CN113303805A - Automatic electrocardiogram diagnosis method - Google Patents
Automatic electrocardiogram diagnosis method Download PDFInfo
- Publication number
- CN113303805A CN113303805A CN202010119455.0A CN202010119455A CN113303805A CN 113303805 A CN113303805 A CN 113303805A CN 202010119455 A CN202010119455 A CN 202010119455A CN 113303805 A CN113303805 A CN 113303805A
- Authority
- CN
- China
- Prior art keywords
- electrocardiographic
- data
- automated
- diagnostic method
- specific
- 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
- 238000003745 diagnosis Methods 0.000 title abstract description 34
- 238000000034 method Methods 0.000 title abstract description 28
- 208000019622 heart disease Diseases 0.000 claims abstract description 55
- 238000005259 measurement Methods 0.000 claims abstract description 47
- 238000002405 diagnostic procedure Methods 0.000 claims description 16
- 230000033764 rhythmic process Effects 0.000 claims description 10
- 238000007405 data analysis Methods 0.000 claims description 9
- 208000020446 Cardiac disease Diseases 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 claims 1
- 238000002372 labelling Methods 0.000 claims 1
- 206010006578 Bundle-Branch Block Diseases 0.000 description 10
- 238000002565 electrocardiography Methods 0.000 description 9
- 206010006582 Bundle branch block right Diseases 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 201000007916 right bundle branch block Diseases 0.000 description 8
- 201000010099 disease Diseases 0.000 description 7
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 7
- 208000004301 Sinus Arrhythmia Diseases 0.000 description 6
- 206010040741 Sinus bradycardia Diseases 0.000 description 5
- 208000010125 myocardial infarction Diseases 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 3
- 206010003130 Arrhythmia supraventricular Diseases 0.000 description 2
- 206010003658 Atrial Fibrillation Diseases 0.000 description 2
- 206010006580 Bundle branch block left Diseases 0.000 description 2
- 208000007177 Left Ventricular Hypertrophy Diseases 0.000 description 2
- 206010047281 Ventricular arrhythmia Diseases 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000000747 cardiac effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 201000001715 left bundle branch hemiblock Diseases 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000002861 ventricular Effects 0.000 description 2
- 206010047302 ventricular tachycardia Diseases 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 206010029470 Nodal rhythm Diseases 0.000 description 1
- BQCADISMDOOEFD-UHFFFAOYSA-N Silver Chemical compound [Ag] BQCADISMDOOEFD-UHFFFAOYSA-N 0.000 description 1
- 208000001871 Tachycardia Diseases 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 206010003119 arrhythmia Diseases 0.000 description 1
- 208000006218 bradycardia Diseases 0.000 description 1
- 230000036471 bradycardia Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013467 fragmentation Methods 0.000 description 1
- 238000006062 fragmentation reaction Methods 0.000 description 1
- 208000012727 heart conduction disease Diseases 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 229910052709 silver Inorganic materials 0.000 description 1
- 239000004332 silver Substances 0.000 description 1
- 230000006794 tachycardia 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
- A61B5/72—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
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Public Health (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Medical Informatics (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Pathology (AREA)
- Psychiatry (AREA)
- Animal Behavior & Ethology (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Signal Processing (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Veterinary Medicine (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Heart & Thoracic Surgery (AREA)
Abstract
The invention discloses an automatic electrocardiogram diagnosis method, which comprises the following steps of firstly reading a plurality of heart disease data, obtaining specific electrocardiogram measuring records corresponding to patients according to the heart disease data, and eliminating non-specific electrocardiogram measuring records corresponding to the patients. By using mechanical learning, the corresponding characteristic data is identified in the specific electrocardiogram measurement record corresponding to the patient according to the heart disease data. Then, the electrocardiogram measuring record to be identified is read, and whether the electrocardiogram measuring record to be identified has heart diseases or not is judged by utilizing the characteristic data, so that the accuracy of judging the heart diseases can be effectively improved.
Description
Technical Field
The present invention relates to an electrocardiogram diagnosis method, and more particularly, to an automatic electrocardiogram diagnosis method.
