CN111110221A - Time-frequency-nonlinear multidimensional body surface gastric electrical feature extraction method - Google Patents
Time-frequency-nonlinear multidimensional body surface gastric electrical feature extraction method Download PDFInfo
- Publication number
- CN111110221A CN111110221A CN201911366170.0A CN201911366170A CN111110221A CN 111110221 A CN111110221 A CN 111110221A CN 201911366170 A CN201911366170 A CN 201911366170A CN 111110221 A CN111110221 A CN 111110221A
- Authority
- CN
- China
- Prior art keywords
- frequency
- body surface
- gastric
- time
- nonlinear
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/42—Detecting, measuring or recording for evaluating the gastrointestinal, the endocrine or the exocrine systems
- A61B5/4222—Evaluating particular parts, e.g. particular organs
- A61B5/4238—Evaluating particular parts, e.g. particular organs stomach
-
- 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
-
- 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
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
Abstract
The invention discloses a time-frequency-nonlinear multidimensional body surface stomach electrical characteristic extraction method, which comprises the following steps: 1) collecting body surface gastric electrical signals; 2) preprocessing the acquired body surface gastric electrical signals to obtain required gastric electrical signal data; 3) extracting time domain features from the preprocessed stomach electrical signal data; extracting frequency domain features; extracting nonlinear features; 4) performing statistical analysis on the extracted time domain characteristics, frequency domain characteristics and nonlinear characteristics, screening out optimal characteristics, and combining the optimal characteristics into a time-frequency-nonlinear multidimensional characteristic vector; 5) and classifying the time-frequency-nonlinear multidimensional feature vector by using a Support Vector Machine (SVM) classifier. The invention can extract the body surface gastric potential characteristics from multiple dimensions, and the extracted characteristic parameters can better represent the gastric potential, thereby providing a feasible technical scheme for the early diagnosis and evaluation of gastrointestinal diseases such as functional dyspepsia, irritable bowel syndrome and the like.
Description
Technical Field
The invention relates to the technical field of gastric electrology data analysis, in particular to a time-frequency-nonlinear multidimensional body surface gastric electrology feature extraction method.
Background
Body surface gastric motility (EGG) is the most ideal index for evaluating the gastrointestinal function of human body and is an important characteristic of gastrointestinal peristalsis and contraction. As a nonlinear and non-stable signal, the ultrasonic probe has the characteristics of extremely low frequency (3 times/minute) and extremely weak signal, and is easily interfered by signals such as electrocardio and respiration in the signal recording process, so that the data analysis of the body surface gastric electrical signals is a difficult point.
The characteristic extraction is an important link in data analysis, most of the existing body surface gastric electrograph analysis is based on short-time Fourier transform (STFT) to extract time-domain and frequency-domain characteristic parameters, the examination result of the body surface gastric electrograph is clinically evaluated mainly according to gastric electrograph examination and judgment standards (draft) issued by the Chinese medical society in 1999 at present, the technical conditions are limited to the time, the characteristic parameters extracted by the method are single, only comprise characteristic parameters such as main frequency, main power percentage, normal slow wave rhythm percentage and gastric tachycardia/bradyrhythmia percentage, and the whole information of body surface gastric electrograph signals cannot be comprehensively reflected. Therefore, there is a need for improvement of the existing body surface gastric electrical feature extraction method.
Disclosure of Invention
In view of the above problems, an object of the present invention is to provide a time-frequency-nonlinear multidimensional body surface stomach electrical characteristic extraction method, which is a method for extracting body surface stomach electrical characteristic parameters based on three dimensions of time domain, frequency domain and nonlinearity, and compared with a conventional method (conventional method) for extracting stomach electrical parameters based on short-time fourier transform and 1999 evaluation standard draft, the method has the advantages of strong representation and high stability, and can effectively improve the recognition rate of body surface stomach electrical signals.
