CN117828511B - Anesthesia depth electroencephalogram signal data processing method - Google Patents
Anesthesia depth electroencephalogram signal data processing method Download PDFInfo
- Publication number
- CN117828511B CN117828511B CN202410238690.8A CN202410238690A CN117828511B CN 117828511 B CN117828511 B CN 117828511B CN 202410238690 A CN202410238690 A CN 202410238690A CN 117828511 B CN117828511 B CN 117828511B
- Authority
- CN
- China
- Prior art keywords
- frequency
- target
- value
- target cluster
- electroencephalogram signal
- 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.)
- Active
Links
- 206010002091 Anaesthesia Diseases 0.000 title claims abstract description 28
- 230000037005 anaesthesia Effects 0.000 title claims abstract description 28
- 238000003672 processing method Methods 0.000 title claims description 12
- 230000002159 abnormal effect Effects 0.000 claims abstract description 55
- 230000008878 coupling Effects 0.000 claims abstract description 53
- 238000010168 coupling process Methods 0.000 claims abstract description 53
- 238000005859 coupling reaction Methods 0.000 claims abstract description 53
- 238000001228 spectrum Methods 0.000 claims abstract description 32
- 238000000034 method Methods 0.000 claims abstract description 28
- 238000001514 detection method Methods 0.000 claims abstract description 26
- 230000005856 abnormality Effects 0.000 claims abstract description 12
- 238000012545 processing Methods 0.000 claims abstract description 12
- 238000012937 correction Methods 0.000 claims description 37
- 238000010606 normalization Methods 0.000 claims description 37
- 210000004556 brain Anatomy 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 16
- 238000003064 k means clustering Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 abstract description 2
- 238000011895 specific detection Methods 0.000 abstract 1
- 230000000875 corresponding effect Effects 0.000 description 50
- 230000002902 bimodal effect Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 239000003193 general anesthetic agent Substances 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 208000012902 Nervous system disease Diseases 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000003925 brain function Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003183 myoelectrical effect Effects 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 238000013450 outlier detection Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Landscapes
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention relates to the technical field of abnormal data screening based on frequency characteristics, in particular to a method for processing electroencephalogram data at anesthesia depth, which comprises the steps of obtaining a plurality of electroencephalogram segments obtained by dividing an original electroencephalogram into frequency spectrum data, clustering the frequency spectrum data to obtain two target clusters, obtaining energy characteristics of each target cluster based on power values of each frequency in each target cluster and frequency differences of clustering center points, further obtaining coupling degree among the target clusters, correcting original abnormal scores obtained based on the integral difference condition of data points and corresponding electroencephalogram segments to obtain target abnormal scores of the data points, carrying out abnormal detection on the data points according to the target abnormal scores, reducing the influence of low-frequency signals on abnormal detection, participating actual conditions of specific detection objects into the abnormal detection, and further improving the detection accuracy of data abnormality.
Description
Technical Field
The invention relates to the technical field of abnormal data screening based on frequency characteristics, in particular to an anesthesia depth electroencephalogram data processing method.
Background
The brain electrical signal is a physiological signal for recording brain electrical activity and plays a vital role in clinical anesthesia monitoring. The brain electrical signal has very wide function, such as medical staff evaluates the anesthesia depth of the patient by monitoring the brain electrical signal, ensures that the patient is in a proper anesthesia state during the operation, thereby improving the safety of the operation, and further, for example: the brain electrical signal can also be used in the fields of researching brain functions, diagnosing nervous system diseases and the like.
However, there are often various disturbances and abnormalities in the electroencephalogram signal, such as poor contact of the detection electrode, myoelectric disturbance, electro-oculogram activity, etc., which may cause abnormal values in the electroencephalogram signal. The presence of these outliers may mislead the physician's judgment of the patient's depth of anesthesia, reducing the accuracy and reliability of the monitoring, and therefore an effective processing method is needed to improve the quality of the brain electrical signal.
The existing widely used brain electrical signal abnormality detection method comprises the following steps: determining an abnormal state of the data point to be detected according to the difference degree of the overall situation of the data point to be detected and the electroencephalogram signal in the electroencephalogram signal, for example, determining an abnormal score of the data point to be detected according to the difference degree between the data point to be detected and the average value of the data points, and then performing abnormal detection according to the abnormal score. The greater the difference degree is, the higher the abnormality score of the data point to be measured is, and the higher the abnormality degree of the data point to be measured is. However, in the abnormal value detection method, the consideration factor of the abnormal score is only the difference condition of the data point to be detected and the overall condition of the data point, and the comparison is solidified and single, so that the abnormal score is obtained inaccurately, and the accuracy of abnormal detection is affected.
Disclosure of Invention
In view of the above, the invention provides a method for processing electroencephalogram data at anesthesia depth in order to solve the technical problem of low abnormality detection accuracy of the existing electroencephalogram abnormal value detection method.
The adopted technical scheme is as follows:
an anesthesia depth electroencephalogram data processing method comprises the following steps:
Acquiring a plurality of electroencephalogram signal segments obtained by dividing an original electroencephalogram signal, and converting the electroencephalogram signal segments into frequency spectrum data;
Clustering the spectrum data to obtain two target clusters, and obtaining energy characteristics of each target cluster based on power values of each frequency in each target cluster and the difference between the power values and the frequencies of the corresponding cluster center points;
Based on the association between the local energy characteristics of the preset frequency range of each target cluster and the energy characteristics, the coupling degree between the two target clusters is obtained;
Combining the coupling degree and the energy characteristics of the two target clusters, correcting the original anomaly score obtained based on the difference condition of the data points and the whole corresponding electroencephalogram signal segments to obtain the target anomaly score of the data points;
And carrying out anomaly detection on the data points according to the target anomaly score.
Further, based on the power value of each frequency in each target cluster and the difference between the power value and the frequency of the center point of the corresponding cluster, the energy characteristic of each target cluster is obtained, and the method comprises the following steps:
For any target cluster, acquiring an original power value corresponding to each frequency value in the target cluster, and acquiring a maximum original power value in the target cluster;
According to the original power values corresponding to the frequency values, combining the difference between the frequency values and the frequency values of the clustering center points, and weighting the original power values corresponding to the frequency values to obtain weighted power values corresponding to the frequency values, wherein the weighted power values and the difference between the corresponding frequency values and the frequency values of the clustering center points are in an inverse correlation relationship;
and obtaining the energy characteristic of the target cluster according to the maximum original power value and the average value of the weighted power values corresponding to all frequency values, wherein the energy characteristic, the maximum original power value and the average value of the weighted power values are in positive correlation.
Further, the calculation formula of the energy characteristic is as follows:
；
Wherein, Representing the energy characteristics of the target cluster,/>Representing a preset parameter greater than or equal to 1,/>Representing the frequency value/>, of the target clusterThe original power value in the spectrogram corresponding to the target cluster,Representing the maximum original power value of the target cluster,/>Representing the minimum frequency value in the target cluster,/>Represents the maximum frequency value in the target cluster, N represents the number of frequency values in the target cluster,Representation pair/>Performing negative correlation normalization,/>And the cluster center point frequency value of the target cluster is represented.