Background
With the increasing progress of science and technology and the progress of medical technology, the life span of human beings is gradually increased. At present, China gradually enters an aging society, the proportion of the population of the old is increased year by year, and heart diseases also become one of the silver hair family first invisible killers.
Cardiac arrhythmia is a very common condition in heart-related diseases, and monitoring of heart-related diseases is an important issue for the elderly. Currently, the heart rate/electrocardiogram detection is mainly monitored by an electrocardiogram measuring device and then interpreted by a doctor.
However, because each individual has a different natural constitution, it is often determined that it is inappropriate to rely on the conventional electrocardiogram comparison method to interpret the electrocardiogram and to use it as a basis for disease diagnosis.
In addition, the conventional way of interpreting the electrocardiogram of the user by using large data analysis not only requires huge data, but also often causes interpretation errors due to data differences, so that the accuracy of electrocardiogram interpretation is reduced. The electrocardiogram judgment result obtained by the computer is also presented in a different way from the doctor judgment result, which also causes troubles for doctors and patients.
Disclosure of Invention
In view of the above, the present invention discloses an automatic electrocardiogram diagnosis method, which can reduce the learning time and software and hardware resources, so as to rapidly and correctly perform disease marking and diagnosis of electrocardiogram, thereby improving the medical quality.
According to one embodiment of the present disclosure, an automatic electrocardiogram diagnosis method includes the steps of reading a plurality of heart disease data, obtaining a specific electrocardiogram measurement record corresponding to a patient according to the heart disease data, and excluding the specific electrocardiogram measurement record corresponding to the patient. By using mechanical learning, the corresponding characteristic data is identified in the specific electrocardiogram measurement record corresponding to the patient according to the heart disease data. Then, the electrocardiogram measuring record to be identified is read, and whether the electrocardiogram measuring record to be identified has heart diseases or not is judged by using the characteristic data.
In some embodiments, the excluding of non-specific electrocardiographic records corresponding to patients is based on big data analysis, excluding electrocardiographic records of leads with low correlation to heart disease data.
In some embodiments, the excluding of the ecg measurement records of leads with low correlation to heart disease data based on big data analysis is recording ecg measurement records of leads with recognition accuracy lower than 50% as non-specific ecg measurement records.
In some embodiments, the excluding the ecg measurement records of leads with low correlation to the heart disease data based on the big data analysis further comprises recording ecg measurement records of leads with an identification accuracy higher than 90% as the specific ecg measurement records.
In some embodiments, the characterization data is used to determine whether the electrocardiographic recording to be evaluated has a heart condition, and the electrocardiographic recording to be evaluated is determined to have a heart condition based only on a particular electrocardiographic recording of the electrocardiographic recordings to be evaluated.
In some embodiments, the use of machine learning further includes tagging the corresponding feature data. Wherein, marking the corresponding characteristic data is to record specific electrocardiogram measurement and perform cutting by the time coordinate of the maximum value of the absolute value of each heartbeat voltage.
In some embodiments, marking the corresponding characteristic data further includes recording the specific electrocardiographic measurement with a cutting position where the maximum value of the absolute value of the heartbeat voltage corresponding to the heart disease data is located as an origin, taking two thirds of the cutting position before and two thirds of the cutting position after the cutting position as a marking interval of the corresponding characteristic data.
In some embodiments, the marking the corresponding feature data further comprises adding a time to the front-back average when the marking interval is less than 2.5 seconds, so that the marking interval is equal to 2.5 seconds.
In some embodiments, marking the corresponding feature data further includes randomly multiplying the time coordinate of the marking interval by any one of values from 0.75 to 1.25 to generate the auxiliary feature data, and when the marking interval of the auxiliary feature data is less than 2.5 seconds, zeroing the subsequent data of the auxiliary feature data to 2.5 seconds.
In some embodiments, identifying the corresponding feature data utilizes a shallow neural network model to identify the corresponding feature data.
In some embodiments, the automated electrocardiographic diagnostic method further comprises classifying the cardiac signals according to heart beat, rhythm and pattern levels and outputting the electrocardiographic recording of the cardiac disease to be identified.
In some embodiments, the automated electrocardiographic diagnosis method further comprises learning with multiple tags, and learning and determining multiple heart diseases simultaneously.
In some embodiments, the automated electrocardiographic diagnosis method further comprises outputting a plurality of heart diseases and corresponding probabilities.
Therefore, the automatic electrocardiogram diagnosis method disclosed by the invention can improve the correctness of learning and judgment, reduce the quantity of electrocardiogram measurement records required, reduce the software and hardware requirements of required operation data, save a large amount of resources such as electric power and the like, facilitate doctors and patients to quickly and correctly know heart diseases, obtain related suggestions and further effectively improve the medical quality. In addition, various heart diseases and corresponding probabilities can be output through multi-label learning so as to be referred by doctors and patients, the accuracy of disease judgment of doctors is effectively improved, and the use will of doctors is further improved.
Drawings
In order to make the aforementioned and other objects, features, and advantages of the present invention comprehensible, embodiments accompanied with figures are described as follows:
fig. 1 is a schematic diagram of an automatic electrocardiographic diagnosis method according to an embodiment of the present invention.
FIG. 2 is a schematic diagram of steps for identifying characteristic data according to the automatic electrocardiogram diagnosis method of FIG. 1.
FIG. 3 is a schematic view of a specific electrocardiographic recording.
Fig. 4 is a schematic diagram of signature data.
Description of the main reference numerals:
100-automatic electrocardiogram diagnosis method, 110-160-step, 210-270-step, 300-specific electrocardiogram measurement record, 310-lead V1 electrocardiogram waveform, 320-positive S Valley waveform, 330-lead II electrocardiogram waveform, 401-block, 410-cutting line, 420-origin, 430-X point, 440-Z point, 450-mark section, 451-front section, 452-rear section.
Detailed Description
The following detailed description of the embodiments with reference to the accompanying drawings is provided for purposes of illustration, and is not intended to limit the scope of the invention, which is defined by the claims, as the description of the illustrated embodiments should not be read as limiting the order in which the features of the embodiments are performed, nor should any structure that results in a structure that is a subcombination of elements, but which has the equivalent effect. In addition, the drawings are for illustrative purposes only and are not drawn to scale. For ease of understanding, the same or similar components will be described with the same reference numerals in the following description.
Further, the term (terms) used throughout the specification and claims has the ordinary meaning as commonly understood by one of ordinary skill in the art to which this term pertains, in the context of this disclosure, and in the context of a particular application, unless otherwise indicated. Certain terms used to describe the invention are discussed below or elsewhere in this specification to provide additional guidance to those skilled in the art in describing the invention.
As used herein, the terms "first," "second," …, etc., are not intended to be limited to the exact order or sequence presented, nor are they intended to be limiting, but rather are intended to distinguish one element from another or from another element or operation described by the same technical term.
Furthermore, as used herein, the terms "comprising," "including," "having," "containing," and the like are open-ended terms that mean including, but not limited to.
Fig. 1 is a schematic diagram of an automatic electrocardiographic diagnosis method according to an embodiment of the present invention, fig. 2 is a schematic diagram of steps of identifying characteristic data according to the automatic electrocardiographic diagnosis method of fig. 1, fig. 3 is a schematic diagram of a specific electrocardiographic measurement record, and fig. 4 is a schematic diagram of marking characteristic data.
As shown in fig. 1, an automatic electrocardiographic diagnosis method 100 comprises the following steps, firstly, step 110, reading heart disease data, and then, step 120, obtaining a specific electrocardiographic measurement record corresponding to a patient according to the heart disease data, and excluding a non-specific electrocardiographic measurement record corresponding to the patient. Wherein, the specific electrocardiographic measurement record refers to electrocardiographic data of a plurality of leads which are related to the heart disease patient.
Referring to fig. 3, a diagram of a specific ecg measurement record 300, when the heart condition is Right Bundle Branch Block (RBBB), the ecg data of 12 leads measured simultaneously only produce significant waveform characteristics in the ecg waveforms at the lead V1 and the lead V6, such as the positive S Valley waveform 320 in the ecg waveform 310 at the lead V1, but have no significant or corresponding change in the ecg data at other leads, such as the ecg waveform 330 at the lead II. In this case, when the subsequent mechanical learning process and the determination of the heart disease are performed in the conventional manner, the learning and determination accuracy is reduced due to the influence of other ecg lead data.
Thus, the automated ECG diagnostic method 100 can exclude such ECG data from being recorded as non-specific ECG records, not only before learning, but also when subsequently determining the ECG record to be identified. Therefore, not only the learning accuracy and the judgment accuracy can be effectively improved, but also the conventional mechanical learning needs to read 12 leads of ECG data, and the automatic ECG diagnosis method 100 only needs to read the specific ECG record of the patient, so the amount of ECG data required in the learning process is greatly reduced.
Taking Right Bundle Branch Block (RBBB) as an example, only 2 leads of electrocardiogram data, namely, the leads V1 and V6 of electrocardiogram data, need to be determined, so as to effectively learn feature data, and in the subsequent determination, when determining whether there is a heart disease with Right Bundle Branch Block (RBBB), only two leads of electrocardiogram data need to be determined, and the interference of other leads of electrocardiogram data is eliminated, so as to not only speed up the determination efficiency, but also improve the determination accuracy and the automatic electrocardiogram diagnosis accuracy.
In some embodiments, a specific electrocardiographic measurement record corresponding to the patient is obtained according to the heart disease data, and a non-specific electrocardiographic measurement record corresponding to the patient is excluded, so that the screening can be performed according to the setting of the doctor.
In some embodiments, a specific electrocardiographic measurement record corresponding to the patient is obtained and a non-specific electrocardiographic measurement record corresponding to the patient is excluded according to the heart disease data, so that the electrocardiographic data of some heart diseases can be set as the specific electrocardiographic measurement record and other electrocardiographic data can be set as the non-specific electrocardiographic measurement record and excluded through big data analysis.
In some embodiments, the ecg measurement records for leads with low correlation to heart disease data are excluded from the big data analysis by recording ecg measurement records for leads with an identified accuracy of less than 50% as non-specific ecg measurement records, and may also be recorded for leads with an identified accuracy of greater than 90% as specific ecg measurement records.
Next, in step 130, machine learning is used to identify corresponding characteristic data in the specific ecg measurement record corresponding to the patient according to the heart disease data. In which, steps 110 to 130 are steps of learning feature data, which can be continuously and repeatedly performed to improve the accuracy of the determination.
Referring to fig. 2 and 4, as shown in the figures, the mechanical learning of the identification feature data of step 130 further includes marking the corresponding feature data, and first, step 210, a specific ecg recording is cut, which is to cut the specific ecg measurement recording with the time coordinate of the maximum value of the absolute value of each heartbeat voltage.
Referring to fig. 4, the characteristic data is labeled with the recorded ventricular early contraction (PVC) of a specific electrocardiograph measurement. The electrocardiogram is divided into a plurality of blocks 401 by using the cutting lines 410 as boundaries. In step 220, the cutting line 410 of the heartbeat voltage corresponding to the heart disease data is used as the origin 420. Next, step 230, two thirds of the distance from the previous cutting line 410, such as the X point 430 in the figure, to the origin 420 is the starting point of the mark section 450, and step 240, two thirds of the distance from the next cutting line 410, such as the Z point 440 in the figure, to the origin 420 is the end point of the mark section 450, so as to form the mark section 450 enclosed by the dashed line in the middle of the figure.
In some embodiments, when the length of the mark section 450 is less than 2.5 seconds, the time is increased by an average amount to expand the time length of the mark section 450 to 2.5 seconds. Referring to step 250, a front zone 451 and a back zone 452 are added to increase the time for marking zone 450 to 2.5 seconds. Therefore, the mark interval 450 can effectively cover the data and time of the whole feature data, and effectively avoid the distortion or fragmentation of the feature data caused by the conventional standard-length cutting method.
In some embodiments, identifying the feature data may further include generating supplementary feature data 260, which may generate supplementary feature data using random numbers, thereby increasing the effective data of the machine learning process.
For example, the time coordinate of the original data may be compressed or expanded to any value of 0.75 to 1.25, for example, the original time coordinate (1+ (random value (0-1) -0.5)/2), so as to compress or expand the time coordinate of the original data, thereby generating the auxiliary feature data, increasing the effective samples of the mechanical learning, and further improving the accuracy of the identification.
In some embodiments, when the mark interval of the auxiliary feature data is less than 2.5 seconds, the subsequent data of the auxiliary feature data is zero-padded to 2.5 seconds.
In some embodiments, when the mark interval of the auxiliary feature data is greater than 2.5 seconds, the first 2.5 seconds of data are taken as auxiliary feature data.
In addition, in step 270, the specific electrocardiogram record is analyzed by using the shallow neural network model to improve the accuracy of the identification. Since the automatic electrocardiogram diagnosis method 100 employs a specific electrocardiogram record for learning and identification, the learning complexity can be simplified, and the identification accuracy can be effectively improved without employing deep learning. In addition, since the automatic electrocardiogram diagnosis method 100 uses a specific electrocardiogram record for learning and identification, it is able to achieve high-precision identification capability with a small amount of calculation, effectively reduce the requirements of software and hardware required for calculation, and save a large amount of resources such as power.
Referring again to fig. 1, step 140 reads the ecg measurement record to be identified, and step 150 determines whether the ecg measurement record to be identified has heart disease using the characteristic data. When the determination is made, the automatic electrocardiographic diagnosis method 100 further uses the specific electrocardiographic measurement record to make the determination when the possible heart disease is identified, and excludes the specific electrocardiographic measurement record, so as to further improve the accuracy of the determination of the heart disease without being affected by the non-specific electrocardiographic measurement record and reduce the accuracy of the determination.
Then, in step 160, a diagnosis result is outputted, which includes the heart disease data recorded by electrocardiography to be identified in a manner of hierarchical classification according to the heart beat, rhythm and pattern. The heartbeat classification may include results of normal heartbeat, tachycardia, or bradycardia. The Rhythm-related classifications include Normal Sinus Rhythm (Normal Sinus Rhythm), Sinus Arrhythmia (Sinus Arrhythmia), Atrial Rhythm (Atrial Rhythm), Atrial Fibrillation (Atrial Fibrillation), Sinus bradycardia (Sinus bradycardia), junction Rhythm (Junctional Rhythm), Ventricular Rhythm (Ventricular Rhythm), Ventricular Tachycardia (Ventricular Tachycardia) or cardiac Rhythm (Ventricular Pacing Rhythm). The type classification includes ST change (ST change), T-wave inversion (T-wave inversion), Right Bundle Branch Block (RBBB), Left Bundle Branch Block (LBBB), Myocardial Infarction (Myocardial Infarction), Left ventricular hyper-trophy (Left ventricular hypertrophy), and other heart diseases.
In some embodiments, the automated electrocardiographic diagnosis method 100 may further output the occurrence probability of each heart disease through Multi-Label Learning (Multi-Label Learning), adjust the probability threshold according to the requirement to control the specificity (specificity) and sensitivity (sensitivity) of the heart disease, and finally output the possible rhythm and type of the electrocardiogram from the probability distribution.
For example, multiple cardiac diseases can be simultaneously learned and judged through multi-label learning, and multiple possible cardiac diseases and corresponding probabilities thereof are output for reference of doctors and patients by adjusting probability threshold values. For example, when the probability threshold is set at 75%, when the probability of the electrocardiographic data is judged to have heart rhythm diseases, such as Sinus Arrhythmia (Sinus Arrhythmia) and Sinus bradycardia (Sinus bradycardia), is greater than 75%, the diagnosis result is output to further output the probability that the patient has Sinus Arrhythmia and Sinus bradycardia correspond to the Sinus Arrhythmia. When the electrocardiogram data is interpreted to show that the probability of heart-type diseases such as T-wave inversion (T-wave inversion) and Myocardial Infarction (Myocardial Infarction) is more than 75%, the diagnosis result is output to further output the probability of T-wave inversion and Myocardial Infarction corresponding to the T-wave inversion. Therefore, the doctor can judge the heart disease and physical condition of the patient more clearly and accurately.
Therefore, the output of the related diagnosis results is closer to the diagnosis manner of the doctor, so that the diagnosis results of the automatic electrocardiographic diagnosis method 100 are closer to the diagnosis results of the doctor than the combination of the possibilities of the heart diseases, but the diagnosis accuracy cannot be improved.
In summary, the automatic electrocardiogram diagnosis method disclosed by the present invention can improve the correctness of learning and judgment, reduce the number of electrocardiogram measurement records required, and reduce the software and hardware requirements of the required operation data, thereby saving a large amount of resources such as power, facilitating doctors and patients to quickly and correctly know heart diseases and obtain related suggestions, and further effectively improving the quality of medical treatment. In addition, various heart diseases and corresponding probabilities can be output through multi-label learning so as to be referred by doctors and patients, the accuracy of disease judgment of doctors is effectively improved, and the use will of doctors is further improved.
Although the present invention has been described with reference to the above embodiments, it should be understood that various changes and modifications can be made therein by those skilled in the art without departing from the spirit and scope of the invention.
Claims (14)
1. An automated electrocardiographic diagnostic method comprising:
reading a plurality of heart disease data;
obtaining a specific electrocardiogram measuring record of a corresponding patient according to the plurality of heart disease data, and excluding a non-specific electrocardiogram measuring record of the corresponding patient;
identifying corresponding characteristic data in the specific electrocardiogram measurement record of the corresponding patient according to the plurality of heart disease data by using mechanical learning;
reading an electrocardiogram measuring record to be identified; and
and judging whether the electrocardiogram measuring record to be identified has heart diseases or not by using the plurality of characteristic data.
2. The automated electrocardiographic diagnostic method according to claim 1, wherein said excluding of said non-specific electrocardiographic recording of said corresponding patient is based on big data analysis to exclude electrocardiographic recording of leads with lower correlation of said plurality of heart disease data.
3. The automated electrocardiographic diagnostic method according to claim 2, wherein said excluding from said plurality of electrocardiographic measurements of leads with low correlation to heart disease data based on big data analysis is recording electrocardiographic measurements of leads with accuracy of identification lower than 50% as non-specific electrocardiographic measurements.
4. The automated electrocardiographic diagnostic method according to claim 3 wherein said excluding from said lead electrocardiographic recording of said plurality of cardiac disease data based on big data analysis further comprises: and recording the electrocardiogram measuring record with the lead identification accuracy rate higher than 90% as a specific electrocardiogram measuring record.
5. The automated electrocardiographic diagnostic method according to claim 4, wherein said determining whether said electrocardiographic measurement record to be evaluated has a heart disease using said plurality of characteristic data is performed based only on a specific electrocardiographic measurement record of said electrocardiographic measurement records to be evaluated.
6. The automated electrocardiographic diagnostic method according to claim 1 wherein said utilizing mechanical learning further comprises labeling said corresponding characteristic data.
7. The automated electrocardiographic diagnostic method according to claim 6, wherein said marking of said corresponding characteristic data is a recording of said plurality of specific electrocardiographic measurements, and wherein said cutting is performed with a time coordinate of a maximum value of an absolute value of each heartbeat voltage.
8. The automated electrocardiographic diagnostic method according to claim 7 wherein said marking of said corresponding characteristic data further comprises: and taking the cutting position where the maximum value of the absolute value of the heartbeat voltage corresponding to the heart disease data is located as an origin, taking two thirds of the cutting position in the front direction and two thirds of the cutting position in the back direction as a mark interval of the corresponding characteristic data.
9. The automated electrocardiographic diagnostic method according to claim 8 wherein said marking of said corresponding characteristic data further comprises: when the mark interval is less than 2.5 seconds, a time is added to the front and back average so that the mark interval is equal to 2.5 seconds.
10. The automated electrocardiographic diagnostic method according to claim 9 wherein said marking of said corresponding characteristic data further comprises: randomly multiplying the mark interval by any value from 0.75 to 1.25 to generate auxiliary feature data, and when the mark interval of the auxiliary feature data is less than 2.5 seconds, zero-filling the subsequent data of the auxiliary feature data to 2.5 seconds.
11. The automated electrocardiographic diagnostic method according to claim 1 wherein the identifying the corresponding feature data is performed using a model of a superficial neural network to identify the plurality of corresponding feature data.
12. The automated electrocardiographic diagnostic method according to claim 1, further comprising: classifying by heart beat, rhythm and pattern, and outputting the heart disease recorded by the electrocardiogram measurement to be identified.
13. The automated electrocardiographic diagnostic method according to claim 12, further comprising: and the multi-label learning is utilized, and various heart diseases are learned and judged simultaneously.
14. The automated electrocardiographic diagnostic method according to claim 13, further comprising: a plurality of heart diseases and corresponding probabilities are output.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010119455.0A CN113303805A (en) | 2020-02-26 | 2020-02-26 | Automatic electrocardiogram diagnosis method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010119455.0A CN113303805A (en) | 2020-02-26 | 2020-02-26 | Automatic electrocardiogram diagnosis method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113303805A true CN113303805A (en) | 2021-08-27 |
Family
ID=77370698
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010119455.0A Pending CN113303805A (en) | 2020-02-26 | 2020-02-26 | Automatic electrocardiogram diagnosis method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113303805A (en) |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1032288A (en) * | 1987-09-30 | 1989-04-12 | 创建基立有限公司 | The method and apparatus of electrocardiosignal multi-domain processing and correlation analysis |
KR20060117546A (en) * | 2005-05-11 | 2006-11-17 | 인하대학교 산학협력단 | Apparatus and method for a cardiac diseases diagnoses based on ecg using neural network |
CN102779234A (en) * | 2012-06-28 | 2012-11-14 | 中国科学院苏州纳米技术与纳米仿生研究所 | Electrocardiogram classified processing method and device |
CN106326644A (en) * | 2016-08-16 | 2017-01-11 | 沈阳东软熙康医疗系统有限公司 | Method and device for calculating HRV (heart rate variability) parameters and fatigue indexes |
CN107103182A (en) * | 2017-03-28 | 2017-08-29 | 南京医科大学 | A kind of heart disease Warning System and method based on deep learning algorithm |
CN107661097A (en) * | 2016-06-14 | 2018-02-06 | 美商宇心生医股份有限公司 | Electrocardiogram measuring method |
CN108030494A (en) * | 2017-11-08 | 2018-05-15 | 华南理工大学 | Electrocardiosignal error flag training sample recognition methods based on cross validation |
CN108836313A (en) * | 2018-07-13 | 2018-11-20 | 希蓝科技(北京)有限公司 | A kind of intelligence chooses the ambulatory ECG analysis method of lead |
CN108836314A (en) * | 2018-07-13 | 2018-11-20 | 希蓝科技(北京)有限公司 | A kind of ambulatory ECG analysis method and system based on network and artificial intelligence |
CN109770889A (en) * | 2017-11-15 | 2019-05-21 | 深圳市理邦精密仪器股份有限公司 | Electrocardiogram (ECG) data selections method and apparatus |
CN110051345A (en) * | 2018-01-18 | 2019-07-26 | 美商宇心生医股份有限公司 | Electrocardiographic diagnosis suggesting method |
CN110236529A (en) * | 2019-07-19 | 2019-09-17 | 浙江好络维医疗技术有限公司 | A kind of multi-lead arrhythmia cordis intelligent diagnosing method based on MODWT and LSTM |
CN110464339A (en) * | 2019-07-05 | 2019-11-19 | 深圳邦健生物医疗设备股份有限公司 | A kind of ecg analysis method and device |
-
2020
- 2020-02-26 CN CN202010119455.0A patent/CN113303805A/en active Pending
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1032288A (en) * | 1987-09-30 | 1989-04-12 | 创建基立有限公司 | The method and apparatus of electrocardiosignal multi-domain processing and correlation analysis |
KR20060117546A (en) * | 2005-05-11 | 2006-11-17 | 인하대학교 산학협력단 | Apparatus and method for a cardiac diseases diagnoses based on ecg using neural network |
CN102779234A (en) * | 2012-06-28 | 2012-11-14 | 中国科学院苏州纳米技术与纳米仿生研究所 | Electrocardiogram classified processing method and device |
CN107661097A (en) * | 2016-06-14 | 2018-02-06 | 美商宇心生医股份有限公司 | Electrocardiogram measuring method |
CN106326644A (en) * | 2016-08-16 | 2017-01-11 | 沈阳东软熙康医疗系统有限公司 | Method and device for calculating HRV (heart rate variability) parameters and fatigue indexes |
CN107103182A (en) * | 2017-03-28 | 2017-08-29 | 南京医科大学 | A kind of heart disease Warning System and method based on deep learning algorithm |
CN108030494A (en) * | 2017-11-08 | 2018-05-15 | 华南理工大学 | Electrocardiosignal error flag training sample recognition methods based on cross validation |
CN109770889A (en) * | 2017-11-15 | 2019-05-21 | 深圳市理邦精密仪器股份有限公司 | Electrocardiogram (ECG) data selections method and apparatus |
CN110051345A (en) * | 2018-01-18 | 2019-07-26 | 美商宇心生医股份有限公司 | Electrocardiographic diagnosis suggesting method |
CN108836313A (en) * | 2018-07-13 | 2018-11-20 | 希蓝科技(北京)有限公司 | A kind of intelligence chooses the ambulatory ECG analysis method of lead |
CN108836314A (en) * | 2018-07-13 | 2018-11-20 | 希蓝科技(北京)有限公司 | A kind of ambulatory ECG analysis method and system based on network and artificial intelligence |
CN110464339A (en) * | 2019-07-05 | 2019-11-19 | 深圳邦健生物医疗设备股份有限公司 | A kind of ecg analysis method and device |
CN110236529A (en) * | 2019-07-19 | 2019-09-17 | 浙江好络维医疗技术有限公司 | A kind of multi-lead arrhythmia cordis intelligent diagnosing method based on MODWT and LSTM |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109171712B (en) | Atrial fibrillation identification method, atrial fibrillation identification device, atrial fibrillation identification equipment and computer readable storage medium | |
Babaeizadeh et al. | Improvements in atrial fibrillation detection for real-time monitoring | |
US9314177B2 (en) | System and method of detecting abnormal movement of a physical object | |
CN108937915B (en) | Method for identifying premature beat in dynamic electrocardiogram | |
WO2021037102A1 (en) | Electrocardiogram analysis method and apparatus based on picture and heartbeat information, and device and medium | |
CN106815570B (en) | Electrocardiosignal ST-T segment identification method based on dynamic pattern identification | |
US20100217144A1 (en) | Diagnostic and predictive system and methodology using multiple parameter electrocardiography superscores | |
CN112237431A (en) | Electrocardio parameter calculation method based on deep learning | |
Zhou et al. | ECG quality assessment using 1D-convolutional neural network | |
US8750973B2 (en) | Method and system for detecting P-waves in the surface ECG signal | |
Jones et al. | Improving ECG classification interpretability using saliency maps | |
Rohmantri et al. | Arrhythmia classification using 2D convolutional neural network | |
CN116503673B (en) | Arrhythmia identification and detection method and system based on electrocardiogram | |
Kuila et al. | Feature extraction and classification of MIT-BIH arrhythmia database | |
Zimmerman et al. | A reconstructed phase space approach for distinguishing ischemic from non-ischemic ST changes using Holter ECG data | |
TWI727678B (en) | Automatic electrocardiogram data processing method | |
WO2024098553A1 (en) | Method and system for analyzing and identifying electrocardiogram, and storage medium | |
TWI827239B (en) | Method for assisting identification of symptoms | |
Donnelly et al. | Lead selection: old and new methods for locating the most electrocardiogram information | |
CN108836312B (en) | Clutter rejection method and system based on artificial intelligence | |
CN113303805A (en) | Automatic electrocardiogram diagnosis method | |
CN113940682B (en) | Atrial fibrillation recognition method based on statistical characteristics | |
KR102437348B1 (en) | Method for wearable ECG signal analysis | |
KR102605130B1 (en) | Method for classifying cardiac arrhythmia from standard 12-lead Electrocardiogram signal using artificial neural network and classification device using the same | |
Bashir et al. | Highlighting the current issues with pride suggestions for improving the performance of real time cardiac health monitoring |
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 |