Represents the time consumed by the algorithm within a specified time period, EP (for short);
the generalization strength can intuitively reflect the effectiveness of the algorithm during the detection of different events by the high frequency oscillation signal, as well as the effectiveness on different lead channels.
Referring to fig. 2, the system for automatically detecting a high-frequency oscillation signal based on a waveform feature template comprises:
the user module is used for acquiring the electroencephalogram signals of the patient and setting system parameters;
the electroencephalogram signal preprocessing module is used for preprocessing data of the electroencephalogram signals of the patient according to set system parameters;
the high-frequency oscillation signal waveform template base establishing module is used for detecting the preprocessed patient electroencephalogram signal according to the set system parameters to obtain suspected high-frequency oscillation signal fragments and establishing a high-frequency oscillation signal waveform template base;
the waveform characteristic template matching module is used for detecting the preprocessed patient electroencephalogram signal according to the high-frequency oscillation signal waveform template library and outputting a high-frequency oscillation signal segment;
and the comprehensive evaluation index module is used for evaluating the effectiveness of the high-frequency oscillation signal segments.
The contents in the above method embodiments are all applicable to the present system embodiment, the functions specifically implemented by the present system embodiment are the same as those in the above method embodiment, and the beneficial effects achieved by the present system embodiment are also the same as those achieved by the above method embodiment.
While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (9)
1. The high-frequency oscillation signal automatic detection algorithm based on the waveform characteristic template is characterized by comprising the following steps of:
s1, acquiring electroencephalogram signals of a patient and setting system parameters;
s2, preprocessing data of the electroencephalogram signals of the patient according to set system parameters;
s3, detecting the preprocessed patient electroencephalogram signals according to set system parameters to obtain suspected high-frequency oscillation signal segments;
s4, based on the waveform template, performing feature positioning processing on the suspected high-frequency oscillation signal segment and storing the suspected high-frequency oscillation signal segment to form a high-frequency oscillation signal waveform template library;
and S5, detecting the preprocessed patient electroencephalogram signal according to the high-frequency oscillation signal waveform template library, and outputting a high-frequency oscillation signal segment.
2. The waveform characteristic template-based high-frequency oscillation signal automatic detection algorithm according to claim 1, wherein the setting of the system parameters in step S1 further specifically includes setting a filtering frequency range, a type of a filter, and adjusting a display range of a horizontal axis and a vertical axis of a waveform display interface.
3. The waveform characteristic template-based high-frequency oscillation signal automatic detection algorithm of claim 2, wherein the step S2 of preprocessing the data of the patient electroencephalogram signal specifically comprises the following steps:
s21, performing ocular noise artifact processing on the electroencephalogram signal of the patient, and separating artifact noise in the electroencephalogram signal of the patient;
s22, eliminating artifact noise in the electroencephalogram signals of the patient to obtain patient electroencephalogram signals with noise eliminated;
s23, carrying out data normalization processing on the patient electroencephalogram signals with the noise removed to obtain the patient electroencephalogram signals with a certain frequency range;
s24, performing band-pass filtering processing on the patient electroencephalogram signals with a certain frequency range, and selecting the patient electroencephalogram signals which accord with a preset frequency range;
and S25, dressing and filtering the electroencephalogram signals of the patient according with the preset frequency range to obtain the preprocessed electroencephalogram signals of the patient.
4. The waveform characteristic template-based high-frequency oscillation signal automatic detection algorithm according to claim 3, wherein the step S3 of detecting the preprocessed patient electroencephalogram signal specifically comprises the following steps:
s31, carrying out feature extraction processing on the preprocessed patient electroencephalogram signal to obtain the features of the patient electroencephalogram signal;
s32, setting a characteristic threshold range, screening out the electroencephalogram signal characteristics of the patient belonging to the characteristic threshold range, and outputting suspected high-frequency oscillation signal segments.
5. The waveform characteristic template-based high-frequency oscillation signal automatic detection algorithm according to claim 4, further comprising a step of performing perfection processing on the constructed high-frequency oscillation signal waveform template library, and the method comprises the following specific steps:
detecting a suspected high-frequency oscillation signal segment by a threshold method, observing whether the time-frequency domain bright spots of the signal have the island phenomenon or not by a wavelet time-frequency graph, and supplementing the detection result of the island phenomenon into a high-frequency oscillation signal waveform template library;
analyzing the high-frequency oscillation signal based on an empirical model, extracting waveform template data according to an analysis result, and supplementing the waveform template data to a high-frequency oscillation signal waveform template library;
and supplementing the high-frequency oscillation signals detected by the waveform characteristic template matching module to a high-frequency oscillation signal waveform template library, and increasing a data set of the high-frequency oscillation signals in the high-frequency oscillation signal waveform template library.
6. The waveform characteristic template-based automatic detection algorithm for the high-frequency oscillation signals according to claim 4, wherein the step S5 of detecting the preprocessed electroencephalogram signals of the patient according to the waveform template library of the high-frequency oscillation signals specifically comprises the following steps:
s51, setting a sliding window, a step length and a correlation coefficient threshold;
s52, circularly calculating a correlation coefficient between waveforms of the preprocessed patient electroencephalogram signal of the sliding window and a suspected high-frequency oscillation signal segment in a high-frequency oscillation signal waveform template library through a Pearson correlation coefficient calculation formula;
and S53, carrying out absolute value processing on the correlation coefficient and judging, judging that the absolute value of the correlation coefficient is greater than the threshold value of the correlation coefficient, and outputting a high-frequency oscillation signal segment.