Further, based on the correlation between the local energy characteristics of the preset frequency range of each target cluster and the energy characteristics, the coupling degree between the two target clusters is obtained, which comprises the following steps:
Calculating the sum value of the local energy characteristics of the preset frequency ranges of the two target clusters to obtain a first sum value, and calculating the sum value of the energy characteristics of the two target clusters to obtain a second sum value;
And taking the ratio of the first sum value to the second sum value as the coupling degree between the two target clusters.
Further, the calculation formula of the coupling degree between the two target clusters is as follows:
；
Wherein, Represents the/>Degree of coupling between two target clusters of individual electroencephalogram segments,/>Represents the/>Local energy characteristics of preset frequency range of target cluster with lower frequency in electroencephalogram signal segment,/>Represent the firstLocal energy characteristics of preset frequency range of target cluster with higher frequency in electroencephalogram signal segment,/>Represents the/>Energy characteristics of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>The energy characteristics of the target cluster with higher frequency in the electroencephalogram signal section;
；
Wherein, Represents the/>A cluster center point frequency value of a target cluster with lower frequency in the electroencephalogram signal section; /(I)Represents the/>The maximum frequency value of the target cluster with lower frequency in the electroencephalogram signal section; /(I)Represents the/>Intermediate frequency value/>, of target cluster with lower frequency in electroencephalogram signal segmentAn original power value in a spectrogram corresponding to the target cluster; /(I)Represents the/>Maximum original power value of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>Clustering center point frequency value/>, of target cluster with lower frequency in electroencephalogram signal segmentAnd maximum frequency value/>Frequency range between,/>Representation pair/>Performing negative correlation normalization,/>Representing a preset parameter greater than or equal to 1;
represents the/> A cluster center point frequency value of a target cluster with higher frequency in the electroencephalogram signal section; represents the/> The minimum frequency value of the target cluster with higher frequency in the electroencephalogram signal section; /(I)Represents the/>Intermediate frequency value/>, of target cluster with higher frequency in electroencephalogram signal segmentAn original power value in a spectrogram corresponding to the target cluster; /(I)Represents the/>Maximum original power value of target cluster with higher frequency in electroencephalogram signal segment,/>Represents the/>Minimum frequency value/>, of target cluster with higher frequency in electroencephalogram signal segmentAnd clustering center frequency value/>Frequency range between,/>Representation pair/>And carrying out negative correlation normalization.
Further, combining the coupling degree and the energy characteristics of the two target clusters, correcting the original anomaly score obtained based on the difference condition of the data points and the whole corresponding electroencephalogram signal segments to obtain the target anomaly score of the data points, including:
Obtaining the original anomaly score of each data point in the electroencephalogram signal segment based on the integral difference condition of the data point and the corresponding electroencephalogram signal segment;
Obtaining a fraction correction coefficient of the corresponding electroencephalogram signal segment based on the coupling degree and the energy characteristics of the two target clusters;
and correcting the original anomaly score according to the score correction coefficient to obtain a target anomaly score of the data point.
Further, based on the coupling degree and the energy characteristics of the two target clusters, obtaining a fraction correction coefficient of the corresponding electroencephalogram signal segment, including:
Acquiring the coupling degree of an electroencephalogram signal segment, and acquiring the energy characteristics of two target clusters corresponding to the electroencephalogram signal segment and the clustering center point frequency value of the target cluster with lower frequency in the electroencephalogram signal segment;
Calculating the absolute value of the difference value of the energy characteristics of two target clusters corresponding to the electroencephalogram signal segment, and taking the sum of the absolute value of the difference value and the energy characteristics of the target clusters with lower frequency in the electroencephalogram signal segment as the energy difference characteristics of the target clusters;
According to the coupling degree of the electroencephalogram signal segment, the energy difference characteristic and the clustering center point frequency value of the target cluster with lower frequency in the electroencephalogram signal segment are used for obtaining the fraction correction coefficient of the electroencephalogram signal segment, and the fraction correction coefficient and the coupling degree and the clustering center point frequency value of the target cluster with lower frequency in the electroencephalogram signal segment are in inverse correlation and in positive correlation with the energy difference characteristic.
Further, the calculation formula of the score correction coefficient is as follows:
；
Wherein, Representing a normalization function,/>Represents the/>Clustering center point frequency value of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>Energy characteristics of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>Energy characteristics of target cluster with higher frequency in electroencephalogram signal segment,/>Represents the/>The degree of coupling between two target clusters of the individual electroencephalogram segments; carrying out negative correlation normalization;
correspondingly, correcting the original anomaly score according to the score correction coefficient to obtain a target anomaly score of the data point, wherein the method comprises the following steps:
normalizing the original anomaly score to obtain an original anomaly score feature;
and multiplying the original abnormal score characteristics by corresponding score correction coefficients to obtain target abnormal score characteristics of the data points in the corresponding electroencephalogram signal segments.
Further, performing anomaly detection on the data points according to the target anomaly score, including:
And judging the data points with the target abnormal score characteristics being greater than or equal to a preset abnormal score characteristic threshold as abnormal data points.
Further, clustering the spectrum data to obtain two target clusters, including:
and setting the number of clustering centers K as 2, and clustering the spectrum data by adopting a K_means clustering algorithm to obtain two target clustering clusters.
The invention has at least the following beneficial effects: the original electroencephalogram signals of the patient under the anesthesia condition are divided to obtain a plurality of electroencephalogram signal segments, so that the actual conditions of the electroencephalogram signal segments can be respectively analyzed, and the accuracy of subsequent abnormality detection can be improved; converting the electroencephalogram signal segment into frequency spectrum data and clustering to obtain two target clusters, wherein the frequencies of the electroencephalogram signals are concentrated in the two frequency segments under normal conditions, so that the frequency spectrum data are clustered into the two target clusters, and the correlation and difference conditions among the target clusters can be combined to obtain accurate energy characteristics of each target cluster respectively, so that the follow-up anomaly detection according to the energy characteristics is facilitated; based on the association between the local energy characteristics and the energy characteristics of the preset frequency range of each target cluster, the coupling degree between the two target clusters is obtained, the coupling degree reflects the energy distribution of the data between the two target clusters, and the data distribution characteristics of the corresponding electroencephalogram segments can be accurately analyzed through the coupling degree; by combining the coupling degree and the energy characteristics of the two target clusters, the original anomaly score obtained based on the integral difference condition of the data points and the corresponding electroencephalogram signal segments can be corrected, the target anomaly score of the data points is obtained, the original anomaly score of the electroencephalogram signals is corrected by combining the energy distribution characteristics in the electroencephalogram signal segments with the electroencephalogram signals with different frequencies, the influence of interference factors on the anomaly score can be reduced, and therefore accurate and reliable anomaly score is obtained, and the accuracy of anomaly detection can be improved when the anomaly detection of the data points is carried out according to the corrected anomaly score.