In order to realize the purpose of the invention, the technical scheme adopted by the invention is as follows:
a time-frequency-nonlinear multidimensional body surface gastric electrical feature extraction method comprises the following steps:
1) collecting body surface gastric electrical signals;
2) preprocessing the acquired body surface gastric electrical signals to obtain required gastric electrical signal data;
3) extracting time domain features from the preprocessed stomach electrical signal data: integrating the gastric value, the mean, the root mean square value and the standard deviation; extracting frequency domain features: dominant frequency, average amplitude frequency and median frequency; extracting nonlinear features: sample entropy, approximate entropy, fuzzy entropy and Lyapunov exponent;
4) performing statistical analysis on the extracted time domain characteristics, frequency domain characteristics and nonlinear characteristics, screening out optimal characteristics, and combining the optimal characteristics into a time-frequency-nonlinear multidimensional characteristic vector;
5) and classifying the time-frequency-nonlinear multidimensional feature vector by using a Support Vector Machine (SVM) classifier.
In the step 1), the acquired body surface stomach electrical signals are preprocessed, the adopted method is empirical mode decomposition, data are decomposed into a plurality of modes through the method, the modes belonging to stomach electrical components are screened out, and the stomach electrical effective modes are added to obtain reconstructed stomach electrical signals.
The key points of the method of the invention are as follows: through a plurality of tests, the time-frequency-nonlinear multidimensional characteristic vector can represent the characteristic information of the gastric electrical signal more comprehensively compared with the characteristic vector extracted by the traditional method, and the accuracy of gastrointestinal disease diagnosis can be effectively improved.
The invention has the beneficial effects that: the method can extract the body surface gastric potential characteristics from multiple dimensions, and the extracted characteristic parameters can better represent the gastric potential, thereby providing a feasible technical scheme for early diagnosis and evaluation of gastrointestinal diseases such as functional dyspepsia and irritable bowel syndrome.
Drawings
FIG. 1 is a flow chart of a method of extracting multi-dimensional time-frequency-nonlinear body surface gastric electrical characteristics according to the present invention;
FIG. 2 is a flow chart of the method of EMD in the present invention.
FIG. 3 is a block diagram of SVM classification and performance evaluation in accordance with the present invention.
Detailed Description
The invention is described in further detail below with reference to the figures and specific examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when used in this specification the singular forms "a", "an" and/or "the" include "specify the presence of stated features, steps, operations, elements, or modules, components, and/or combinations thereof, unless the context clearly indicates otherwise.
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
As shown in fig. 1, a time-frequency-nonlinear multidimensional body surface gastric electrical feature extraction method according to an embodiment of the present invention includes the following steps:
1) collecting body surface stomach electric signals of patients with stomach diseases and normal people before or after meals.
2) Preprocessing the signal data obtained in step 1), wherein the adopted method is Empirical Mode Decomposition (EMD), and the decomposition steps are shown in figure 2. According to the method, data are decomposed into a plurality of modes, the modes belonging to gastric electrical components (with frequency of 1-9 times/minute) are screened out, and the gastric electrical effective modes are added to obtain a reconstructed gastric electrical signal.
3) Extracting time domain, frequency domain and nonlinear characteristics of the reconstructed stomach electrical signal obtained in the step 2).
1. The time domain features include: integrated gastric value, mean, root mean square value, standard deviation.
(1) The integrated gastric electrical value refers to the sum of the absolute values of the amplitudes of the gastric electrical signals x (t). The strength of the gastric electrical signal and the contraction condition of the smooth muscle of the stomach can be analyzed according to the integrated gastric electrical value. The calculation formula is as follows:
(2) the mean value is used to represent the mean value of the gastric electrical signal amplitude over time, xi(t) represents the gastric electrical signal over a certain period of time, and N represents the length of the signal. The mean value can be expressed by the following formula:
(3) the rms value is used to describe the amplitude magnitude of the gastric electrical signal, i.e., the signal strength, which is closely related to the contraction of the gastric smooth muscle due to non-external stimuli, and may be defined as:
(4) the standard deviation, which reflects the degree of dispersion between individuals in a data set, can be used to assess fluctuations in gastric electrical signal,
the standard deviation is the average of the gastric electrical signal over time and can be expressed as:
2. the frequency domain features include: dominant frequency, average amplitude frequency, median frequency.
(1) The dominant frequency refers to the frequency corresponding to the dominant peak in the power spectrum of gastric band signals, i.e., the frequency corresponding to the maximum power. The invention considers the frequency within the range of 0.017-0.15 Hz as the stomach electrical signal dominant frequency.
(2) The average amplitude frequency refers to the frequency corresponding to the average amplitude value, and the characteristic parameter is suitable for frequency-amplitude spectrum analysis.
(3) The median frequency is the frequency corresponding to the division of the frequency spectrum into two regions of equal amplitude, i.e. half of the total power signature, which decreases with increasing excitation time of the smooth muscle of the stomach. PjIs the power spectrum of the gastric band signal corresponding to frequency segment j, N is the length of the frequency segment, and the median frequency can be expressed as:
3. the non-linear characteristics include: sample entropy, approximate entropy, fuzzy entropy, lyapunov exponent.
(1) Sample entropy: for a time series u (1), u (2) of length L,., u (L) or u (j):1 ≦ j ≦ L), an m-dimensional scalar time series is first constructed:
xm(i)＝{u(i+k):0≤k≤m-1},1≤i≤L-m+1 (6)
by mixing xm(i) The chebyshev distance of (chebyshev) is compared with a predetermined value of n while excluding self-comparison and finding a matching template. Next, a variable Y satisfying the above criteria is constructedi：
Wherein the content of the first and second substances,
d|xm(i)-xm(j)|＝max{|u(i+k)-u(j+k)|:0≤k≤m-1} (8)
will Ym(n) is defined as:
z can be obtained by repeating the above process after increasing the dimension m by 1 to m +1m(n)：
Wherein the content of the first and second substances,
the sample entropy can be calculated by the following formula:
wherein, Ym(n) is the probability that two sequences match m points with a similarity tolerance of n, Zm(n) is the probability that two sequences match m +1 points.
(2) Approximate entropy: let the data sequence containing M points be a subsequence of Y (1), Y (2), Y (3),.. times, Y (M), where Y (i) · Y (i), Y (i +1), Y (i +2),. times, Y (i + M-1) ], where i has a value range of 1 ≦ i ≦ M-M, and M represents the number of samples used for prediction.
The noise filter level s is:
s＝k×SD(k＝0,0.1,0.2,0.3,......,0.9) (13)
where SD is the standard deviation of the array Y.
Y (j) represents a group of subsequences resulting from the variation of the variable j of y (j) from 1 to n, and each subsequence y (j) in the set y (j) is compared with y (i), in the process, the result is
And
two parameters, wherein:
definition Km(s) and Km+1(s) are respectively the following formulas:
therefore, the approximate entropy can be expressed by the following formula:
(3) fuzzy entropy: carrying out phase space reconstruction on u (i) according to the sequence by using a sample time sequence { u (i) < 1 > i < L >, wherein the length of the sample time sequence is L, and the u (i) is greater than or equal to 1 < L >, so as to obtain a group of t-dimensional vectors (t < L-2), wherein the reconstruction vectors are as follows:
Yi t＝{u(i),u(i+1),....,u(i+t-1}-u0(i)(i＝1,2,....,L-t+1) (19)
wherein u is0(i) As an average, the following formula can be used:
for a certain vector Yi tIs a reaction of Yi tAnd
the distance between
Defined as the maximum absolute difference between the corresponding scalar components.
According to fuzzy functions
Calculating Yi tAnd
similarity between them
Wherein the fuzzy function
For an exponential function, m and r are the gradient and width of the exponential function, respectively.
Defining functions
Comprises the following steps:
according to the above calculation process, a set of t + 1-dimensional vectors can be reconstructed.
Finally, the fuzzy entropy parameter of the sequence is defined as
And
the negative natural logarithm of the deviation.
Where t and r are the phase space dimension and the similarity tolerance dimension, respectively.
(4) Lyapunov index: assuming that the M-point scalar time series is represented by an embedding dimension n and a delay time t, the phase space can be reconstructed as follows:
Y(t)＝(y(t),y(t+τ),....,y(t+(n-1)τ))(t＝1,2,....,N) (26)
where N ═ M- (N-1) τ, τ is the step size.
If the distance to the nearest neighbor is defined as follows:
where T is the period of the spectral data (T1 when there is no periodicity),
is a hypothetical point other than t that,
is a hypothetical curve parallel to Y (t). The lyapunov index may be defined as follows:
wherein i is the calibration index of the different Lyapunov indices lambda (i),
y (t) is a point in phase space,
is the point of minimum distance to the chosen reference point y (t), and λ (i) represents the overall characteristic of the data sequence, whose value may be positive, negative, or zero.
4) Feature screening and feature fusion. And screening out optimal characteristic parameters by a statistical method, and performing characteristic fusion to obtain a time-frequency-nonlinear multidimensional characteristic vector.
5) The time-frequency-nonlinear multidimensional feature vectors of the patients with the gastric diseases and normal persons are used as the input of a classifier, in this embodiment, the adopted classifier is an SVM, and as shown in FIG. 3, the kernel function type is a radial basis kernel function, so that the classification of the body surface gastric electrical features of the patients with the gastric diseases is realized.
The method used in the embodiment of the present invention is not limited unless specifically described, and any method may be used as long as the above process is completed. Those skilled in the art will appreciate that the drawings are merely schematic representations of a preferred embodiment and are not intended to represent the merits of the embodiment.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (2)
1. A time-frequency-nonlinear multidimensional body surface stomach electrical feature extraction method, which is characterized by comprising the following steps:
1) collecting body surface gastric electrical signals;
2) preprocessing the acquired body surface gastric electrical signals to obtain required gastric electrical signal data;
3) extracting time domain features from the preprocessed stomach electrical signal data: integrating the gastric value, the mean, the root mean square value and the standard deviation; extracting frequency domain features: dominant frequency, average amplitude frequency and median frequency; extracting nonlinear features: sample entropy, approximate entropy, fuzzy entropy and Lyapunov exponent;
4) performing statistical analysis on the extracted time domain characteristics, frequency domain characteristics and nonlinear characteristics, screening out optimal characteristics, and combining the optimal characteristics into a time-frequency-nonlinear multidimensional characteristic vector;
5) and classifying the time-frequency-nonlinear multidimensional feature vectors by using a support vector machine classifier.
2. The method for extracting gastric electrical characteristics of a time-frequency-nonlinear multidimensional body surface according to claim 1, wherein in step 1), the acquired gastric electrical signals of the body surface are preprocessed by empirical mode decomposition, data are decomposed into a plurality of modes by the empirical mode decomposition, modes belonging to gastric electrical components are screened out, and the gastric electrical effective modes are added to obtain reconstructed gastric electrical signals.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911366170.0A CN111110221A (en) | 2019-12-26 | 2019-12-26 | Time-frequency-nonlinear multidimensional body surface gastric electrical feature extraction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911366170.0A CN111110221A (en) | 2019-12-26 | 2019-12-26 | Time-frequency-nonlinear multidimensional body surface gastric electrical feature extraction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111110221A true CN111110221A (en) | 2020-05-08 |
Family
ID=70503105
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911366170.0A Pending CN111110221A (en) | 2019-12-26 | 2019-12-26 | Time-frequency-nonlinear multidimensional body surface gastric electrical feature extraction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111110221A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030166996A1 (en) * | 2002-01-11 | 2003-09-04 | Samsung Electronics Co., Ltd. | Method and apparatus for measuring animal's condition by acquiring and analyzing its biological signals |
| WO2009063463A2 (en) * | 2007-11-14 | 2009-05-22 | Medasense Biometrics Ltd | Pain monitoring using multidimensional analysis of physiological signals |
| CN107713988A (en) * | 2017-10-10 | 2018-02-23 | 天津大学 | A kind of obese degree detection means based on the extraction of stomach electrical feature |
| CN109492546A (en) * | 2018-10-24 | 2019-03-19 | 广东工业大学 | A kind of bio signal feature extracting method merging wavelet packet and mutual information |
| CN109674468A (en) * | 2019-01-30 | 2019-04-26 | 大连理工大学 | It is a kind of singly to lead brain electricity sleep mode automatically method by stages |
| CN109953755A (en) * | 2019-03-15 | 2019-07-02 | 度特斯(大连)实业有限公司 | A kind of extracting method and device of electrocardial vector data characteristics |
| CN110279410A (en) * | 2019-06-26 | 2019-09-27 | 天津大学 | Portable minisize is wireless stomach electricity collector |
-
2019
- 2019-12-26 CN CN201911366170.0A patent/CN111110221A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030166996A1 (en) * | 2002-01-11 | 2003-09-04 | Samsung Electronics Co., Ltd. | Method and apparatus for measuring animal's condition by acquiring and analyzing its biological signals |
| WO2009063463A2 (en) * | 2007-11-14 | 2009-05-22 | Medasense Biometrics Ltd | Pain monitoring using multidimensional analysis of physiological signals |
| CN107713988A (en) * | 2017-10-10 | 2018-02-23 | 天津大学 | A kind of obese degree detection means based on the extraction of stomach electrical feature |
| CN109492546A (en) * | 2018-10-24 | 2019-03-19 | 广东工业大学 | A kind of bio signal feature extracting method merging wavelet packet and mutual information |
| CN109674468A (en) * | 2019-01-30 | 2019-04-26 | 大连理工大学 | It is a kind of singly to lead brain electricity sleep mode automatically method by stages |
| CN109953755A (en) * | 2019-03-15 | 2019-07-02 | 度特斯(大连)实业有限公司 | A kind of extracting method and device of electrocardial vector data characteristics |
| CN110279410A (en) * | 2019-06-26 | 2019-09-27 | 天津大学 | Portable minisize is wireless stomach electricity collector |
Non-Patent Citations (8)
| Title |
|---|
| M. CURILEM ET AL.: "Comparison of artificial neural networks an support vector machines for feature selection in electrogastrography signal processing", 《2010 ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY》 * |
| 何峰,李刚,周振理,于向阳,张居元: "基于独立成分分析的胃电信号检测", 《仪器仪表学报》 * |
| 姜爱萍: "《混沌时间序列的小波神经网络预测方法及其优化研究》", 31 December 2013 * |
| 李海石: "胃电磁信号测量及信号特征提取方法研究", 《中国优秀硕士学位论文全文数据库》 * |
| 杨孝敬: "《基于ACT-R与fMRI融合的情绪与认知计算的信息加工过程研究》", 30 November 2017 * |
| 潘虹，张飞，黄时锋: "《基于局域均值分解的水力机组故障特征提取方法研究》", 31 December 2018 * |
| 石峰，莫忠息: "《信息论基础》", 30 September 2014 * |
| 陆阿明，张秋霞: "《人体运动动作测量与分析实践指导》", 31 July 2017 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109934089B (en) | Automatic multi-stage epilepsia electroencephalogram signal identification method based on supervised gradient raiser | |
| Gupta et al. | A critical review of feature extraction techniques for ECG signal analysis | |
| Zhu et al. | Epileptic seizure detection in EEGs signals using a fast weighted horizontal visibility algorithm | |
| CN104102915B (en) | Personal identification method based on ECG multi-template matching under a kind of anomalous ecg state | |
| Sharmila et al. | Wavelet-based feature extraction for classification of epileptic seizure EEG signal | |
| Choi | Detection of valvular heart disorders using wavelet packet decomposition and support vector machine | |
| CN111310570B (en) | Electroencephalogram signal emotion recognition method and system based on VMD and WPD | |
| Nishad et al. | Classification of epileptic electroencephalogram signals using tunable-Q wavelet transform based filter-bank | |
| Yuan et al. | A novel wavelet-based model for eeg epileptic seizure detection using multi-context learning | |
| CN109948396B (en) | Heart beat classification method, heart beat classification device and electronic equipment | |
| Güler et al. | A modified mixture of experts network structure for ECG beats classification with diverse features | |
| Taran et al. | Detection of sleep apnea events using electroencephalogram signals | |
| CN111027488A (en) | Signal classification method and device | |
| Sharathappriyaa et al. | Auto-encoder based automated epilepsy diagnosis | |
| Jadhav et al. | Automated sleep stage scoring using time-frequency spectra convolution neural network | |
| Pant et al. | Sleep apnea detection using electrocardiogram signal input to fawt and optimize ensemble classifier | |
| CN116049735A (en) | Uterine electromyographic signal classification method based on multivariate empirical mode decomposition | |
| Liu et al. | Automatic sleep staging algorithm based on random forest and hidden markov model | |
| CN114578963A (en) | Electroencephalogram identity recognition method based on feature visualization and multi-mode fusion | |
| CN113011330B (en) | Electroencephalogram signal classification method based on multi-scale neural network and cavity convolution | |
| Lucena et al. | The performance of short-term heart rate variability in the detection of congestive heart failure | |
| Singhal et al. | GSMD-SRST: Group Sparse Mode Decomposition and Superlet Transform Based Technique for Multi-Level Classification of Cardiac Arrhythmia | |
| Turan et al. | Empirical mode decomposition of throat microphone recordings for intake classification | |
| CN117017297A (en) | Method for establishing prediction and identification model of driver fatigue and application thereof | |
| Zhao | Skin cancer classification based on convolutional neural networks and vision transformers |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200508 |