7. The waveform signature template-based HF oscillation signal automatic detection algorithm as claimed in claim 1 further includes evaluating the effectiveness of the HF oscillation signal segment.
8. The waveform feature template-based high-frequency oscillation signal automatic detection algorithm according to claim 7, wherein the evaluation of the effectiveness of the high-frequency oscillation signal segment specifically comprises:
calculating the sensitivity, specificity, accuracy, instantaneity and generalization of the high-frequency oscillation signal segments by corresponding index algorithms to obtain evaluation indexes;
and evaluating the effectiveness of the high-frequency oscillation signal segment according to the evaluation index.
9. High frequency oscillation signal automatic check out system based on wave form characteristic template, its characterized in that includes the following module:
the user module is used for acquiring the electroencephalogram signals of the patient and setting system parameters;
the electroencephalogram signal preprocessing module is used for preprocessing data of the electroencephalogram signals of the patient according to set system parameters;
the high-frequency oscillation signal waveform template base establishing module is used for detecting the preprocessed patient electroencephalogram signal according to the set system parameters to obtain suspected high-frequency oscillation signal fragments and establishing a high-frequency oscillation signal waveform template base;
the waveform characteristic template matching module is used for detecting the preprocessed patient electroencephalogram signal according to the high-frequency oscillation signal waveform template library and outputting a high-frequency oscillation signal segment;
and the comprehensive evaluation index module is used for evaluating the effectiveness of the high-frequency oscillation signal segments.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210907931.4A CN115299963A (en) | 2022-07-29 | 2022-07-29 | High-frequency oscillation signal automatic detection algorithm and system based on waveform characteristic template |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210907931.4A CN115299963A (en) | 2022-07-29 | 2022-07-29 | High-frequency oscillation signal automatic detection algorithm and system based on waveform characteristic template |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115299963A true CN115299963A (en) | 2022-11-08 |
Family
ID=83858315
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210907931.4A Pending CN115299963A (en) | 2022-07-29 | 2022-07-29 | High-frequency oscillation signal automatic detection algorithm and system based on waveform characteristic template |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115299963A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115470833A (en) * | 2022-11-14 | 2022-12-13 | 博睿康科技（常州）股份有限公司 | Signal detection method and signal detection device |
| CN117547286A (en) * | 2023-12-29 | 2024-02-13 | 中国人民解放军东部战区总医院 | Electroencephalogram signal data analysis management system based on intelligent repair material |
-
2022
- 2022-07-29 CN CN202210907931.4A patent/CN115299963A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115470833A (en) * | 2022-11-14 | 2022-12-13 | 博睿康科技（常州）股份有限公司 | Signal detection method and signal detection device |
| CN117547286A (en) * | 2023-12-29 | 2024-02-13 | 中国人民解放军东部战区总医院 | Electroencephalogram signal data analysis management system based on intelligent repair material |
| CN117547286B (en) * | 2023-12-29 | 2024-05-28 | 中国人民解放军东部战区总医院 | Electroencephalogram signal data analysis management system based on intelligent repair material |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115299963A (en) | High-frequency oscillation signal automatic detection algorithm and system based on waveform characteristic template | |
| US7809433B2 (en) | Method and system for limiting interference in electroencephalographic signals | |
| CN109907752B (en) | Electrocardiogram diagnosis and monitoring system for removing motion artifact interference and electrocardio characteristic detection | |
| CN113786204A (en) | Epilepsia intracranial electroencephalogram early warning method based on deep convolution attention network | |
| CN113066502B (en) | Heart sound segmentation positioning method based on VMD and multi-wavelet | |
| CN110123304B (en) | Dynamic electrocardio noise filtering method based on multi-template matching and correlation coefficient matrix | |
| CN114391846B (en) | Emotion recognition method and system based on filtering type feature selection | |
| CN114010207B (en) | Time domain data classification method based on zero crossing coefficient and implantable stimulation system | |
| CN112869716B (en) | Pulse feature identification system and method based on two-channel convolutional neural network | |
| CN111783715B (en) | Identity recognition method based on pulse signal feature extraction | |
| CN111067513A (en) | Sleep quality detection key brain area judgment method based on characteristic weight self-learning | |
| CN108433719B (en) | Curve driving capability evaluation method based on electrocardio and control data | |
| CN113729653A (en) | Human body pulse wave signal acquisition method | |
| CN106618486B (en) | Sleep state identification method and system in intelligent sleep assistance | |
| CN112022151B (en) | Method for processing and identifying brain electricity spike slow wave | |
| CN113591769B (en) | Non-contact heart rate detection method based on photoplethysmography | |
| CN114532994A (en) | Automatic detection method for unsupervised electroencephalogram high-frequency oscillation signals based on convolution variational self-encoder | |
| CN113143291B (en) | Electroencephalogram feature extraction method under rapid sequence visual presentation | |
| CN115067878A (en) | EEGNet-based resting state electroencephalogram consciousness disorder classification method and system | |
| Tun et al. | Analysis of computer aided identification system for ECG characteristic points | |
| CN113180705A (en) | Fatigue detection method and system based on EEG brain waves | |
| CN113208633A (en) | Emotion recognition method and system based on EEG brain waves | |
| CN111783857A (en) | Motor imagery brain-computer interface based on nonlinear network information graph | |
| CN118121176B (en) | Data analysis method and system for noninvasive heart row monitoring | |
| CN112183331B (en) | System and method for identifying electrocardiographic artifact of neonatal brain electrical signal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |