CN112914545A - Non-contact intestinal disease diagnostic equipment - Google Patents
Non-contact intestinal disease diagnostic equipment Download PDFInfo
- Publication number
- CN112914545A CN112914545A CN201911233821.9A CN201911233821A CN112914545A CN 112914545 A CN112914545 A CN 112914545A CN 201911233821 A CN201911233821 A CN 201911233821A CN 112914545 A CN112914545 A CN 112914545A
- Authority
- CN
- China
- Prior art keywords
- intestinal
- signals
- module
- machine learning
- extracting
- 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
- 208000028774 intestinal disease Diseases 0.000 title claims abstract description 66
- 230000008991 intestinal motility Effects 0.000 claims abstract description 50
- 238000010801 machine learning Methods 0.000 claims abstract description 39
- 238000013507 mapping Methods 0.000 claims abstract description 29
- 230000000968 intestinal effect Effects 0.000 claims abstract description 21
- 238000000926 separation method Methods 0.000 claims abstract description 18
- 238000000605 extraction Methods 0.000 claims abstract description 17
- 238000000034 method Methods 0.000 claims abstract description 17
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 238000003745 diagnosis Methods 0.000 claims abstract description 12
- 238000012545 processing Methods 0.000 claims abstract description 11
- 238000012544 monitoring process Methods 0.000 claims abstract description 8
- 238000012549 training Methods 0.000 claims description 26
- 238000005070 sampling Methods 0.000 claims description 20
- 230000033001 locomotion Effects 0.000 claims description 17
- 238000001228 spectrum Methods 0.000 claims description 16
- 230000002496 gastric effect Effects 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 210000001035 gastrointestinal tract Anatomy 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 10
- 230000029058 respiratory gaseous exchange Effects 0.000 claims description 10
- 238000001914 filtration Methods 0.000 claims description 9
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 8
- 238000005311 autocorrelation function Methods 0.000 claims description 6
- 201000010099 disease Diseases 0.000 claims description 6
- 230000030135 gastric motility Effects 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 210000000936 intestine Anatomy 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 5
- 238000003066 decision tree Methods 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 4
- 238000009825 accumulation Methods 0.000 claims description 2
- 230000002763 arrhythmic effect Effects 0.000 claims description 2
- 230000033764 rhythmic process Effects 0.000 claims description 2
- 230000009286 beneficial effect Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 11
- 210000002784 stomach Anatomy 0.000 description 7
- 238000001514 detection method Methods 0.000 description 4
- RAICQOARWDPXIZ-UHFFFAOYSA-N [Gd].[Cr] Chemical compound [Gd].[Cr] RAICQOARWDPXIZ-UHFFFAOYSA-N 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 235000013305 food Nutrition 0.000 description 3
- 238000002595 magnetic resonance imaging Methods 0.000 description 3
- 208000018522 Gastrointestinal disease Diseases 0.000 description 2
- 230000005176 gastrointestinal motility Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000005481 NMR spectroscopy Methods 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000035565 breathing frequency Effects 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 239000002872 contrast media Substances 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000001121 heart beat frequency Effects 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 235000021056 liquid food Nutrition 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007659 motor function Effects 0.000 description 1
- 238000012633 nuclear imaging Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000000241 respiratory effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/05—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves
- A61B5/055—Detecting, measuring or recording for diagnosis by means of electric currents or magnetic fields; Measuring using microwaves or radio waves involving electronic [EMR] or nuclear [NMR] magnetic resonance, e.g. magnetic resonance imaging
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0033—Features or image-related aspects of imaging apparatus classified in A61B5/00, e.g. for MRI, optical tomography or impedance tomography apparatus; arrangements of imaging apparatus in a room
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Engineering & Computer Science (AREA)
- Animal Behavior & Ethology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Veterinary Medicine (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Radiology & Medical Imaging (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- High Energy & Nuclear Physics (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The invention discloses a non-contact intestinal disease diagnosis device, comprising: the data acquisition module is used for acquiring radar echo signals by using a non-contact vital sign monitoring radar; the preprocessing module is used for processing the radar echo signal by adopting a band-pass filter; the signal separation module is used for processing the preprocessed signals by adopting a signal separation algorithm to separate intestinal motility signals; the characteristic extraction module is used for extracting the characteristics of the separated intestinal motility signals to obtain characteristic parameters which are beneficial to judging different intestinal conditions, and the machine learning module is used for determining a mapping model between the intestinal conditions and the characteristic parameters by utilizing a machine learning algorithm; and the intestinal disease judgment module judges the intestinal diseases according to the mapping model. The method is effective and feasible, has reliable performance, and can accurately judge the intestinal diseases.
Description
Technical Field
The invention belongs to the field of medical equipment, and particularly relates to non-contact intestinal disease diagnosis equipment.
Background
With the acceleration of the pace of daily life and the increase of work and study pressure of people, gastrointestinal diseases become common diseases and high morbidity of modern society gradually, and the importance of evaluating gastrointestinal motility is more and more concerned by people. In the diagnosis of gastrointestinal diseases, patients who have developed organic lesions can be examined by means of radiology, nuclear imaging, nuclear magnetic resonance, electrogastrogram, and the like.
The radionuclide imaging method is based on the principle that a radionuclide-labeled drug is mixed with common food, and as the motion process of the radionuclide in the stomach is consistent with that of the food, a gamma camera is used for continuously photographing in an examination area, the gastrointestinal motility is evaluated according to the amount of the radionuclide in the stomach, a dynamic functional image of the stomach and the intestine can be obtained, a time-radioactivity curve of the stomach and the intestine is obtained through computer processing, and the half-emptying time of the stomach and the emptying rate of different times of the stomach and the intestine are calculated. Radionuclide imaging has disadvantages in that patients receive small doses of radiation and are expensive.
The Magnetic Resonance Imaging (MRI) uses gadolinium-chromium complex as a contrast agent, multi-layer cross section scanning is carried out by the MRI after the gadolinium-chromium complex is taken in, so that a main body image can be displayed, and a gastrointestinal main body image displayed by the MRI is changed in a series manner along with the gadolinium-chromium complex and food are discharged from the stomach together, so that a gastrointestinal emptying result is obtained. However, the examination must be performed in a supine position, which is different from the physiological standing position, and there is a certain error in the detection of liquid food using gravity as the main power, and the examination cost is high.
Electrogastrogram the mucosal electrodes and body surface electrodes were placed through an endoscope while the gastrointestinal electrical activity was recorded. The electrogastrogram is simple to operate, and can carry out long-time non-invasive tracing on the gastrointestinal electrical activity, but the gastrointestinal electrical activity only reflects the gastrointestinal contraction frequency and does not directly reflect the gastrointestinal motor function.
From the above, the prior art has the following disadvantages: 1) the intestinal motion condition and the intestinal disease cannot be accurately judged, and 2) most detection equipment is expensive and has certain harm to human bodies.
Disclosure of Invention
The purpose of the present invention is to provide a noncontact diagnostic apparatus for intestinal diseases, which is capable of determining intestinal diseases.
The technical solution for realizing the purpose of the invention is as follows: a non-contact intestinal disease diagnosis apparatus comprising:
the non-contact vital sign monitoring radar is used for acquiring radar echo signals;
the preprocessing module is used for preprocessing the acquired radar echo signals, acquiring human body vital sign signals, and then filtering interference to obtain preprocessed signals, wherein the acquired human body vital sign signals comprise heartbeat signals, respiration signals, gastric motility signals and intestinal motility signals;
the signal separation module is used for processing the preprocessed signals by adopting a signal separation algorithm to separate intestinal motility signals of the human body;
the characteristic extraction module is used for extracting the characteristics of the separated intestinal motility signals and extracting the characteristic parameters of the intestinal movement;
the machine learning module is used for inputting the extracted characteristic parameters and the corresponding labels into machine learning and determining a mapping model between the intestinal diseases and the characteristic parameters by a machine learning method;
and the intestinal disease judgment module judges the intestinal diseases according to the mapping model.
Compared with the prior art, the invention has the following remarkable advantages: 1) the invention can extract the characteristic parameters which are beneficial to judging the intestinal motion condition; 2) the invention selects proper characteristic parameters to represent the intestinal motion condition, and trains a model capable of distinguishing different intestinal diseases by a machine learning method. Therefore, the intestinal disease can be judged through the characteristic parameters of the intestinal motility signals; 3) the non-contact measurement is more convenient and is easy to be accepted by patients; 4) the invention is simple and effective, the equipment is simple and easy to realize, the cost is low, the operation is easy, and the performance is reliable.
Drawings
Fig. 1 is a flowchart of the non-contact intestinal disease diagnosis apparatus according to the present invention.
Fig. 2 is a diagram of a radar echo signal obtained from the acquired data in the present invention.
Fig. 3 is a signal diagram of a radar echo signal after band-pass filtering in the present invention.
Fig. 4 is a time domain waveform diagram obtained after signal separation algorithm processing in the present invention, wherein, diagram (a) is a respiration signal waveform diagram, diagram (b) is a heartbeat signal waveform diagram, diagram (c) is an intestinal movement signal waveform diagram, and diagram (d) is a gastric movement signal waveform diagram.
Detailed Description
Referring to fig. 1, a non-contact diagnosis apparatus for intestinal diseases of the present invention includes:
a data acquisition module, which acquires radar echo signals by using a non-contact vital sign monitoring radar, as shown in fig. 2;
the preprocessing module is used for preprocessing the radar echo signal x (t), acquiring a human body vital sign signal, filtering interference, and acquiring a preprocessed signal f (t), as shown in fig. 3, wherein the acquired human body vital sign signal comprises a heartbeat signal, a respiration signal, a gastric motility signal and an intestinal motility signal, and specifically comprises:
the radar echo signal x (t) is subjected to band-pass filtering, and three band-pass filters h are introduced1(n)、h2(n) and h3(n) wherein h1(n) the frequency range is 0.025-0.12 Hz, which includes the common intestinal motility and gastric motility frequency bands; h is2(n) the frequency range is 0.2-0.4 Hz, and the frequency range is a basic breathing frequency band; h is3(n) the frequency range is 1-1.33 Hz, and the frequency range is a basic heartbeat frequency band; will pass through a band-pass filter h1(n)、h2(n) and h3(n) adding the filtered signals to obtain a preprocessed signal f (t);
the signal separation module is used for processing the preprocessed signals by adopting a signal separation algorithm to separate intestinal motility signals, gastric motility signals, respiration signals and heartbeat signals of the human body, and the signal separation module is shown in fig. 4; the method comprises the following specific steps:
step 3-1, performing fast Fourier transform on the preprocessed signal f (t) to obtain a frequency spectrum of the preprocessed signal
Omega is frequency;
step 3-2, the frequency spectrum of the preprocessed signal
Performing spectrum detection, sequencing and normalizing local maximum values in the detected spectrum in a descending manner to obtain a maximum value sequence, wherein the maximum value sequence is
Step 3-3, in maximum value sequence
In (1), setting a threshold value Mmin+α(Mmax-Mmin) Wherein M isminIs the minimum value in the maximum sequence, MmaxIs the maximum value in the maximum value sequence, alpha is a constant, the number which is larger than the threshold value in the maximum value sequence is reserved, and the number M of the maximum value sequences is obtained at the momentNA selected number of modes N-M for separating respiration, heartbeat, intestinal movement and gastric movement signalsN+D，D≥4；
Step 3-4, after determining the number of modes N, carrying out frequency spectrum treatment on the preprocessed signal
Performing spectrum detection, and limiting the spectrum range to [0, pi ] according to Shannon's theorem]Dividing the frequency spectrum into N adjacent regions to generate N +1 boundary lines omeganN is the nth, N is more than or equal to 0 and less than or equal to N, and a transition interval T is defined around each boundary linenWidth of 2 τn；
Step 3-5, determining a segmentation interval ^n＝[ωn-1,ωn]Adding wavelet window coefficients
Obtaining a wavelet empirical scale function
Empirical wavelet
Step 3-6, determining detail coefficient
In the formula F-1Representing an inverse fourier transform;
step 3-7, determining approximation coefficient
In the formula F-1Representing an inverse fourier transform;
step 3-8, decomposing the preprocessed signal f (t) obtained in the step 2 into N inherent modal components through empirical wavelet transform, and expressing the modal function of each frequency scale decomposed by the preprocessed signal f (t) as
And 3-9, performing descending order arrangement on the N modal time domain functions obtained by the empirical wavelet transform algorithm according to the energy size, and selecting the first four modes with the maximum energy, namely a respiratory signal, a heartbeat signal, a gastric motility signal and an intestinal motility signal.