Drawings
FIG. 1 is a flow chart of an anesthesia depth EEG signal data processing method provided by the invention;
FIG. 2 is a flow chart for obtaining fractional correction coefficients;
Fig. 3 is a flowchart of an anomaly score acquisition method for anesthesia depth electroencephalogram data processing.
Detailed Description
An embodiment of an anesthesia depth electroencephalogram data processing method is provided:
The embodiment provides a method for processing electroencephalogram data at anesthesia depth, as shown in fig. 1, which comprises the following steps:
step 1: acquiring a plurality of electroencephalogram signal segments obtained by dividing an original electroencephalogram signal, and converting the electroencephalogram signal segments into frequency spectrum data:
The application scene of the anesthesia depth electroencephalogram data processing method provided by the invention is to detect the electroencephalogram of a patient under the anesthesia condition. As a specific implementation mode, when a patient is under anesthesia, an electrode plate is correctly installed or an electrode cap is worn on the patient, then acquisition equipment is connected and zeroing treatment is carried out, and then the acquisition of the brain electrical signals of the patient, called as the original brain electrical signals, can be started, and the acquisition frequency is set as follows . As another embodiment, the acquisition frequency of the electroencephalogram signal may be set to other values according to actual needs.
Then dividing the original electroencephalogram signals to obtain a plurality of electroencephalogram signal segments, and in a specific implementation mode, segmenting the original electroencephalogram signals according to preset time lengths, wherein the lengths of the obtained electroencephalogram signal segments are the same, specific numerical values of the preset time lengths are determined by actual needs or acquisition frequencies, the preset time lengths are not too long or too short, the too short data points in the electroencephalogram signal segments are too few, the characteristics of the electroencephalogram signal segments cannot be reflected, the too long data points in the electroencephalogram signal segments are too many, and the data points with weak relevance are involved. As a specific implementation manner, if the preset duration is 1 second, the original electroencephalogram signals are segmented by taking 1 second as a unit to obtain a plurality of electroencephalogram signal segments, and the length of each electroencephalogram signal segment is 1 second.
In this embodiment, for any one electroencephalogram segment, fourier transformation is performed on the electroencephalogram segment to obtain spectral data corresponding to the electroencephalogram segment, and then the spectral data is used as a unit of time to obtain a corresponding spectral three-dimensional graph, wherein an X-axis is time, a Y-axis is frequency, and a Z-axis is power.
Step 2: clustering the spectrum data to obtain two target clusters, and obtaining the energy characteristics of each target cluster based on the power value of each frequency in each target cluster and the difference between the power value and the frequency of the center point of the corresponding cluster:
the original electroencephalogram signal can be divided into different wave bands according to the oscillation frequency: slow wave, frequency is: <1 Hz; the wave, the frequency is: 1-4 Hz or 0-4 Hz; /(I) The wave, the frequency is: 5-8 Hz; /(I)The wave, the frequency is: 9-12 Hz; /(I)The wave, the frequency is: 13-25 Hz; /(I)The wave, the frequency is: 26-80 Hz. While most anesthetic drugs are administered in slow waves// >, during maintenance of anesthesiaWave/>The waves are mainly, namely, the frequency spectrum data of the brain electrical signals are often in bimodal distribution. This results in two different major frequency variations in the brain electrical signal. Accordingly, in using existing outlier detection means, such as/>When the method detects abnormal values, the obtained statistical data is influenced by the bimodal distribution to generate distortion, the specific distortion degree is related to the bimodal distribution in the frequency spectrum data, and the calculated abnormal degree of the data is inaccurate, so that the frequency spectrum distribution condition at different moments needs to be obtained in a frequency spectrum diagram. Also, because of the differences in the human body, the characteristics of different anesthetic agents, and the different degrees of anesthesia, there is a certain difference in the specific distribution of the electroencephalogram signals, so specific analysis and treatment are required.
Clustering operation is carried out on each spectrum data, and the clustering processing modes of each spectrum data are the same. Due to the slow wave of the spectrum dataWave/>The waves are mainly, namely, the frequency spectrum data of the brain electrical signals are often in bimodal distribution. Then, for any one spectrum data, the spectrum data is clustered to obtain two target clusters, and the two target clusters are the main frequency distribution. As a specific implementation manner, the number of clustering centers K is set to be 2, and the frequency spectrum data is clustered by adopting a K_means clustering algorithm to obtain two target clustering clusters. The two target clusters have different corresponding frequency ranges, wherein the target cluster with smaller frequency is called a first target cluster, and the target cluster with larger frequency is called a second target cluster. The first target cluster is target cluster/>Its cluster center is/>; The second target cluster is the target clusterIts cluster center is/>. Wherein/>Represents the/>Individual brain electrical signal segments, i.e./>And frequency spectrum data.
And then, obtaining the energy characteristics of each target cluster based on the power value of each frequency in each target cluster and the difference between the power value and the frequency of the cluster center point. For any one target cluster, the energy characteristics of the target cluster are influenced by the power value of each frequency in the target cluster and the difference condition between each frequency and the frequency of the cluster center point in the target cluster, in addition, other influencing factors can be added according to the accuracy requirement, the energy characteristics of the target cluster reflect the energy distribution condition of the electroencephalogram signals corresponding to the target cluster, and the energy distribution condition is used for representing the electroencephalogram spectrum characteristics of a patient, so that the represented characteristics are more accurate.
As a specific implementation manner, for any one target cluster, each frequency value in the target cluster is obtained, and an original power value corresponding to each frequency value in the target cluster is obtained, wherein the original power value is an initial power value corresponding to each frequency value and not subjected to further data processing. And obtaining the maximum original power value in the original power values corresponding to the frequency values in the target cluster. And obtaining the cluster center point frequency value of the target cluster.
And weighting the original power values corresponding to the frequency values according to the original power values corresponding to the frequency values, combining the differences of the frequency values and the frequency values of the clustering center points, and obtaining weighted power values corresponding to the frequency values, wherein the weighted power values and the differences of the corresponding frequency values and the frequency values of the clustering center points are in an inverse correlation relation. The inverse relationship may be a division relationship, a subtraction relationship, or a product of the corresponding data and other data after the corresponding data is negatively correlated and normalized, etc.
And obtaining the energy characteristic of the target cluster according to the maximum original power value in the target cluster and the average value of the weighted power values corresponding to all frequency values in the target cluster, wherein the energy characteristic, the maximum original power value and the average value of the weighted power values are in positive correlation. The positive correlation relationship may be an addition relationship, a multiplication relationship, or the like. As a specific embodiment, the calculation formula of the energy characteristic is given as follows:
；
Wherein, The larger the value is, the larger the total power of the target cluster is, and the stronger the total energy of the target cluster is; />, in the present embodimentRepresenting a preset parameter greater than or equal to 1, and setting the preset parameter/>The purpose of the placement in denominator is to prevent the data in the numerator from being too large, thereby leading to the excessive calculation result and affecting the subsequent data calculation, and presetting the parameter/>The specific value of (3) is set according to the actual requirement, if the data in the molecule is not very large, the method is that/>Can be equal to 1, i.e. equivalent to nothing, in this embodiment, where/>Taking 10 as an example; /(I)Representing the frequency value/>, of the target clusterAn original power value in a spectrogram corresponding to the target cluster; /(I)Representing the maximum original power value of the target cluster; /(I)Representing the minimum frequency value in the target cluster,/>Representing the maximum frequency value in the target cluster,/>And/>Mainly represents the upper and lower frequency boundaries of the target cluster,/>And/>The range between the two is the frequency range of the target cluster; n represents the number of frequency values in the target cluster, i.e. the minimum frequency value/>, comprising the target clusterMaximum frequency value/>Minimum frequency value/>And maximum frequency value/>The total number of individual frequency values that occur in between; /(I)A cluster center point frequency value representing the target cluster; /(I)Representation pair/>And carrying out negative correlation normalization. The negative correlation normalization involved in this embodiment may use an existing negative correlation normalization manner, for example: first negative correlation normalization:
；
Or: second negative correlation normalization approach:
；
wherein e is a natural constant, x is an input quantity requiring negative correlation normalization, and y is an output quantity subjected to negative correlation normalization.
In this embodiment, in order to normalize each calculation process, all the negative correlation normalization manners in this embodiment are the first negative correlation normalization manner in the foregoing description.
Maximum original power value of the target clusterThe larger the number of (2), the stronger the energy in the target cluster; /(I)The larger the frequency difference is, the larger the difference between the frequency value and the clustering center is, and the larger the reduction amplitude of the corresponding original power value is.
And representing the average value of the weighted power values corresponding to all the frequency values, wherein the larger the average value is, the stronger the energy in the target cluster is.
Step 2, by means of the characteristic that most of anesthetic drugs can cause the spectrum data of the electroencephalogram signals to be in bimodal distribution in the anesthesia maintenance process, the electroencephalogram signal spectrum is classified into two types, the energy distribution condition of the clustering result is obtained according to the power distribution in the clustering result, the electroencephalogram spectrum characteristics of a patient are represented through the energy distribution data, and the represented characteristics are more accurate.
Step 3: based on the association between the local energy characteristics of the preset frequency range of each target cluster and the energy characteristics, the coupling degree between the two target clusters is obtained:
for any one electroencephalogram segment, the coupling degree between the two target clusters of the electroencephalogram segment is obtained based on the correlation between the local energy characteristics of the preset frequency range of the two target clusters of the electroencephalogram segment and the energy characteristics of the two target clusters of the electroencephalogram segment.
As a specific implementation manner, for any one electroencephalogram signal segment, calculating the sum value of local energy characteristics of preset frequency ranges of two target clusters of the electroencephalogram signal segment to obtain a first sum value, and calculating the sum value of energy characteristics of the two target clusters to obtain a second sum value; then, taking the ratio of the first sum value to the second sum value as the coupling degree between two target clusters of the electroencephalogram signal segment, and calculating the following formula:
；
Wherein, Represents the/>Degree of coupling between two target clusters of individual electroencephalogram segments,/>Represents the/>Local energy characteristics of a preset frequency range of a target cluster with lower frequency in the electroencephalogram signal segment, namely local energy characteristics of a preset frequency range of a first target cluster,/>Represents the/>Local energy characteristics of a preset frequency range of a target cluster with higher frequency in the electroencephalogram signal segment, namely local energy characteristics of a preset frequency range of a second target cluster,Represents the/>Energy characteristics of target clusters with lower frequency in the electroencephalogram signal segment, namely energy characteristics of first target clusters,/>, are obtainedRepresents the/>And the energy characteristic of the target cluster with higher frequency in the electroencephalogram signal section, namely the energy characteristic of the second target cluster.
In this embodiment, for a first target cluster with a lower frequency, the higher the frequency, the more the coupling degree between the first target cluster and the second target cluster can be reflected, so that the preset frequency range of the first target cluster is a frequency range between the cluster center point frequency value of the first target cluster and the maximum frequency value of the first target cluster, and for a second target cluster with a higher frequency, the lower the frequency, the more the coupling degree between the first target cluster and the second target cluster can be reflected, so that the preset frequency range of the second target cluster is a frequency range between the minimum frequency value of the second target cluster and the cluster center point frequency value of the second target cluster. Based on this, the following gives the firstLocal energy characteristic/>, of preset frequency range of first target cluster in electroencephalogram segmentAnd local energy characteristics/>, of a preset frequency range of the second target clusterIs calculated according to the formula:
；
Wherein, Represents the/>The frequency value of the clustering center point of the first target cluster in the electroencephalogram signal segment,Represents the/>Maximum frequency value of first target cluster in electroencephalogram signal segment,/>Represents the/>Frequency value/>, of clustering center point of first target cluster in electroencephalogram segmentAnd maximum frequency value/>The number of frequency values in between, i.e. comprising the cluster center frequency value/>Maximum frequency value/>Cluster center frequency value/>And maximum frequency value/>The total number of individual frequency values that occur in between; /(I)Represents the/>Frequency value/>, in first target cluster in electroencephalogram segmentAn original power value in a spectrogram corresponding to the first target cluster; /(I)Represent the firstMaximum original power value of a first target cluster in the electroencephalogram signal section; /(I)Representation pairThe specific manner of performing the negative correlation normalization can be referred to above, and will not be described in detail.
Represents the/>A cluster center point frequency value of a second target cluster in the electroencephalogram signal section; /(I)Represents the/>Minimum frequency value of second target cluster in electroencephalogram signal segment,/>Represents the/>Minimum frequency value/>, of target cluster with higher frequency in electroencephalogram signal segmentAnd clustering center frequency value/>The number of frequency values in between, i.e. comprising the minimum frequency value/>Cluster center frequency value/>Minimum frequency value/>And clustering the center point frequency valuesThe total number of individual frequency values that occur in between; /(I)Represents the/>Intermediate frequency value/>, of second target cluster in electroencephalogram segmentAn original power value in a spectrogram corresponding to the second target cluster; /(I)Represents the/>Maximum original power value of the second target cluster in the electroencephalogram signal section; /(I)Representation pair/>The specific manner of performing the negative correlation normalization can be referred to above, and will not be described in detail.
Local energy featuresRepresents the/>The energy condition of the upper half part of the first target cluster in the electroencephalogram signal section; local energy features/>Represents the/>The energy condition of the lower half part of the second target cluster in the electroencephalogram signal section; represents the/> Energy between a first target cluster and a second target cluster in the electroencephalogram signal segment; represents the/> Total energy of two types of data of the individual brain electrical signal segments. Thus, the degree of coupling/>Represents the/>The ratio relation between the energy of two types of data and the total energy of the two types of data in the electroencephalogram signal section is larger, which means that the stronger the energy distribution between the two types of data is, the closer the energy distribution between the two types of data is, and the greater the coupling degree between the two types of data is.
According to the method, the coupling phenomenon between the electroencephalogram signals with different frequencies, which can occur during electroencephalogram monitoring, is considered, the electroencephalogram signal coupling degree at different moments is obtained through the energy distribution of the electroencephalogram signals in a spectrogram, the data distribution characteristics of the electroencephalogram signals are reflected through the electroencephalogram signal coupling degree, and the data distribution characteristics are more accurately represented.
Step 4: and correcting the original anomaly score obtained based on the difference condition of the data points and the whole corresponding electroencephalogram signal segments by combining the coupling degree and the energy characteristics of the two target clusters to obtain the target anomaly score of the data points:
In the process of carrying out anomaly identification on the electroencephalogram signals, the larger the energy difference between two types of data in the electroencephalogram signals is, the more the change of the electroencephalogram signals is affected by one type of frequency, the smaller the influence of the other type of data on the electroencephalogram signals is, the more accurate the original anomaly score obtained through the existing anomaly detection mode is, and the smaller the adjustment amplitude of the original anomaly score is. The lower the frequency, the slower the change of the data, the greater the effect of the data of that portion on the original anomaly score of the electroencephalogram signal. Moreover, the larger the coupling degree between two target clusters, namely two types of electroencephalograms, the more inaccurate the distinction between the two types of electroencephalograms is, the less obvious the bimodal distribution of the electroencephalograms is, the more disordered the frequency in the electroencephalograms is, the more disordered the distribution of the electroencephalograms is, the more likely the superposition of the electroencephalograms with different real frequencies is generated, and the more easily recognized the electroencephalograms are as abnormal data.
Based on the logic, the original anomaly score of each data point in the electroencephalogram signal segment is obtained firstly based on the difference condition of the data point and the whole corresponding electroencephalogram signal segment. As a specific embodiment, the present embodiment adopts the conventional methodThe method carries out abnormal score calculation, and the calculation formula is as follows:
；
Wherein: Represents the/>, of the brain electrical signal Data point,/>Represents the/>The/>Raw anomaly scores for individual data points; /(I)Represents the/>The/>Data values for data points, i.e., voltage values; /(I)Represents the/>An average value of data values of data points in the electroencephalogram signal sections; /(I)Represents the/>Standard deviation of data values of data points in individual electroencephalogram segments.
According to the analysis, the coupling degree and the energy characteristics of the two target clusters influence the real abnormal scores of the data points, so that the score correction coefficients of the corresponding electroencephalogram segments are required to be obtained based on the coupling degree and the energy characteristics of the two target clusters. As a specific embodiment, as shown in fig. 2, a specific implementation procedure of the score correction coefficient is given as follows:
Step 4-1: and for any one electroencephalogram signal segment, according to the coupling degree of the electroencephalogram signal segment, the energy characteristics of two target clusters corresponding to the electroencephalogram signal segment and the cluster center point frequency value of the target cluster with lower frequency in the electroencephalogram signal segment are obtained.
Step 4-2: and calculating the absolute value of the difference value of the energy characteristics of the two target clusters corresponding to the electroencephalogram signal segment, and taking the sum value of the absolute value of the difference value and the energy characteristics of the target clusters with lower frequency in the electroencephalogram signal segment as the energy difference characteristics of the target clusters.
Step 4-3: and (3) obtaining the fraction correction coefficient of the electroencephalogram segment according to the coupling degree of the electroencephalogram segment, wherein the fraction correction coefficient of the electroencephalogram segment and the coupling degree of the electroencephalogram segment and the frequency value of the clustering center point of the target cluster with lower frequency in the electroencephalogram segment are in inverse correlation, and the fraction correction coefficient of the electroencephalogram segment and the frequency value of the clustering center point of the target cluster with lower frequency in the electroencephalogram segment are in positive correlation with the energy difference feature. As a specific embodiment, the firstFractional correction coefficient of individual electroencephalogram segments/>The calculation formula of (2) is as follows:
;
Wherein, Representing a normalization function,/>Represents the/>Clustering center point frequency value of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>Energy characteristics of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>Energy characteristics of target cluster with higher frequency in electroencephalogram signal segment,/>Represents the/>The degree of coupling between two target clusters of the individual electroencephalogram segments; /(I)Representation pairAnd carrying out negative correlation normalization.
As other embodiments, the first is based on satisfying the same logicFractional correction coefficient of individual electroencephalogram segments/>The calculation formula of (2) can also be as follows:
;
Representation pair/> And carrying out negative correlation normalization.
Normalization function in the present embodimentThe specific implementation of (a) can adopt the existing normalization mode, such as: first normalization:
;
Or: second normalization:
;
or: third normalization mode:
;
In order to normalize each calculation procedure in this embodiment, all normalization manners in this embodiment are the first normalization manner described above.
First, theClustering center point frequency value/>, of target cluster with lower frequency in electroencephalogram signal segmentMain frequency distribution for representing low-frequency target cluster, frequency value/>The smaller the target cluster with lower frequency is, the lower the frequency of the target cluster is, the smaller the influence of the data trend on the anomaly score is, the larger the score correction coefficient is, and the lower the original anomaly score is; /(I)Represents the/>The larger the sum value of the energy of the target cluster with lower frequency in the electroencephalogram signal section and the energy difference characteristics of the target cluster, the larger the value is, which shows that the larger the energy ratio of the low-frequency data is, and the larger the energy difference of the two target clusters is, which shows that the smaller the influence on the anomaly score is, the larger the score correction coefficient is, and the smaller the reduction degree of the original anomaly score is. Degree of coupling/>The larger the data is, the less accurate the distinction between the two target clusters is, the less obvious the bimodal distribution of the electroencephalogram signals is, the more disordered the frequency in the electroencephalogram signals is, the more disordered the distribution of the electroencephalogram signals is, the more the electroencephalogram signals are likely to be superimposed with the electroencephalogram signals with different real frequencies, the easier the data is identified into abnormal data, the smaller the score correction coefficient is, the greater the degree of lowering the original abnormal score is, and the erroneous identification is avoided.
And finally, correcting the original anomaly score according to the obtained score correction coefficient to obtain the target anomaly score of the data point. It should be understood that, since each electroencephalogram segment corresponds to a fractional correction coefficient, the fractional correction coefficient of an electroencephalogram segment is taken as the fractional correction coefficient of all data points in the electroencephalogram segment, such as: will be the firstFractional correction coefficient of individual electroencephalogram segments/>As/>Fractional correction coefficients for all data points within the individual brain electrical signal segments.
For any data point in any electroencephalogram segment, e.g. the firstThe/>, of the individual EEG signal segmentsData points, as one embodiment, first will be the/>The/>, of the individual EEG signal segmentsRaw anomaly score for individual data pointsNormalizing to obtain the original anomaly score feature/>。
Then, the original anomaly score featureMultiplying by the corresponding fractional correction factor/>Obtain the/>The/>, of the individual EEG signal segmentsTarget anomaly score feature for individual data points/>。
According to the embodiment, the original anomaly score is normalized, the target anomaly score characteristic is obtained according to the score correction coefficient, the specific value of the subsequent preset anomaly score characteristic threshold can be conveniently obtained, the subsequent anomaly detection according to the preset anomaly score characteristic threshold is facilitated, and as other embodiments, the obtained original anomaly score is directly multiplied by the score correction coefficient without normalization processing, so that the target anomaly score is obtained.
By the method, the target abnormal score characteristic of each data point in the original electroencephalogram signal can be obtained. The higher the target abnormality score feature, the higher the degree of abnormality.
And 4, correcting abnormal conditions of the electroencephalogram signals by combining energy distribution characteristics in the electroencephalogram signals with different influences of the electroencephalogram signals with different frequencies, so that the detection accuracy of abnormal data values is improved.
Step 5: performing anomaly detection on the data points according to the target anomaly score:
Presetting an anomaly score characteristic threshold, wherein the specific value of the anomaly score characteristic threshold is set according to actual needs, for example, the severity of anomaly judgment is determined, the more severe the anomaly detection is, the larger the anomaly score characteristic threshold can be set, and the specific mode is related to the normalization mode of the original anomaly score in the calculation process of the target anomaly score characteristic and the specific mode of normalization/negative correlation normalization of each parameter in the score correction coefficient, when the adopted specific modes of normalization/negative correlation normalization are different, the specific value of the preset anomaly score characteristic threshold may need to be correspondingly adjusted according to actual conditions, so that the judgment accuracy is met, for example, 0.6 is taken as an example in the embodiment.
Comparing the target abnormal score characteristic of each data point with a preset abnormal score characteristic threshold, judging the data points with the target abnormal score characteristic being greater than or equal to the preset abnormal score characteristic threshold as abnormal data points, and judging the data points with the target abnormal score characteristic being smaller than the preset abnormal score characteristic threshold as normal data points.
So far, all abnormal data points in the original electroencephalogram signal can be detected.
All abnormal data points can be removed later, interpolation processing is carried out through a least square method, and the processed electroencephalogram signals are obtained.
An embodiment of an anomaly score acquisition method for anesthesia depth electroencephalogram data processing is provided:
The existing widely used brain electrical signal abnormality detection method comprises the following steps: determining an abnormal state of the data point to be detected according to the difference degree of the overall situation of the data point to be detected and the electroencephalogram signal in the electroencephalogram signal, for example, determining an abnormal score of the data point to be detected according to the difference degree between the data point to be detected and the average value of the data points, and then performing abnormal detection according to the abnormal score. However, the consideration factor of the abnormal score is only the difference condition between the data point to be measured and the overall condition of the data point, and the comparison is solidified and single, so that the abnormal score is obtained inaccurately. In order to solve the technical problem of low accuracy in acquiring abnormal scores of data points of electroencephalogram signals in the prior art, the embodiment provides an abnormal score acquiring method for processing data of electroencephalogram signals at anesthesia depth, as shown in fig. 3, which comprises the following steps:
Acquiring a plurality of electroencephalogram signal segments obtained by dividing an original electroencephalogram signal, and converting the electroencephalogram signal segments into frequency spectrum data;
Clustering the spectrum data to obtain two target clusters, and obtaining energy characteristics of each target cluster based on power values of each frequency in each target cluster and the difference between the power values and the frequencies of the corresponding cluster center points;
Based on the association between the local energy characteristics of the preset frequency range of each target cluster and the energy characteristics, the coupling degree between the two target clusters is obtained;
And correcting the original anomaly score obtained based on the difference condition of the data points and the whole corresponding electroencephalogram signal segments by combining the coupling degree and the energy characteristics of the two target clusters to obtain the target anomaly score of the data points.
Specific implementation processes of each step in the method are described in the embodiment of the anesthesia depth electroencephalogram data processing method, and are not repeated.
The above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the application and are intended to be included within the scope of the application.
Claims (6)
1. The method for processing the brain electrical signal data at the anesthesia depth is characterized by comprising the following steps of:
Acquiring a plurality of electroencephalogram signal segments obtained by dividing an original electroencephalogram signal, and converting the electroencephalogram signal segments into frequency spectrum data;
Clustering the spectrum data to obtain two target clusters, and obtaining energy characteristics of each target cluster based on power values of each frequency in each target cluster and the difference between the power values and the frequencies of the corresponding cluster center points;
Based on the association between the local energy characteristics of the preset frequency range of each target cluster and the energy characteristics, the coupling degree between the two target clusters is obtained;
Combining the coupling degree and the energy characteristics of the two target clusters, correcting the original anomaly score obtained based on the difference condition of the data points and the whole corresponding electroencephalogram signal segments to obtain the target anomaly score of the data points;
Performing anomaly detection on the data points according to the target anomaly score;
based on the power value of each frequency in each target cluster and the difference between the power value and the frequency of the center point of the corresponding cluster, the energy characteristic of each target cluster is obtained, and the method comprises the following steps:
For any target cluster, acquiring an original power value corresponding to each frequency value in the target cluster, and acquiring a maximum original power value in the target cluster;
According to the original power values corresponding to the frequency values, combining the difference between the frequency values and the frequency values of the clustering center points, and weighting the original power values corresponding to the frequency values to obtain weighted power values corresponding to the frequency values, wherein the weighted power values and the difference between the corresponding frequency values and the frequency values of the clustering center points are in an inverse correlation relationship;
obtaining energy characteristics of the target cluster according to the maximum original power value and the average value of the weighted power values corresponding to all frequency values, wherein the energy characteristics, the maximum original power value and the average value of the weighted power values are in positive correlation;
The calculation formula of the energy characteristic is as follows:
；
Wherein, Representing the energy characteristics of the target cluster,/>Representing a preset parameter greater than or equal to 1,/>Representing the frequency value/>, of the target clusterOriginal power value in spectrogram corresponding to target cluster,/>Representing the maximum original power value of the target cluster,/>Representing the minimum frequency value in the target cluster,/>Represents the maximum frequency value in the target cluster, N represents the number of frequency values in the target cluster,/>Representation pair/>Performing negative correlation normalization,/>A cluster center point frequency value representing the target cluster;
Based on the association between the local energy characteristics of the preset frequency range of each target cluster and the energy characteristics, the coupling degree between the two target clusters is obtained, and the method comprises the following steps:
Calculating the sum value of the local energy characteristics of the preset frequency ranges of the two target clusters to obtain a first sum value, and calculating the sum value of the energy characteristics of the two target clusters to obtain a second sum value;
Taking the ratio of the first sum value to the second sum value as the coupling degree between the two target clusters;
The calculation formula of the coupling degree between the two target clusters is as follows:
；
Wherein, Represents the/>Degree of coupling between two target clusters of individual electroencephalogram segments,/>Represents the/>Local energy characteristics of preset frequency range of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>Local energy characteristics of preset frequency range of target cluster with higher frequency in electroencephalogram signal segment,/>Represents the/>Energy characteristics of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>The energy characteristics of the target cluster with higher frequency in the electroencephalogram signal section;
；
Wherein, Represents the/>A cluster center point frequency value of a target cluster with lower frequency in the electroencephalogram signal section; represents the/> The maximum frequency value of the target cluster with lower frequency in the electroencephalogram signal section; /(I)Represents the/>Intermediate frequency value/>, of target cluster with lower frequency in electroencephalogram signal segmentAn original power value in a spectrogram corresponding to the target cluster; /(I)Represents the/>Maximum original power value of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>Clustering center point frequency value/>, of target cluster with lower frequency in electroencephalogram signal segmentAnd maximum frequency value/>Frequency range between,/>Representation pair/>Performing negative correlation normalization,/>Representing a preset parameter greater than or equal to 1;
represents the/> A cluster center point frequency value of a target cluster with higher frequency in the electroencephalogram signal section; /(I)Represents the/>The minimum frequency value of the target cluster with higher frequency in the electroencephalogram signal section; /(I)Represents the/>Intermediate frequency value/>, of target cluster with higher frequency in electroencephalogram signal segmentAn original power value in a spectrogram corresponding to the target cluster; /(I)Represents the/>Maximum original power value of target cluster with higher frequency in electroencephalogram signal segment,/>Represents the/>Minimum frequency value/>, of target cluster with higher frequency in electroencephalogram signal segmentAnd clustering center frequency value/>Frequency range between,/>Representation pair/>And carrying out negative correlation normalization.
2. The anesthesia depth electroencephalogram data processing method according to claim 1, wherein correcting the original anomaly score obtained based on the difference condition of the data point and the whole of the corresponding electroencephalogram segment by combining the coupling degree and the energy characteristics of the two target clusters to obtain the target anomaly score of the data point comprises:
Obtaining the original anomaly score of each data point in the electroencephalogram signal segment based on the integral difference condition of the data point and the corresponding electroencephalogram signal segment;
Obtaining a fraction correction coefficient of the corresponding electroencephalogram signal segment based on the coupling degree and the energy characteristics of the two target clusters;
and correcting the original anomaly score according to the score correction coefficient to obtain a target anomaly score of the data point.
3. The anesthesia depth electroencephalogram data processing method according to claim 2, wherein obtaining a score correction coefficient of a corresponding electroencephalogram segment based on the coupling degree and energy characteristics of two target clusters, comprises:
Acquiring the coupling degree of an electroencephalogram signal segment, and acquiring the energy characteristics of two target clusters corresponding to the electroencephalogram signal segment and the clustering center point frequency value of the target cluster with lower frequency in the electroencephalogram signal segment;
Calculating the absolute value of the difference value of the energy characteristics of two target clusters corresponding to the electroencephalogram signal segment, and taking the sum of the absolute value of the difference value and the energy characteristics of the target clusters with lower frequency in the electroencephalogram signal segment as the energy difference characteristics of the target clusters;
According to the coupling degree of the electroencephalogram signal segment, the energy difference characteristic and the clustering center point frequency value of the target cluster with lower frequency in the electroencephalogram signal segment are used for obtaining the fraction correction coefficient of the electroencephalogram signal segment, and the fraction correction coefficient and the coupling degree and the clustering center point frequency value of the target cluster with lower frequency in the electroencephalogram signal segment are in inverse correlation and in positive correlation with the energy difference characteristic.
4. The method for processing the electroencephalogram data at the anesthesia depth according to claim 3, wherein the calculation formula of the score correction coefficient is as follows:
；
Wherein, Representing a normalization function,/>Represents the/>Clustering center point frequency value of target cluster with lower frequency in electroencephalogram signal segment,/>Represents the/>The energy characteristics of the target cluster with lower frequency in the individual electroencephalogram signal segment,Represents the/>Energy characteristics of target cluster with higher frequency in electroencephalogram signal segment,/>Represents the/>The degree of coupling between two target clusters of the individual electroencephalogram segments; carrying out negative correlation normalization;
correspondingly, correcting the original anomaly score according to the score correction coefficient to obtain a target anomaly score of the data point, wherein the method comprises the following steps:
normalizing the original anomaly score to obtain an original anomaly score feature;
and multiplying the original abnormal score characteristics by corresponding score correction coefficients to obtain target abnormal score characteristics of the data points in the corresponding electroencephalogram signal segments.
5. The method according to claim 4, wherein abnormality detection is performed on data points according to the target abnormality score, comprising:
And judging the data points with the target abnormal score characteristics being greater than or equal to a preset abnormal score characteristic threshold as abnormal data points.
6. The anesthesia depth electroencephalogram data processing method according to claim 1, wherein clustering the spectrum data to obtain two target clusters comprises:
and setting the number of clustering centers K as 2, and clustering the spectrum data by adopting a K_means clustering algorithm to obtain two target clustering clusters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410238690.8A CN117828511B (en) | 2024-03-04 | 2024-03-04 | Anesthesia depth electroencephalogram signal data processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410238690.8A CN117828511B (en) | 2024-03-04 | 2024-03-04 | Anesthesia depth electroencephalogram signal data processing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117828511A CN117828511A (en) | 2024-04-05 |
| CN117828511B true CN117828511B (en) | 2024-05-10 |
Family
ID=90504392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410238690.8A Active CN117828511B (en) | 2024-03-04 | 2024-03-04 | Anesthesia depth electroencephalogram signal data processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117828511B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117997352B (en) * | 2024-04-07 | 2024-05-31 | 中国医学科学院阜外医院 | Optimized storage method for monitoring data of anesthesia machine |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111400543A (en) * | 2020-03-20 | 2020-07-10 | 腾讯科技（深圳）有限公司 | Audio segment matching method, device, equipment and storage medium |
| CN111931129A (en) * | 2020-07-23 | 2020-11-13 | 杭州电子科技大学 | Inter-muscle coupling network analysis method based on Gaussian Copula transfer entropy |
| CN114942133A (en) * | 2022-05-20 | 2022-08-26 | 大连理工大学 | Optimal rank non-negative matrix factorization-based early fault diagnosis method for planetary gearbox |
| CN116687423A (en) * | 2023-05-29 | 2023-09-05 | 中国科学院上海微系统与信息技术研究所 | Neuron spike potential classification method and device |
| CN116898455A (en) * | 2023-07-06 | 2023-10-20 | 湖北大学 | Sleep electroencephalogram signal detection method and system based on deep learning model |
| WO2023206888A1 (en) * | 2022-04-25 | 2023-11-02 | 广东玖智科技有限公司 | Ppg signal cluster center acquisition method and apparatus, and ppg signal processing method and apparatus |
| CN117040983A (en) * | 2023-09-28 | 2023-11-10 | 联通(江苏)产业互联网有限公司 | Data sharing method and system based on big data analysis |
| CN117100241A (en) * | 2023-08-24 | 2023-11-24 | 中科心感(南京)医疗电子科技有限公司 | Heartbeat interval measurement method and device |
| CN117349220A (en) * | 2023-12-04 | 2024-01-05 | 大连致胜科技有限公司 | Data processing method and system based on PCI bus |
| CN117576823A (en) * | 2023-11-29 | 2024-02-20 | 上海徽视科技集团有限公司 | Queuing and calling system terminal |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11125653B2 (en) * | 2018-10-11 | 2021-09-21 | Palo Alto Research Center Incorporated | Motion-insensitive features for condition-based maintenance of factory robots |
-
2024
- 2024-03-04 CN CN202410238690.8A patent/CN117828511B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111400543A (en) * | 2020-03-20 | 2020-07-10 | 腾讯科技（深圳）有限公司 | Audio segment matching method, device, equipment and storage medium |
| CN111931129A (en) * | 2020-07-23 | 2020-11-13 | 杭州电子科技大学 | Inter-muscle coupling network analysis method based on Gaussian Copula transfer entropy |
| WO2023206888A1 (en) * | 2022-04-25 | 2023-11-02 | 广东玖智科技有限公司 | Ppg signal cluster center acquisition method and apparatus, and ppg signal processing method and apparatus |
| CN114942133A (en) * | 2022-05-20 | 2022-08-26 | 大连理工大学 | Optimal rank non-negative matrix factorization-based early fault diagnosis method for planetary gearbox |
| CN116687423A (en) * | 2023-05-29 | 2023-09-05 | 中国科学院上海微系统与信息技术研究所 | Neuron spike potential classification method and device |
| CN116898455A (en) * | 2023-07-06 | 2023-10-20 | 湖北大学 | Sleep electroencephalogram signal detection method and system based on deep learning model |
| CN117100241A (en) * | 2023-08-24 | 2023-11-24 | 中科心感(南京)医疗电子科技有限公司 | Heartbeat interval measurement method and device |
| CN117040983A (en) * | 2023-09-28 | 2023-11-10 | 联通(江苏)产业互联网有限公司 | Data sharing method and system based on big data analysis |
| CN117576823A (en) * | 2023-11-29 | 2024-02-20 | 上海徽视科技集团有限公司 | Queuing and calling system terminal |
| CN117349220A (en) * | 2023-12-04 | 2024-01-05 | 大连致胜科技有限公司 | Data processing method and system based on PCI bus |
Non-Patent Citations (5)
| Title |
|---|
| "Clustering Sparse Swarm Decomposition for Automated Recognition of Upper Limb Movements From Nonhomogeneous Cross-Channel EEG Signals ";Shailesh Vitthalrao Bhalerao 等;《 IEEE Sensors Letters》;20231227;第8卷(第1期);全文 * |
| "Novel feature for identification of focal EEG signals with k-Means and fuzzy c-means algorithms";Khushnandan Rai等;2015 IEEE International Conference on Digital Signal Processing (DSP);20150910;全文 * |
| "基于改进EMD与IMF能量熵的单次脑电信号分类研究";罗松;《中国优秀硕士学位论文全文数据库 (基础科学辑)》;20220315;全文 * |
| 基于改进K均值聚类及其距离修正的睡眠分期方法;于莹;王蓓;马家睿;王行愚;;计算机应用;20200710(第S1期);全文 * |
| 用于构建脑磁图网络的信号提取方法;杨春兰;吴文晓;吴水才;任洁钏;;北京工业大学学报;20200710(第07期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117828511A (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117828511B (en) | Anesthesia depth electroencephalogram signal data processing method | |
| US7734334B2 (en) | Assessment of sleep quality and sleep disordered breathing based on cardiopulmonary coupling | |
| KR101868888B1 (en) | Classification of Sleep/Wakefulness using Nasal Pressure for Patients with Sleep-disordered Breathing | |
| US9386951B2 (en) | Method and apparatus for monitoring sleep apnea severity | |
| CN113180691B (en) | Three-channel sleep apnea and hypopnea syndrome recognition device | |
| US20170188865A1 (en) | Method And System To Calculate qEEG | |
| US20220096022A1 (en) | System, method, and computer algorithm for measuring, displaying, and accurately detecting changes in electrophysiological evoked potentials | |
| CA2536516A1 (en) | Eeg seizure analysis | |
| AU2019101756A4 (en) | Method for analyzing heart rate variability signal based on extremum energy decomposition method | |
| US8862195B2 (en) | Method, system, and apparatus for automatic detection of obstructive sleep apnea from oxygen saturation recordings | |
| CN106901726A (en) | A kind of epilepsy localization system based on brain electricity high frequency brain electric oscillation rhythm and pace of moving things comentropy | |
| CN112971724A (en) | Method for detecting sleeping point | |
| CN116138745A (en) | Sleep respiration monitoring method and device integrating millimeter wave radar and blood oxygen data | |
| WO2014158921A1 (en) | Method and system to calculate qeeg | |
| Luján et al. | EEG based schizophrenia and bipolar disorder classification by means of deep learning methods | |
| Bali et al. | Performance comparison of ANN classifiers for sleep apnea detection based on ECG signal analysis using hilbert transform | |
| Li et al. | Crackles detection method based on time-frequency features analysis and SVM | |
| CN106264479A (en) | body temperature data processing method | |
| US20220151565A1 (en) | Signal quality index evaluation circuit | |
| Sofiah et al. | Calculation of Quantitative Parameters of Clinical EEG Signals by Adopting Visual Reading Methods | |
| Medina-Ibarra et al. | Assessment of Singularities in the EEG during A-phases of Sleep based on Wavelet Decomposition | |
| Fahruzi et al. | A New Approach to Predict Potential Sleep Disorder based on Fractal Analysis from Non-overlapping Single Lead ECG Using Support Vector Machine. | |
| Herrero et al. | Electrocardiogram-based detection of central sleep apnea: a full-record signal processing approach | |
| Huang et al. | Optimizing blood oxygen saturation measurement accuracy and consistency using fingertip acceleration and blood perfusion index | |
| Yang et al. | An Adaptive CUSUM Approach for Automating Sleep Apnoea Analysis Based on Pulse and Oximetry Data |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |