CN115299956B - Myocardial ischemia detection method and system based on definite learning and electrocardiogram - Google Patents
Myocardial ischemia detection method and system based on definite learning and electrocardiogram Download PDFInfo
- Publication number
- CN115299956B CN115299956B CN202210999189.4A CN202210999189A CN115299956B CN 115299956 B CN115299956 B CN 115299956B CN 202210999189 A CN202210999189 A CN 202210999189A CN 115299956 B CN115299956 B CN 115299956B
- Authority
- CN
- China
- Prior art keywords
- electrocardiogram
- ring
- myocardial ischemia
- electrocardiographic
- enhanced
- 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
- 208000031225 myocardial ischemia Diseases 0.000 title claims abstract description 50
- 238000001514 detection method Methods 0.000 title claims abstract description 27
- 238000010586 diagram Methods 0.000 claims abstract description 38
- 239000013598 vector Substances 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000013145 classification model Methods 0.000 claims abstract description 21
- 238000012545 processing Methods 0.000 claims abstract description 19
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 15
- 238000012360 testing method Methods 0.000 claims abstract description 15
- 230000000747 cardiac effect Effects 0.000 claims abstract description 9
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 238000000034 method Methods 0.000 claims description 29
- 238000000605 extraction Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 description 12
- 238000002565 electrocardiography Methods 0.000 description 10
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 230000002107 myocardial effect Effects 0.000 description 3
- 238000005312 nonlinear dynamic Methods 0.000 description 3
- 230000000737 periodic effect Effects 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000002586 coronary angiography Methods 0.000 description 2
- 239000006185 dispersion Substances 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000010412 perfusion Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000036770 blood supply Effects 0.000 description 1
- 238000013184 cardiac magnetic resonance imaging Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010968 computed tomography angiography Methods 0.000 description 1
- 210000004351 coronary vessel Anatomy 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000012631 diagnostic technique Methods 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 208000019622 heart disease Diseases 0.000 description 1
- 230000004217 heart function Effects 0.000 description 1
- 208000028867 ischemia Diseases 0.000 description 1
- 230000004807 localization Effects 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
- 210000002569 neuron Anatomy 0.000 description 1
- 231100000915 pathological change Toxicity 0.000 description 1
- 230000036285 pathological change Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000010854 vegetative compatibility group Methods 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
- A61B5/346—Analysis of electrocardiograms
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
-
- 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
- A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
-
- 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
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Pathology (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physiology (AREA)
- Psychiatry (AREA)
- Cardiology (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Fuzzy Systems (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention provides a myocardial ischemia detection method and system based on definite learning and electrocardiogram, which respectively acquire electrocardiogram data of normal individuals and myocardial ischemia patients, and divide the acquired electrocardiogram data into a test set and a training set; preprocessing an electrocardiogram in the training set, converting the electrocardiogram into a three-dimensional electrocardiogram vector diagram, and modeling the three-dimensional electrocardiogram vector diagram by using a discrete determination algorithm to obtain an enhanced electrocardiogram dynamic diagram; extracting ring features on the enhanced electrocardiographic dynamic diagram based on the complete cardiac beat, processing the ring features extracted by the cardiac beats to obtain final feature values, and inputting the final feature values into a classification model for training to obtain a trained classification model; and processing the electrocardiogram in the test set, inputting the processed electrocardiogram into a trained classification model, and outputting a classification result. The characteristics capable of effectively reflecting myocardial ischemia are extracted, and the classifier is trained to obtain better myocardial ischemia detection performance.
Description
Technical Field
The invention belongs to the technical field of medical detection, and particularly relates to a myocardial ischemia detection method and system based on definite learning and electrocardiogram.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
At present, a plurality of means are clinically used for detecting or diagnosing myocardial ischemia, such as coronary CT angiography and coronary angiography, and detecting and evaluating the degree of pathological changes of coronary arteries on physiological structures; methods such as nuclear myocardial perfusion imaging, cardiac magnetic resonance imaging, and fractional coronary flow reserve can be used to detect and assess whether the function of myocardial blood supply is normal. Diagnostic techniques for myocardial ischemia continue to evolve, but such as the methods described above suffer from the disadvantage of being costly or invasive.
Electrocardiogram records electrical signals transmitted from the electrical activity of the heart to the body surface, and is an important basis in assessing heart function and diagnosing various heart diseases. Electrocardiogram is easy to obtain, low in cost and non-invasive, and studies have pointed out that the electrocardiogram contains a large amount of information reflecting myocardial ischemia.
Based on electrocardiography, a great number of myocardial ischemia detection technologies are proposed, such as extracting various transform domain features from electrocardiography based on time domain, frequency domain and time-frequency domain analysis technologies, and then implementing detection and classification of myocardial ischemia by combining a machine learning method. With the application and success of deep learning in more and more fields, many studies at home and abroad use the deep learning method for detecting myocardial ischemia. In addition, in recent years, the dynamic analysis method of the electrocardiosignal is also applied to the detection of myocardial ischemia, such as a definite learning algorithm, and can be used for locally and accurately modeling or learning the nonlinear dynamics inherent in the electrocardiosignal and the time sequence, thereby providing a way for better understanding and analyzing the myocardial ischemia electrocardiosignal from the aspect of dynamics.
Although the above-described methods have made good progress in myocardial ischemia detection, limitations remain. The machine learning-based method extracts effective time domain and time-frequency domain features from the electrocardiosignal through artificial feature engineering, and the features are easy to explain and understand and apply by doctors; efficient feature extraction typically relies on accurate localization of the P-wave, QRS-wave, ST-segment and T-wave in the electrocardiographic signal, which is often difficult. The method based on deep learning can automatically learn the electrocardio characteristics from a large amount of data to improve the accuracy of myocardial ischemia detection, but the basis for detecting myocardial ischemia is difficult to explain at present; and the method often requires a large-scale data set, and acquiring large-scale medical data in reality is expensive and takes a long time. Firstly extracting effective features from an electrocardiographic dynamic diagram containing deep ischemia information based on a method for determining learning, and secondly establishing a myocardial ischemia detection model by combining a machine learning method; studies have shown that by modeling the ST-T segment of electrocardiographic data and extracting the time dispersion and the space dispersion characteristics thereof, myocardial ischemia can be accurately detected when the electrocardiogram has no specific change, i.e. myocardial ischemia cannot be clearly diagnosed only by visual inspection of the electrocardiogram. At present, the characteristics extracted from the electrocardiographic dynamic diagram are fewer, namely, the information utilization degree of the electrocardiographic dynamic diagram is limited. In order to improve the performance of myocardial ischemia detection based on an electrocardiographic map and expand the application range of the electrocardiographic map, further potential exploitation is needed.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides a myocardial ischemia detection method and a myocardial ischemia detection system based on definite learning and electrocardiogram, which are based on definite learning algorithm and electrocardiogram, and the method and the system extract new characteristics capable of effectively reflecting myocardial ischemia from the perspective of dynamics and train a classifier to obtain better myocardial ischemia detection performance.
To achieve the above object, one or more embodiments of the present invention provide the following technical solutions: a myocardial ischemia detection method based on deterministic learning and electrocardiogram, comprising the steps of:
respectively acquiring electrocardiogram data of a normal individual and a myocardial ischemia patient, and dividing the acquired electrocardiogram data into a test set and a training set;
preprocessing an electrocardiogram in the training set, converting the electrocardiogram into a three-dimensional electrocardiogram vector diagram, and modeling the three-dimensional electrocardiogram vector diagram by using a discrete determination algorithm to obtain an enhanced electrocardiogram dynamic diagram;
Extracting ring features on the enhanced electrocardiographic dynamic diagram based on the complete cardiac beat, processing the ring features extracted by the cardiac beats to obtain final feature values, and inputting the final feature values into a classification model for training to obtain a trained classification model;
and processing the electrocardiogram in the test set, inputting the processed electrocardiogram into a trained classification model, and outputting a classification result.
A second aspect of the present invention provides a myocardial ischemia detection system based on deterministic learning and electrocardiogram, characterized by comprising:
The data acquisition module is used for respectively acquiring electrocardiogram data of a normal individual and a myocardial ischemia patient and dividing the acquired electrocardiogram data into a test set and a training set;
The data processing module is used for preprocessing the electrocardiograms in the training set and converting the electrocardiograms into a three-dimensional electrocardio vector diagram, and modeling the three-dimensional electrocardio vector diagram by using a discrete determination algorithm to obtain an enhanced electrocardio dynamic diagram;
The model training module is used for extracting ring features in a ring shape on the enhanced electrocardiographic dynamic diagram based on the complete heart beat, processing the ring features extracted by the heart beats to be used as final feature values, and inputting the final feature values into the classification model for training to obtain a trained classification model;
And the classification module is used for processing the electrocardiograms in the test set, inputting the processed electrocardiograms into the trained classification model and outputting a classification result.
A third aspect of the invention provides a computer readable storage medium storing computer instructions which, when executed by a processor, perform the steps of the method described above.
A fourth aspect of the invention provides an electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of the method described above.
The one or more of the above technical solutions have the following beneficial effects:
On the one hand, from the dynamics point of view, the characteristics capable of effectively reflecting myocardial ischemia are extracted, and the characteristics can reflect the change of electrocardiosignals caused by deeper myocardial ischemia, thereby being beneficial to obtaining better myocardial ischemia detection performance; on the other hand, compared with methods such as coronary angiography, nuclide myocardial perfusion imaging and the like, the method has the advantages that the used data are simple and easy to obtain, and the patient is not invasive in the acquisition process.
Additional aspects of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
FIG. 1 is a flow chart of a first embodiment of the present invention;
FIG. 2 is an ECG with noise removed in accordance with a first embodiment of the present invention;
FIG. 3 is an enhanced CDG computationally generated from electrocardiographic data according to an embodiment of the present invention;
FIG. 4 is a visual representation of feature 1 in a first embodiment of the invention;
FIG. 5 is a convex hull visualization of the feature 2 calculation process in accordance with the first embodiment of the present invention;
FIG. 6 is a visual representation of projections during the computation of features 3, 5 and 6 in accordance with an embodiment of the present invention;
Fig. 7 is a visual representation of feature 4 in a first embodiment of the invention.
Detailed Description
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the invention. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
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 exemplary embodiments according to the present invention.
Embodiments of the invention and features of the embodiments may be combined with each other without conflict.
Example 1
As shown in fig. 1, the present embodiment discloses a myocardial ischemia detection method based on deterministic learning and electrocardiogram, comprising the steps of:
Step 1: respectively acquiring electrocardiogram data of a normal individual and a myocardial ischemia patient, and dividing the acquired electrocardiogram data into a test set and a training set;
step 2: preprocessing an electrocardiogram in the training set, converting the electrocardiogram into a three-dimensional electrocardiogram vector diagram, and modeling the three-dimensional electrocardiogram vector diagram by using a discrete determination algorithm to obtain an enhanced electrocardiogram dynamic diagram;
Step 3: extracting ring features on the enhanced electrocardiographic dynamic diagram based on the complete cardiac beat, processing the ring features extracted by the cardiac beats to obtain final feature values, and inputting the final feature values into a classification model for training to obtain a trained classification model;
Step 4: and processing the electrocardiogram in the test set, inputting the processed electrocardiogram into a trained classification model, and outputting a classification result.
In step 1 of this example, electrocardiographic data was used from the Shandong Qilu hospital and Shandong province hospital, and 333 electrocardiographic recordings were collected in total, including conventional 12-lead electrocardiographic recordings of 141 myocardial ischemia patients in Qilu hospital, conventional 12-lead electrocardiographic recordings of 143 normal individuals, and conventional 12-lead electrocardiographic recordings of 49 normal individuals in Qilu hospital. The electrocardiographic data collected in this example was 10 seconds long and the sampling rate was 1000HZ.
Conventional 12-lead electrocardiography refers to numerical form data of I, II, III, aVR, aVL, aVF, V1, V2, V3, V4, V5, and V6.
In step 2 of this example, a3 rd order high pass butterworth filter with a cut-off frequency of 0.5Hz was used to remove baseline drift in electrocardiographic data, and a 50Hz notch filter was used to remove power frequency interference.
As shown in fig. 2, which is a noise-removed ECG, a 12-lead ECG (8 of the individual leads i, ii, V1, V2, V3, V4, V5, and V6) is converted to a 3-lead VCG based on a Kors matrix, wherein Kors matrix is as follows:
The VCG is then modeled based on the determined learning to obtain an enhanced electrocardiographic map. The electrocardiographic signals may be regarded as periodic or periodic-like signals generated by a nonlinear system, and thus the deterministic learning algorithm may dynamically model electrocardiographic data to obtain an enhanced electrocardiographic map. Better signal combinations for kinetic modeling can be explored by dimensionality reduction of the ECG or picking partial lead signals in the ECG.
Specifically, the specific method for modeling by utilizing the determined learning is as follows:
first, consider the ECG as a periodic non-stationary signal generated by a complex nonlinear dynamic system:
Wherein: x (t) = [ X 1(t),x2(t),…,x12(t)]T ] is the system state, i.e. the human standard 12-lead electrocardiographic signal; p is a system parameter, F (x (t))= [ F 1(x(t)),f2(x(t)),…,f12(x(t))]T is an unknown nonlinear dynamics.
Second, the 12-lead electrocardiographic signal is converted into a 3-lead electrocardiographic vector signal (VCG), and the complete electrocardiographic cycle data is modeled based on the deterministic learning. Studies have shown that 3-lead VCGs are a compact set of orthogonal measurements of cardiac spatiotemporal electrical activity that can overcome the dimensional problems associated with 12-lead electrocardiographic signals while preserving electrocardiographic information. The 12-lead ECG (8 independent leads I, II, V1, V2, V3, V4, V5, and V6) is converted to a 3-lead VCG by a Kors matrix.
The following three-dimensional dynamic system is considered:
Wherein, F v(V(t))＝[Fv1(V(t)),Fv2(V(t)),Fv3 (V (t)) ] is nonlinear system dynamics, i.e. is the dynamics law inherent in VCG, V (t) = [ V 1(t),v2(t),v3(t)]T∈R3 is VCG. And calculating a discrete electrocardio vector signal V (k) by using the 12-lead electrocardio signal.
And thirdly, dynamically modeling the VCG sampling data. The discrete sampling model of the system (3) can be approximated by the following euler model:
V(k+1)＝V(k)+TsF(V(k)；p)+ε(k；Ts),V(0)＝V0 (4)
Wherein T s is a sampling period, epsilon (k; ts) represents modeling error of the Euler sampling model, V (0) is an initial sampling point, V (k) represents sampling data at time T k＝t0+kTs, namely, a sampling point obtained after k (k is equal to or greater than 1 and is equal to or less than N-1, N is sampling data length) sampling periods from the initial sampling point.
A neural network identifier is adopted to learn dynamics rules:
Wherein, Is RBF neural network for identifying dynamics system, S (V (k)) is regression vector of RBF neural network obtained when input is V (k), and is used for identifying dynamics systemIs the state of the neural network identifier, v i (k) is the system true state,/>Representing the estimation of RBF neural network weight, S: R n→Rm is the regression vector of RBF neural network, alpha i represents the identifier parameter satisfaction/>Where ω i is the convergence speed of the passive parameter estimation error system.
Based on Liapunov theory designIs an adaptive update law of (a):
Wherein, Is state estimation error, gamma is the learning gain parameter satisfiesWherein II S 2max is the upper bound of the 2-norm of the RBF neural network regression vector.
According to a definite learning theory, the subvectors formed by neuron functions along a periodic sampling data sequence in the RBF neural network meet partial continuous excitation conditions, so that an electrocardiograph dynamics system F (V (k); p) can be locally and accurately modeled, and finally, a dynamics law F v (V (k))ofelectrocardiograph vector signals is obtained:
Finally, output And (5) visualizing in a three-dimensional space to obtain the enhanced electrocardiographic dynamic diagram.
In this embodiment, in step 3, the heart beat is divided by locating the R point position in the VCG using the Pan-Tompkins method. The single integral heart beat is presented as a ring or approximate ring in the enhanced electrocardiographic dynamic diagram, and 6 characteristics are extracted from the enhanced electrocardiographic dynamic diagram of the single heart beat, wherein the characteristics comprise a maximum vector module value of the ring, a convex hull volume of the ring, an area of the ring on an optimal projection surface, a maximum distance from a point on the ring to a center of mass of the ring, an included angle between the optimal projection surface and an XY plane of the ring and a perimeter of the ring on the optimal projection surface.
Specifically, as shown in fig. 4, for the extraction of the feature 1 loop maximum vector modulus feature: and calculating the modular value of all vectors in the ring in the enhanced electrocardiographic dynamic diagram, namely, the Euclidean distance from the point on the ring to the origin of the three-dimensional coordinate system, and taking the maximum value as the modular value of the maximum vector of the ring.
Extracting the volume of the feature 2-ring convex hull: since points on a ring in the enhanced electrocardiographic map can be regarded as a set of points in three-dimensional space, the volume of a single ring therein can be quantified, and the flatness of the ring is reflected to some extent.
As shown in fig. 5, in this embodiment, convex hull points in a point set are calculated by a convex hull algorithm, a three-dimensional convex polyhedron is obtained by surface modeling, all points in the point set are enveloped in the polyhedron, and finally, the volume of the three-dimensional convex polyhedron is calculated to obtain the minimum convex hull volume of the point set.
In this embodiment, the convex hull points in the point set are calculated by QuickHull algorithm, the three-dimensional convex polyhedron is obtained by surface modeling, all the points in the point set are enveloped in the polyhedron, and finally the convex hull volume of the point set can be obtained by calculating the volume of the three-dimensional convex polyhedron. As for the convex polyhedron a, it can be considered that it is surrounded by triangular faces S 1,S2,…,Sm, and the apex of triangular face S i is a i,Bi,Ci (three apexes are arranged counterclockwise when viewed from the outside), and the coordinate values thereof can be formed as follows:
The volume of the convex polyhedron is:
It should be noted that convex hull refers to a smallest convex polyhedron that contains a series of known vertices in spatial geometry.
As shown in fig. 6, the extraction of the area features of the feature 3 ring on the optimal projection plane: firstly, a best fitting plane of a point set is obtained by using a least square method and is taken as a best projection plane, and the best fitting plane is shown as a formula (2):
Ax+By+Gz+D＝0 (2)
And (3) projecting the points on the ring onto an optimal projection surface according to a formula (3), converting the three-dimensional coordinates of the projection points into two-dimensional coordinates according to a formula (4), and finally calculating the convex hull area of the two-dimensional point set through a convex hull algorithm.
Wherein A, B, C, D are parameters of the optimal projection plane, (x 0,y0,z0) are points on the ring in the three-dimensional space, and (x, y, z) are points on the ring projected to the optimal projection plane.
Points on the ring in the enhanced electrocardiographic graph form a point set V= [ V 1,v2,…,vm ], wherein m is the number of points on the ring, one point is taken as an origin O, and a unit vector of a vector from any point of the point set divided by the origin to the origin is calculated firstlyThen find the vector/>, and the vector in the projection planePerpendicular Unit vector/>And (4) obtaining the two-dimensional coordinates corresponding to the three-dimensional points in the projection plane according to the formula (4).
Wherein (a, b) is the two-dimensional coordinates after conversion.
As shown in fig. 7, extraction of maximum distance features from points on the feature 4 ring to the ring centroid: and calculating the Euclidean distance from the point on the ring to the mass center of the ring in the enhanced electrocardiographic dynamic graph, and taking the maximum value of the Euclidean distance as the maximum distance from the point on the ring to the mass center of the ring.
The centroid calculation is shown in formula (5), wherein the density ρ i =1, m is the number of points on the ring, (x i,yi,zi) is the three-dimensional coordinates of the points on the ring, and (x c,yc,zc) is the three-dimensional coordinates of the centroid of the ring.
As shown in fig. 6, the extraction of the feature of the angle between the best projection plane of the feature 5 ring and the XY plane: the calculation formula is shown as formula (6), vectorAnd/>Normal vectors to the best projection plane and XY plane of the ring, respectively.
As shown in fig. 6, the perimeter feature of the feature 6 ring on the optimal projection plane is extracted: aiming at a point set V= [ V 1,v2,…,vm ] formed by points on a ring in the enhanced electrocardiographic dynamic diagram, corresponding projection points on a projection plane are obtained according to a formula (3), euclidean distance between every two adjacent projection points is calculated, and the calculated Euclidean distances are summed to obtain the circumference of the ring on the optimal projection plane.
Based on the scheme, 6 features are extracted from the enhanced electrocardiographic dynamic images of a single heart beat, the operations are repeated on the enhanced electrocardiographic dynamic images of a plurality of heart beats, a plurality of sets of feature data are obtained, the obtained plurality of sets of feature data are calculated to obtain the average value and variance, and the average value and the variance value are used as final feature values.
The obtained 12 eigenvalues are input into a support vector machine model for training, 10 times of 5-fold cross validation are adopted, the average accuracy of the trained support vector machine is 88.83%, the sensitivity is 86.80%, and the specificity is 90.30%.
After the same processing is carried out on the electrocardiographic data in the test set, 12 eigenvalues are obtained and input into a trained support vector machine model, and classification results are output, wherein the detection results can be used as references for clinically diagnosing myocardial ischemia or for primarily screening myocardial ischemia patients.
Example two
It is an object of the present embodiment to provide a myocardial ischemia detection system based on deterministic learning and electrocardiogram, comprising:
The data acquisition module is used for respectively acquiring electrocardiogram data of a normal individual and a myocardial ischemia patient and dividing the acquired electrocardiogram data into a test set and a training set;
The data processing module is used for preprocessing the electrocardiograms in the training set and converting the electrocardiograms into a three-dimensional electrocardio vector diagram, and modeling the three-dimensional electrocardio vector diagram by using a discrete determination algorithm to obtain an enhanced electrocardio dynamic diagram;
The model training module is used for extracting ring features in a ring shape on the enhanced electrocardiographic dynamic diagram based on the complete heart beat, processing the ring features extracted by the heart beats to be used as final feature values, and inputting the final feature values into the classification model for training to obtain a trained classification model;
And the classification module is used for processing the electrocardiograms in the test set, inputting the processed electrocardiograms into the trained classification model and outputting a classification result.
Example III
It is an object of the present embodiment to provide a computing device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, which processor implements the steps of the method described above when executing the program.
Example IV
An object of the present embodiment is to provide a computer-readable storage medium.
A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of the above method.
The steps involved in the devices of the second, third and fourth embodiments correspond to those of the first embodiment of the method, and the detailed description of the embodiments can be found in the related description section of the first embodiment. The term "computer-readable storage medium" should be taken to include a single medium or multiple media including one or more sets of instructions; it should also be understood to include any medium capable of storing, encoding or carrying a set of instructions for execution by a processor and that cause the processor to perform any one of the methods of the present invention.
It will be appreciated by those skilled in the art that the modules or steps of the invention described above may be implemented by general-purpose computer means, alternatively they may be implemented by program code executable by computing means, whereby they may be stored in storage means for execution by computing means, or they may be made into individual integrated circuit modules separately, or a plurality of modules or steps in them may be made into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (7)
1. A method for detecting myocardial ischemia based on deterministic learning and electrocardiogram, comprising the steps of:
respectively acquiring electrocardiogram data of a normal individual and a myocardial ischemia patient, and dividing the acquired electrocardiogram data into a test set and a training set;
preprocessing an electrocardiogram in the training set, converting the electrocardiogram into a three-dimensional electrocardiogram vector diagram, and modeling the three-dimensional electrocardiogram vector diagram by using a discrete determination algorithm to obtain an enhanced electrocardiogram dynamic diagram;
Extracting ring features on the enhanced electrocardiographic dynamic diagram based on the complete cardiac beat, processing the ring features extracted by the cardiac beats to obtain final feature values, and inputting the final feature values into a classification model for training to obtain a trained classification model;
Processing the electrocardiogram in the test set, inputting the processed electrocardiogram into a trained classification model, and outputting a classification result;
The ring features extracted in a ring shape on the enhanced electrocardiographic dynamic diagram based on the complete heart beat comprise a ring maximum vector module value, a ring convex hull volume, an area of a ring on an optimal projection surface, a maximum distance from a point on the ring to a ring centroid, an included angle between the optimal projection surface and an XY plane of the ring and a perimeter of the ring on the optimal projection surface;
the extraction of the volume of the annular convex hull is as follows: taking points in the ring of the enhanced electrocardiographic dynamic graph as a point set in a three-dimensional space, calculating convex hull points in the point set by using a convex hull algorithm, modeling the convex hull points to obtain a convex polyhedron, and calculating the volume of the convex polyhedron to obtain the minimum convex hull volume in the point set;
The extraction of the area of the ring on the optimal projection surface is as follows: calculating the best fitting plane of the point set by using a least square method as an optimal projection plane, converting the three-dimensional coordinates in the point set into two-dimensional coordinates based on the optimal projection plane, and calculating the convex hull area in the point set formed by the two-dimensional coordinates by using a convex hull algorithm;
The circumference of the ring on the optimal projection surface is extracted as follows: based on a point set formed by points on the ring in the enhanced electrocardiographic dynamic graph, calculating the Euclidean distance of all adjacent two points based on the corresponding projection points of the optimal projection plane, and summing the Euclidean distances of every two adjacent points to obtain the circumference of the ring on the optimal projection plane.
2. A method for myocardial ischemia detection based on deterministic learning and electrocardiogram as in claim 1 wherein preprocessing the electrocardiogram comprises: a butterworth filter is used to remove baseline wander in electrocardiographic data and a notch filter is used to remove power frequency interference.
3. The myocardial ischemia detection method based on deterministic learning and electrocardiogram as defined in claim 1, wherein the position of the R point in the three-dimensional electrocardiographic vector diagram is located by Pan-Tompkins method, and heart beat division is performed.
4. A myocardial ischemia detection method based on deterministic learning and electrocardiogram as recited in claim 1, wherein the mean and variance of the ring features extracted from a plurality of beats are calculated as final feature values.
5. A myocardial ischemia detection system based on deterministic learning and electrocardiogram, employing a method for detecting myocardial ischemia based on deterministic learning and electrocardiogram according to claim 1, comprising the steps of:
the data acquisition module is used for respectively acquiring electrocardiogram data of a normal individual and a myocardial ischemia patient and dividing the acquired electrocardiogram data into a test set and a training set;
The data processing module is used for preprocessing the electrocardiograms in the training set and converting the electrocardiograms into a three-dimensional electrocardio vector diagram, and modeling the three-dimensional electrocardio vector diagram by using a discrete determination algorithm to obtain an enhanced electrocardio dynamic diagram;
The model training module is used for extracting ring features in a ring shape on the enhanced electrocardiographic dynamic diagram based on the complete heart beat, processing the ring features extracted by the heart beats to be used as final feature values, and inputting the final feature values into the classification model for training to obtain a trained classification model;
And the classification module is used for processing the electrocardiograms in the test set, inputting the processed electrocardiograms into the trained classification model and outputting a classification result.
6. A computer-readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of a method for detecting myocardial ischemia based on deterministic learning and electrocardiogram as claimed in any one of claims 1 to 4.
7. A processing device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of a method for myocardial ischemia detection based on deterministic learning and electrocardiogram according to any one of claims 1-4 when the program is executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210999189.4A CN115299956B (en) | 2022-08-19 | 2022-08-19 | Myocardial ischemia detection method and system based on definite learning and electrocardiogram |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210999189.4A CN115299956B (en) | 2022-08-19 | 2022-08-19 | Myocardial ischemia detection method and system based on definite learning and electrocardiogram |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115299956A CN115299956A (en) | 2022-11-08 |
| CN115299956B true CN115299956B (en) | 2024-06-25 |
Family
ID=83863316
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210999189.4A Active CN115299956B (en) | 2022-08-19 | 2022-08-19 | Myocardial ischemia detection method and system based on definite learning and electrocardiogram |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115299956B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103549949A (en) * | 2013-10-21 | 2014-02-05 | 华南理工大学 | Myocardial ischemia auxiliary detecting method based on deterministic learning theory |
| CN110151165A (en) * | 2019-04-30 | 2019-08-23 | 杭州电子科技大学 | A kind of monocardiogram classification method based on Nonlinear Dynamical Characteristics |
| CN111261289A (en) * | 2018-11-30 | 2020-06-09 | 上海图灵医疗科技有限公司 | Heart disease detection method based on artificial intelligence model |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009153677A (en) * | 2007-12-26 | 2009-07-16 | Konica Minolta Medical & Graphic Inc | Kinetic image processing system |
| US10342620B2 (en) * | 2014-04-15 | 2019-07-09 | Siemens Healthcare Gmbh | Efficient treatment of atrial fibrillation using three-dimensional electrical potential model |
| CN106815570B (en) * | 2016-12-30 | 2020-10-27 | 广东技术师范大学 | Electrocardiosignal ST-T segment identification method based on dynamic pattern identification |
| WO2018184779A1 (en) * | 2017-04-06 | 2018-10-11 | Koninklijke Philips N.V. | Coronary artery disease metric based on estimation of myocardial microvascular resistance from ecg signal |
| CN107260161B (en) * | 2017-07-18 | 2019-02-19 | 上海图灵医疗科技有限公司 | A kind of electrocardio dynamics data quantitative analysis method |
| CN108786074B (en) * | 2018-06-14 | 2019-06-25 | 浙江大学 | Training load monitoring device |
| CN109512423A (en) * | 2018-12-06 | 2019-03-26 | 杭州电子科技大学 | A kind of myocardial ischemia Risk Stratification Methods based on determining study and deep learning |
| 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 |
| CA3165746A1 (en) * | 2019-12-23 | 2021-07-01 | Analytics For Life Inc. | Method and system for signal quality assessment and rejection using heart cycle variability |
| EP4266986A1 (en) * | 2020-12-23 | 2023-11-01 | Analytics For Life Inc. | Method and system for engineering cycle variability-related features from biophysical signals for use in characterizing physiological systems |
-
2022
- 2022-08-19 CN CN202210999189.4A patent/CN115299956B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103549949A (en) * | 2013-10-21 | 2014-02-05 | 华南理工大学 | Myocardial ischemia auxiliary detecting method based on deterministic learning theory |
| CN111261289A (en) * | 2018-11-30 | 2020-06-09 | 上海图灵医疗科技有限公司 | Heart disease detection method based on artificial intelligence model |
| CN110151165A (en) * | 2019-04-30 | 2019-08-23 | 杭州电子科技大学 | A kind of monocardiogram classification method based on Nonlinear Dynamical Characteristics |
Non-Patent Citations (1)
| Title |
|---|
| 基于确定学习及多特征融合的心肌缺血检测研究;王志远;万方学位论文;20230423;1-48 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115299956A (en) | 2022-11-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11089988B2 (en) | Non-invasive method and system for estimating arterial flow characteristics | |
| Yang | Multiscale recurrence quantification analysis of spatial cardiac vectorcardiogram signals | |
| US11147516B2 (en) | Methods and systems to quantify and remove asynchronous noise in biophysical signals | |
| CN108324263B (en) | Noninvasive cardiac electrophysiology inversion method based on low-rank sparse constraint | |
| Yang et al. | Identification of myocardial infarction (MI) using spatio-temporal heart dynamics | |
| US20190117164A1 (en) | Methods and systems of de-noising magnetic-field based sensor data of electrophysiological signals | |
| CN114901143A (en) | System and method for reduced-lead electrocardiography diagnosis using deep neural networks and rule-based systems | |
| Al-Ani | ECG waveform classification based on P-QRS-T wave recognition | |
| Prabhakararao et al. | A weighted SVM based approach for automatic detection of posterior myocardial infarction using VCG signals | |
| Singhal et al. | A systematic review on artificial intelligence-based techniques for diagnosis of cardiovascular arrhythmia diseases: challenges and opportunities | |
| CN114901145A (en) | System and method for electrocardiographic diagnosis using deep neural networks and rule-based systems | |
| CN118119333A (en) | Methods and systems for engineering conduction bias characteristics from biophysical signals for characterizing physiological systems | |
| Lee et al. | Reconstruction of precordial lead electrocardiogram from limb leads using the state-space model | |
| Tapotee et al. | M2ECG: Wearable mechanocardiograms to electrocardiogram estimation using deep learning | |
| CN115299956B (en) | Myocardial ischemia detection method and system based on definite learning and electrocardiogram | |
| Noponen et al. | Invariant trajectory classification of dynamical systems with a case study on ECG | |
| Huang et al. | A vectorcardiogram-based classification system for the detection of Myocardial infarction | |
| Sadhukhan et al. | Automated identification of Myocardial Infarction using a single Vectorcardiographic feature | |
| Mayapur | Detection and classification of heart defects | |
| Vysiya et al. | Automatic detection of cardiac arrhythmias in ECG signal for IoT application | |
| Mitra et al. | A rough set based approach for ecg classification | |
| Kheirati Roonizi | ADAPTIVE MODELS-BASED CARDIAC SIGNALS ANALYSIS AND FEATURE EXTRACTION | |
| Potekhin et al. | ANALYSIS OF ACUTE MYOCARDIAL INFARCTION USING ECG SIGNALS AND MACHINE LEARNING ALGORITHMS | |
| Munck et al. | Body surface mapping of the mechanical cardiac activity | |
| Rafik et al. | On the Boundedness solutions of the difference equation xn+ 1= axan+ bxan-1, 0< a≤ 2 and its application in medicine |
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 |