The characteristic extraction module is used for carrying out characteristic extraction on the separated intestinal motility signals and extracting characteristic parameters of intestinal movement, and comprises:
(1) average frequency: reflecting the speed of the intestinal motility signal;
(2) average amplitude value: reflecting the amplitude of the waveform;
(3) percentage of intestinal rhythm disorder: percentage of time of occurrence of arrhythmic signals;
(4) wave reaction area: reflecting the magnitude of the signal power;
(5) waveform main frequency: the frequency corresponding to the highest amplitude in the spectrum;
(6) coefficient of variation of dominant frequency: the ratio of the standard deviation of the average dominant frequency to the average dominant frequency is used for evaluating the stability of the dominant frequency of the intestinal motility signal;
(7) coefficient of variation of main power: the ratio of the standard deviation of the average main power to the average main power is used for evaluating the stability of the main power of the intestinal tract signal;
(8) percentage of main power: the percentage of main power to total power;
(9) intestinal motility cycle IC: the difference between the first peak point of the autocorrelation function curve of the intestinal motility signal power spectrum and the abscissa value of the origin in a certain time period;
(10) intestinal motility times IPF: adding one to the number of peak points of an autocorrelation function curve of the power spectrum of the intestinal motility signal in a certain time period,
IPFi＝Npeak+1
in the formula NpeakThe number of peak points of an autocorrelation function curve of the intestinal motility signal power spectrum;
(11) intestinal motility energy difference accumulation: accumulating the difference values of the kinetic energy of all the two adjacent intestines within a certain time period;
the machine learning module inputs the extracted characteristic parameters and the corresponding labels into machine learning, and determines a mapping model between the intestinal diseases and the characteristic parameters by a machine learning method, specifically comprising the following steps:
(1) randomly dividing characteristic data of the intestinal motility signals into two mutually exclusive sets, wherein one set is a training set S, and the other set is a cross check set CV, and classifying the data;
(2) training set S { (F)1,L1),(F2,L2),…(Fi,Li) Inputting the feature matrix of {0,1,2,3, …, N } into the machine learning method, wherein Fi＝(Fi1,Fi2,Fi3,…Fi11) Respectively extracting 11 intestinal motility signal characteristics from the characteristic extraction module; l isiThe method comprises the following steps of (1) obtaining a category label corresponding to an ith sample, wherein the label 0 is normal intestinal tract, the label N represents intestinal tract disease N, and N is more than or equal to 1 and less than or equal to N;
(3) randomly extracting sp1 samples from the training set, then performing feature sampling on sp1 samples, extracting the same sp2 features from each sample to obtain a sampling set, repeating for T times, and totally obtaining T sampling sets. Training a classification-based learner and a secondary learner of the intestinal diseases for each sampling set, and forming a mapping Model1 by using a voting method through the T classification-based learners and the secondary learner;
(4) and further performing parameter adjustment and optimization on the mapping Model1 between the intestinal diseases and the characteristic parameters by using a cross test set CV to obtain a mapping Model 2.
The intestinal disease classification base learner trained in the machine learning module is a multi-class classification SVM, and specifically comprises the following steps:
(1) setting samples containing z categories in each sampling set, dividing the z categories for v times, wherein v is larger than z, dividing one part of categories into positive categories and marking as (+1) by random division each time, dividing the other part of categories into negative categories and marking as (-1), and taking the positive and negative categories corresponding to each group as a training set; generating v training sets in total;
(2) training the SVM respectively by the generated v training sets to obtain v SVM learners, inputting data of each class into the v SVM learners to obtain corresponding codes to form a z x v code matrix;
(3) and predicting the test samples by utilizing v classifiers respectively, forming a code by the prediction marks, comparing the code with the respective code of each class, and returning the class with the minimum distance as a final prediction result, wherein the v SVM classifiers form a multi-class classified SVM learner.
The intestinal disease classification secondary learner trained in the machine learning module is a decision tree, and specifically comprises the following steps:
(1) traversing each feature, and calculating the Keyny value of the feature, wherein the formula is as follows:
wherein, the data set a is a sampling matrix of sp1 × sp2, Gini (a) represents the Keyni index of the data set a, r represents that r samples exist in the data set, and r representsiRepresenting the proportion of the ith type sample in the data set a;
wherein A is 11 intestinal motility signal characteristics obtained from the characteristic extraction module, GiniA(a) The expression is a Gini index divided according to the characteristic A, k represents that the data set a is divided into k parts under the characteristic A, ajRepresenting the j-th data set after the division;
ΔGini(A)＝Gini(a)-GiniA(a)
wherein Δ gini (a) represents the gain value of the kini index before and after the division according to the characteristic a; selecting the characteristic with the maximum gain value of the Gini index as a splitting condition of the node to generate a new node;
(2) and repeating the process by taking the generated node as a root node, selecting new intestinal motility signal characteristics as splitting conditions, and completely dividing the data a.
The parameter adjustment optimization in the machine learning module specifically comprises the following steps:
(1) determining parameters of a classification base learner and a secondary learner of the intestinal diseases, wherein the parameters of a decision tree comprise a maximum characteristic number max _ feature, a maximum depth max _ depth and a minimum sample number min _ samples _ split required by splitting, which are involved in judgment during splitting of a decision number; the parameters of the multi-class classification SVM are a punishment parameter c and a kernel function parameter gamma of the SVM;
(2) performing linear search on each parameter, training a base learner and a secondary learner by utilizing a CV set, and calculating the gradient between the model performance and each parameter;
(3) selecting a parameter with the maximum gradient value as a current tuning parameter according to the performance of the model and the gradient between the parameters, and tuning the current tuning parameter by using a gradient descent method;
(4) repeating the two steps until the gradient of all the parameters is less than a threshold value, which is a positive number close to 0;
(5) adjusting the number T1 of the classified bases of the intestinal diseases and the number T2 of the secondary learners, which is specifically as follows: and respectively calculating gradients between the model performance and the number T1 of the base learners and the number T2 of the secondary learners, and respectively optimizing T1 and T2 by using a gradient descent method to obtain the final number T1 of the intestinal disease classification base learners and the final number T2 of the secondary learners.
And the intestinal disease judgment module judges the intestinal diseases according to the mapping model.
A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the functions of the following modules in a non-contact intestinal disease diagnosis device when executing the computer program:
the preprocessing module is used for preprocessing radar echo signals acquired by the non-contact vital sign monitoring radar to acquire human vital sign signals, and then filtering interference to obtain preprocessed signals, wherein the acquired human vital sign signals comprise heartbeat signals, respiration signals, gastric dynamic signals and intestinal dynamic signals;
the signal separation module is used for processing the preprocessed signals by adopting a signal separation algorithm to separate intestinal motility signals of the human body;
the characteristic extraction module is used for extracting the characteristics of the separated intestinal motility signals and extracting the characteristic parameters of the intestinal movement;
the machine learning module is used for inputting the extracted characteristic parameters and the corresponding labels into machine learning and determining a mapping model between the intestinal diseases and the characteristic parameters by a machine learning method;
and the intestinal disease judgment module judges the intestinal diseases according to the mapping model.
A computer-readable storage medium on which a computer program is stored, which, when being executed by a processor, implements functions of the following modules in a noncontact enteric disease diagnosis apparatus:
the preprocessing module is used for preprocessing radar echo signals acquired by the non-contact vital sign monitoring radar to acquire human vital sign signals, and then filtering interference to obtain preprocessed signals, wherein the acquired human vital sign signals comprise heartbeat signals, respiration signals, gastric dynamic signals and intestinal dynamic signals;
the signal separation module is used for processing the preprocessed signals by adopting a signal separation algorithm to separate intestinal motility signals of the human body;
the characteristic extraction module is used for extracting the characteristics of the separated intestinal motility signals and extracting the characteristic parameters of the intestinal movement;
the machine learning module is used for inputting the extracted characteristic parameters and the corresponding labels into machine learning and determining a mapping model between the intestinal diseases and the characteristic parameters by a machine learning method;
and the intestinal disease judgment module judges the intestinal diseases according to the mapping model.
The device can extract characteristic parameters beneficial to judging the intestinal tract movement condition; the device selects proper characteristic parameters to represent the intestinal motion condition, and trains a model capable of distinguishing different intestinal diseases by a machine learning method. Therefore, the intestinal disease can be judged through the characteristic parameters of the intestinal motility signals.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A non-contact diagnosis apparatus for intestinal diseases, comprising:
the non-contact vital sign monitoring radar is used for acquiring radar echo signals;
the preprocessing module is used for preprocessing the acquired radar echo signals, acquiring human body vital sign signals, and then filtering interference to obtain preprocessed signals, wherein the acquired human body vital sign signals comprise heartbeat signals, respiration signals, gastric motility signals and intestinal motility signals;
the signal separation module is used for processing the preprocessed signals by adopting a signal separation algorithm to separate intestinal motility signals of the human body;
the characteristic extraction module is used for extracting the characteristics of the separated intestinal motility signals and extracting the characteristic parameters of the intestinal movement;
the machine learning module is used for inputting the extracted characteristic parameters and the corresponding labels into machine learning and determining a mapping model between the intestinal diseases and the characteristic parameters by a machine learning method;
and the intestinal disease judgment module judges the intestinal diseases according to the mapping model.
2. The apparatus according to claim 1, wherein the feature extraction module performs feature extraction on the separated intestinal motility signal of the human body, and the extracting of the feature parameters of the intestinal motility specifically includes:
(1) average frequency: reflecting the speed of the intestinal motility signal;
(2) average amplitude value: reflecting the amplitude of the waveform;
(3) percentage of intestinal rhythm disorder: percentage of time of occurrence of arrhythmic signals;
(4) wave reaction area: reflecting the magnitude of the signal power;
(5) waveform main frequency: the frequency corresponding to the highest amplitude in the spectrum;
(6) coefficient of variation of dominant frequency: the ratio of the standard deviation of the average dominant frequency to the average dominant frequency is used for evaluating the stability of the dominant frequency of the intestinal motility signal;
(7) coefficient of variation of main power: the ratio of the standard deviation of the average main power to the average main power is used for evaluating the stability of the main power of the intestinal tract signal;
(8) percentage of main power: the percentage of main power to total power;
(9) intestinal motility cycle IC: the difference between the first peak point of the autocorrelation function curve of the intestinal motility signal power spectrum and the abscissa value of the origin in a certain time period;
(10) intestinal motility times IPF: adding one to the number of peak points of an autocorrelation function curve of the power spectrum of the intestinal motility signal in a certain time period,
IPFi＝Npeak+1
in the formula NpeakThe number of peak points of an autocorrelation function curve of the intestinal motility signal power spectrum;
(11) intestinal motility energy difference accumulation: accumulating the difference values of the kinetic energy of all the two adjacent intestines within a certain time period;
3. the apparatus according to claim 1, wherein the machine learning module determines a mapping model between the intestinal disease and the characteristic parameters by a machine learning method, specifically:
(1) randomly dividing characteristic data of the intestinal motility signals into two mutually exclusive sets, wherein one set is a training set S, and the other set is a cross check set CV, and classifying the data;
(2) training set S { (F)1,L1),(F2,L2),…(Fi,Li) Inputting the feature matrix of L epsilon {0,1,2,3, …, N } into the machine learning model, wherein Fi＝(Fi1,Fi2,Fi3,…Fi11) Extracting 11 enterokinetic signal features for the ith sample in a feature extraction module; l isiThe method comprises the following steps of (1) obtaining a category label corresponding to an ith sample, wherein the label 0 is normal intestinal tract, the label N represents intestinal tract disease N, and N is more than or equal to 1 and less than or equal to N;
(3) randomly extracting sp1 samples from the training set, then performing feature sampling on sp1 samples, extracting the same sp2 features from each sample to obtain a sampling set, repeating for T times, and totally obtaining T sampling sets; training an intestinal disease classification base learning device and a secondary learning device in each sampling set, and forming a mapping Model1 by using a voting method through the T intestinal disease classification base learning devices and the secondary learning device;
(4) and further performing parameter adjustment and optimization on the mapping Model1 between the intestinal diseases and the characteristic parameters by using a cross test set CV to obtain a mapping Model 2.
4. The non-contact intestinal disease diagnostic apparatus according to claim 3, wherein the intestinal disease classification base learner trained in the machine learning module is a multi-class classification SVM, specifically:
(1) setting samples containing z categories in each sampling set, dividing the z categories for v times, wherein v is larger than z, dividing one part of categories into positive categories and marking as (+1) by random division each time, dividing the other part of categories into negative categories and marking as (-1), and taking the positive and negative categories corresponding to each group as a training set; generating v training sets in total;
(2) training the generated v training sets to SVM respectively to obtain v SVM learners, inputting data of each class into the v SVM learners to obtain corresponding codes to form a z x v code matrix;
(3) and predicting the test samples by utilizing v classifiers respectively, forming a code by the prediction marks, comparing the code with the respective code of each class, and returning the class with the minimum distance as a final prediction result, wherein the v SVM classifiers form a multi-class classified SVM learner.
5. The apparatus according to claim 3, wherein the secondary learner for classifying intestinal diseases trained in the machine learning module is a decision tree, specifically:
(1) traversing each feature, and calculating the Keyny value of the feature, wherein the formula is as follows:
wherein, the data set a is a sampling matrix of sp1 × sp2, Gini (a) represents the Keyni index of the data set a, r represents that r samples exist in the data set, and r representsiRepresenting the proportion of the ith type sample in the data set a;
wherein A is 11 intestinal motility signal characteristics obtained from the characteristic extraction module, GiniA(a) The expression is a Gini index divided according to the characteristic A, k represents that the data set a is divided into k parts under the characteristic A, ajRepresenting the j-th data set after the division;
ΔGini(A)＝Gini(a)-GiniA(a)
wherein Δ gini (a) represents the gain value of the kini index before and after the division according to the characteristic a; selecting the characteristic with the maximum gain value of the Gini index as a splitting condition of the node to generate a new node;
(2) and repeating the process by taking the generated node as a root node, selecting new intestinal motility signal characteristics as splitting conditions, and completely dividing the data a.
6. The non-contact intestinal disease diagnosis device according to claim 3, wherein the tuning optimization in the machine learning module is specifically:
(1) determining parameters of a classification base learner and a secondary learner of the intestinal diseases, wherein the parameters of a decision tree comprise a maximum characteristic number max _ feature, a maximum depth max _ depth and a minimum sample number min _ samples _ split required by splitting, which are involved in judgment during splitting of a decision number; the parameters of the multi-class classification SVM are a punishment parameter c and a kernel function parameter gamma of the SVM;
(2) performing linear search on each parameter, training a base learner and a secondary learner by utilizing a CV set, and calculating the gradient between the model performance and each parameter;
(3) selecting a parameter with the maximum gradient value as a current tuning parameter according to the performance of the model and the gradient between the parameters, and tuning the current tuning parameter by using a gradient descent method;
(4) repeating the two steps until the gradient of all the parameters is less than a threshold value, which is a positive number close to 0;
(5) adjusting the number T1 of the classified bases of the intestinal diseases and the number T2 of the secondary learners, which is specifically as follows: and respectively calculating gradients between the model performance and the number T1 of the base learners and the number T2 of the secondary learners, and respectively optimizing T1 and T2 by using a gradient descent method to obtain the final number T1 of the intestinal disease classification base learners and the final number T2 of the secondary learners.
7. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the functions of the following modules in a non-contact diagnosis device for intestinal diseases when executing the computer program:
the preprocessing module is used for preprocessing radar echo signals acquired by the non-contact vital sign monitoring radar to acquire human vital sign signals, and then filtering interference to obtain preprocessed signals, wherein the acquired human vital sign signals comprise heartbeat signals, respiration signals, gastric dynamic signals and intestinal dynamic signals;
the signal separation module is used for processing the preprocessed signals by adopting a signal separation algorithm to separate intestinal motility signals of the human body;
the characteristic extraction module is used for extracting the characteristics of the separated intestinal motility signals and extracting the characteristic parameters of the intestinal movement;
the machine learning module is used for inputting the extracted characteristic parameters and the corresponding labels into machine learning and determining a mapping model between the intestinal diseases and the characteristic parameters by a machine learning method;
and the intestinal disease judgment module judges the intestinal diseases according to the mapping model.
8. The computer device of claim 7, wherein the processor, when executing the computer program to implement the machine learning module function, implements the steps of:
(1) randomly dividing characteristic data of the intestinal motility signals into two mutually exclusive sets, wherein one set is a training set S, and the other set is a cross check set CV, and classifying the data;
(2) training set S { (F)1,L1),(F2,L2),…(Fi,Li) Inputting the feature matrix of L epsilon {0,1,2,3, …, N } into the machine learning model, wherein Fi＝(Fi1,Fi2,Fi3,…Fi11) Extracting 11 enterokinetic signal features for the ith sample in a feature extraction module; l isiThe method comprises the following steps of (1) obtaining a category label corresponding to an ith sample, wherein the label 0 is normal intestinal tract, the label N represents intestinal tract disease N, and N is more than or equal to 1 and less than or equal to N;
(3) randomly extracting sp1 samples from the training set, then performing feature sampling on sp1 samples, extracting the same sp2 features from each sample to obtain a sampling set, repeating for T times, and totally obtaining T sampling sets; training an intestinal disease classification base learning device and a secondary learning device in each sampling set, and forming a mapping Model1 by using a voting method through the T intestinal disease classification base learning devices and the secondary learning device;
(4) and further performing parameter adjustment and optimization on the mapping Model1 between the intestinal diseases and the characteristic parameters by using a cross test set CV to obtain a mapping Model 2.
9. A computer-readable storage medium on which a computer program is stored, the computer program, when being executed by a processor, implementing functions of the following modules in a noncontact bowel disease diagnosis apparatus:
the preprocessing module is used for preprocessing radar echo signals acquired by the non-contact vital sign monitoring radar to acquire human vital sign signals, and then filtering interference to obtain preprocessed signals, wherein the acquired human vital sign signals comprise heartbeat signals, respiration signals, gastric dynamic signals and intestinal dynamic signals;
the signal separation module is used for processing the preprocessed signals by adopting a signal separation algorithm to separate intestinal motility signals of the human body;
the characteristic extraction module is used for extracting the characteristics of the separated intestinal motility signals and extracting the characteristic parameters of the intestinal movement;
the machine learning module is used for inputting the extracted characteristic parameters and the corresponding labels into machine learning and determining a mapping model between the intestinal diseases and the characteristic parameters by a machine learning method;
and the intestinal disease judgment module judges the intestinal diseases according to the mapping model.
10. The computer-readable storage medium of claim 9, wherein the computer program, when executed by a processor to implement the machine learning module function, implements the steps of:
(1) randomly dividing characteristic data of the intestinal motility signals into two mutually exclusive sets, wherein one set is a training set S, and the other set is a cross check set CV, and classifying the data;
(2) training set S { (F)1,L1),(F2,L2),…(Fi,Li) Inputting the feature matrix of L epsilon {0,1,2,3, …, N } into the machine learning model, wherein Fi＝(Fi1,Fi2,Fi3,…Fi11) Extracting 11 enterokinetic signal features for the ith sample in a feature extraction module; l isiThe method comprises the following steps of (1) obtaining a category label corresponding to an ith sample, wherein the label 0 is normal intestinal tract, the label N represents intestinal tract disease N, and N is more than or equal to 1 and less than or equal to N;
(3) randomly extracting sp1 samples from the training set, then performing feature sampling on sp1 samples, extracting the same sp2 features from each sample to obtain a sampling set, repeating for T times, and totally obtaining T sampling sets; training an intestinal disease classification base learning device and a secondary learning device in each sampling set, and forming a mapping Model1 by using a voting method through the T intestinal disease classification base learning devices and the secondary learning device;
(4) and further performing parameter adjustment and optimization on the mapping Model1 between the intestinal diseases and the characteristic parameters by using a cross test set CV to obtain a mapping Model 2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911233821.9A CN112914545A (en) | 2019-12-05 | 2019-12-05 | Non-contact intestinal disease diagnostic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911233821.9A CN112914545A (en) | 2019-12-05 | 2019-12-05 | Non-contact intestinal disease diagnostic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112914545A true CN112914545A (en) | 2021-06-08 |
Family
ID=76161060
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911233821.9A Pending CN112914545A (en) | 2019-12-05 | 2019-12-05 | Non-contact intestinal disease diagnostic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112914545A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114246563A (en) * | 2021-12-17 | 2022-03-29 | 重庆大学 | Intelligent heart and lung function monitoring equipment based on millimeter wave radar |
| CN116913519A (en) * | 2023-07-24 | 2023-10-20 | 东莞莱姆森科技建材有限公司 | Health monitoring method, device, equipment and storage medium based on intelligent mirror |
-
2019
- 2019-12-05 CN CN201911233821.9A patent/CN112914545A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114246563A (en) * | 2021-12-17 | 2022-03-29 | 重庆大学 | Intelligent heart and lung function monitoring equipment based on millimeter wave radar |
| CN114246563B (en) * | 2021-12-17 | 2023-11-17 | 重庆大学 | Heart and lung function intelligent monitoring equipment based on millimeter wave radar |
| CN116913519A (en) * | 2023-07-24 | 2023-10-20 | 东莞莱姆森科技建材有限公司 | Health monitoring method, device, equipment and storage medium based on intelligent mirror |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110840402B (en) | Atrial fibrillation signal identification method and system based on machine learning | |
| Meintjes et al. | Fundamental heart sound classification using the continuous wavelet transform and convolutional neural networks | |
| CN110876626A (en) | Depression detection system based on optimal lead selection of multi-lead electroencephalogram | |
| EP3803709B1 (en) | Detecting abnormalities in ecg signals | |
| Bairy et al. | Automated diagnosis of depression electroencephalograph signals using linear prediction coding and higher order spectra features | |
| EP3585264A1 (en) | Systems and methods of automatic cough identification | |
| CN112826513B (en) | Fetal heart rate detection system based on deep learning and specificity correction on FECG | |
| Khan et al. | Automated Detection and Classification of Gastrointestinal Diseases using surface-EMG Signals | |
| CN112914545A (en) | Non-contact intestinal disease diagnostic equipment | |
| Lo et al. | Deep learning for detection of fetal ECG from multi-channel abdominal leads | |
| Liu et al. | A novel P-QRS-T wave localization method in ECG signals based on hybrid neural networks | |
| Peimankar et al. | Ensemble learning for detection of short episodes of atrial fibrillation | |
| CN116188466B (en) | Method and device for determining in-vivo residence time of medical instrument | |
| Samanta et al. | Neuromuscular disease detection employing deep feature extraction from cross spectrum images of electromyography signals | |
| Zehir et al. | Support vector machine for human identification based on non-fiducial features of the ecg | |
| Putri et al. | Retracted: Prediction Instrument of Diabetes Mellitus based on Iridology Image and Artificial Neural Network: Initial Findings | |
| Kraft et al. | Vibroarthrography using convolutional neural networks | |
| Chen et al. | Breathing pattern recognition of abdominal wall movement by using ensemble empirical mode decomposition | |
| Paila et al. | Detection and Analysis of Cardiac Arrhythmias from Heartbeat Classification | |
| KR102275263B1 (en) | Respiratory measurement system using depth camera | |
| Khan et al. | Detection of acute coronary syndrome using electrocardiogram signal analysis | |
| CN113425298A (en) | Method for analyzing depression degree by collecting data through wearable equipment | |
| Cao et al. | A novel approach analysing the dynamic brain functional connectivity for improved MCI detection | |
| Lopes et al. | Machine Learning and Stereoelectroencephalographic Feature Extraction for Brain Tissue Classification | |
| Cherifi et al. | Automated Diagnosis of Neuromuscular Disorders using EMG Signals |
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 |
Application publication date: 20210608 |
|
| RJ01 | Rejection of invention patent application after publication |