CN110141214A - A kind of mask method of electrocardiogram identification and its application - Google Patents
A kind of mask method of electrocardiogram identification and its application Download PDFInfo
- Publication number
- CN110141214A CN110141214A CN201910327442.XA CN201910327442A CN110141214A CN 110141214 A CN110141214 A CN 110141214A CN 201910327442 A CN201910327442 A CN 201910327442A CN 110141214 A CN110141214 A CN 110141214A
- Authority
- CN
- China
- Prior art keywords
- electrocardiogram
- callout box
- heartbeat
- training
- recognition
- 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
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
-
- 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
- A61B5/349—Detecting specific parameters of the electrocardiograph cycle
- A61B5/352—Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Abstract
The present invention relates to a kind of mask method of electrocardiogram identification and its applications.The present invention is simple, easy using electrocardiogram identification mask method, all callout box cover the complete information of electrocardiogram, it does not interspace between the callout box of adjacent cardiac, the demands such as percentage of head rice, the entire electrocardiogram of covering, the promotion of machine algorithm distinguishing indexes and machine recognition data-handling efficiency to take into account heartbeat message interception, obtain the comprehensive of various aspects of performance and are promoted.
Description
Technical field
The present invention relates to field of image recognition, and in particular to a kind of mask method of electrocardiogram identification and its answers in the field
With.
Background technique
The starting that electrocardiogram (ECG or EKG) has recorded starting to the heartbeat next time of heart from heartbeat each time is produced
Raw electrical activity variation, electrocardiogram is one of clinical most common electrophysiologic study, using very extensive.The image of electrocardiogram
The main purpose of identification and analytical technology is that the electrocardiogram (ECG) data of human body is automatically analyzed and handled, to provide to doctor
The judgment basis of auxiliary diagnosis.
It in the development in each field and popularizes along with technologies such as machine deep learning, neural network algorithms, has produced at present
The approach using neural network algorithm (such as CNN, DNN) training, analyzing ecg signals is given birth to.Relative to traditional artificial reading
For figure, efficiency, in terms of achieve beneficial progress.For example, using a large amount of electrocardio picture construction training sets, testing
Card collection and the set such as test set, further using above-mentioned set implement image labeling, model training, model measurement, image recognition,
Model evaluation and etc. and be finally applied to actually detected.
In the identification of above-mentioned machine algorithm, sensitivity, specificity, accuracy, recall rate, F1 value and accuracy parameter
It is the important indicator for evaluating discrimination efficiency.Inventor has found under study for action, and the Reasonable and mark of electrocardiogram segment be (interception
Electrocardiogram segment is also referred to as callout box) be influence These parameters parameter an important factor for one of.Existing automatic electrocardiogram is known
Other mask method generallys use centered on R wave position, and the electrocardiographic wave conduct of regular length is intercepted before and after center
Callout box, the selection of regular length are often based upon artificial experience, and need constantly to be adjusted according to the length of different heartbeats
It is whole, thus method is cumbersome is unfavorable for automating.
In addition, the general interception for thinking callout box of common viewpoint need to only make its packet in the interception range of callout box
Information containing a heartbeat can (such as P, Q, R, S, T characteristic wave).However, how appropriately designed section under the premise of machine recognition
Range is taken to ensure the percentage of head rice of most heartbeat message interceptions, and whether the ECG information except a heartbeat message is to machine
There are benifit, Shang Xianyou further investigateds for the index parameter promotion of device algorithm identification.
Summary of the invention
The purpose of the present invention is for defect present in background technique, the method for improving existing electrocardiogram image labeling,
A kind of electrocardiogram image labeling method simple and easy to do, recognition performance is excellent is provided.
The present invention provides a kind of mask method for electrocardiogram identification, in the corresponding image position of each heartbeat of electrocardiogram
It sets place and delimit callout box, characterized by comprising:
(1) the peak value high point position of current heartbeat R wave, previous heartbeat R wave and latter heartbeat R wave is positioned；
(2) centered on current heartbeat R crest value high point position, with it between previous heartbeat R crest value high point position
Left margin at the c of horizontal distance as the callout box；
(3) centered on current heartbeat R crest value high point position, with it between latter heartbeat R crest value high point position
Right margin at the d of horizontal distance as the callout box；
Described c, d meet:
1/3 < c≤2/3, and 1/3 < d≤2/3.
Further, 1/2≤c≤2/3, and 1/2≤d≤2/3
Further, the c=d=1/2.
Further, the delimitation of the callout box further includes the delimitation step on upper and lower boundary:
(1) using horizontal line locating for current heartbeat R crest value high point position as the coboundary of the callout box；
(2) with current heartbeat Q crest value low dot location and S crest value low dot location, the lower one institute in position in the two
Locate lower boundary of the horizontal line as the callout box.
The present invention also provides a kind of electrocardiogram recognition training methods, successively include the raw data set for obtaining electrocardiogram, figure
The step of being assessed as mark, model training and test, it is characterised in that: described image annotation step uses above-mentioned mark side
Method.
Further, the training pattern that the model training step uses is Faster R-CNN model.
The present invention also provides a kind of electrocardiogram recognition methods, if including carrying out image labeling formation to electrocardiogram to be identified
The step of dry callout box image, and the step of machine recognition is executed to several callout box images；It is characterized by: utilizing
Above-mentioned mask method carries out image labeling to electrocardiogram to be identified to form several callout box images.
It is described the present invention also provides a kind of electrocardiogram identification device, including electrocardiogram labeling module and picture recognition module
Electrocardiogram labeling module is used to carry out image labeling to electrocardiogram to be identified to form several callout box images；Described image identification
Module is used to execute machine recognition to several callout box images；It is characterized by: the electrocardiogram labeling module is using upper
The mask method stated carries out image labeling to electrocardiogram to be identified to form several callout box images.
Further, the electrocardiogram identification device further includes image training module, and described image training module utilizes upper
Electrocardiogram recognition training method is stated, recognition training is carried out to described image identification module.
The invention has the following advantages:
(1) electrocardiogram of the invention identification callout box can ensure the percentage of head rice of higher multiple groups heartbeat message interception；
(2) electrocardiogram of the invention identification callout box can not only include the complete information of a heartbeat, additionally it is possible to including
The complete information of entire electrocardiogram can obtain sensitivity, specificity, accuracy, recall rate, F1 in machine algorithm identification
The promotion of value, accuracy index parameter；
(3) callout box demarcation method of the invention can take into account percentage of head rice, the entire electrocardio of covering of heartbeat message interception
Figure, machine algorithm distinguishing indexes are promoted and demands, the synthesis for obtaining various aspects of performance such as machine recognition data-handling efficiency mention
It rises.
Detailed description of the invention
Fig. 1 is the mask method schematic diagram of electrocardiogram of the present invention identification；
Specific embodiment
Technical solution of the present invention is clearly and completely described below in conjunction with attached drawing, it is clear that described implementation
Example is a part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill
Personnel's every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
In the description of the present invention, it should be noted that term " center ", "upper", "lower", "left", "right", "vertical",
The orientation or positional relationship of the instructions such as "horizontal", "inner", "outside" be based on the orientation or positional relationship shown in the drawings, merely to
Convenient for description the present invention and simplify description, rather than the device or element of indication or suggestion meaning must have a particular orientation,
It is constructed and operated in a specific orientation, therefore is not considered as limiting the invention.In addition, term " first ", " second ",
" third " is used for descriptive purposes only and cannot be understood as indicating or suggesting relative importance.
In the description of the present invention, it should be noted that unless otherwise clearly defined and limited, term " installation ", " phase
Even ", " connection " shall be understood in a broad sense, for example, it may be being fixedly connected, may be a detachable connection, or be integrally connected；It can
To be mechanical connection, it is also possible to be electrically connected；It can be directly connected, can also can be indirectly connected through an intermediary
Connection inside two elements.For the ordinary skill in the art, above-mentioned term can be understood at this with concrete condition
Concrete meaning in invention.
As shown in Fig. 1, electrocardiogram generally include several heartbeat images and its between gap.Each heartbeat image is usual
Comprising P, Q, R, S, T wave, 5 characteristic waves of heartbeat image are respectively represented, wherein P wave is Atrial depolarization wave, represents left and right disloyalty
The excitement in room；QRS complex reflects that Ventricular removes the variation of electrode potential and time, and first downward wave is Q wave, upward
Wave is R wave, and then downward wave is S wave；T wave is the process of ventricular bipolar.
(1) boundary of callout box delimited
The present invention is used for the mask method of electrocardiogram identification, is included at each heartbeat picture position of electrocardiogram and delimit mark
Frame.Referring to attached drawing 1, three heartbeat images are shown in figure, here by taking the callout box of intermediate heartbeat image delimited as an example, illustrate this hair
Bright main annotation step:
(1) the peak value high point position of current heartbeat R wave, previous heartbeat R wave and latter heartbeat R wave is positioned；
(2) centered on current heartbeat R crest value high point position, with it between previous heartbeat R crest value high point position
Left margin at the c of horizontal distance as the callout box；
(3) centered on current heartbeat R crest value high point position, with it between latter heartbeat R crest value high point position
Right margin at the d of horizontal distance as the callout box；
Thus can with and so on, in fig. 1 left and right two heartbeat picture positions at delimit callout box accordingly.
Callout box of the invention can be determined by only dividing left and right side frame.Since there is no front and backs in upper down space
The extension of heartbeat, thus up-and-down boundary of the up-and-down boundary as callout box of electrocardiogram can be directly used, without volume
Outer delimitation.It is found due to filtering, signal fluctuation, steady noise etc. however, passing through research, sometimes the up-and-down boundary of electrocardiogram
Interior to be mixed into the noise information for being not belonging to heartbeat image, this causes bad shadow to the parameter index of machine recognition callout box image
It rings, therefore present invention preferably comprises the delimitation steps on upper and lower boundary to overcome the problems, such as this:
(1) using horizontal line locating for current heartbeat R crest value high point position as the coboundary of the callout box；
(2) with current heartbeat Q crest value low dot location and S crest value low dot location, the lower one institute in position in the two
Locate lower boundary of the horizontal line as the callout box.
(2) research of boundary value
On the one hand the interception of each heartbeat message is difficult to if research finds that its value is too small about the value of above-mentioned c, d
Ensure good percentage of head rice.Because the expanded width of each heartbeat is not the same, particularly with suffering from heart disease crowd's
More very, this will result directly in the accuracy rate and reliability of subsequent machine recognition to electrocardiogram；Another aspect callout box can include phase
The range in gap is too small between adjacent heartbeat image, to influence several index parameters of machine algorithm identification.Such as when c, d value etc.
When 1/3, although being still able to maintain good each heartbeat message percentage of head rice, it includes gap information it is on the low side, it is difficult to guarantee
Excellent machine algorithm distinguishing indexes parameter.Therefore, the value of the present invention preferred c, d is all larger than 1/3.
In addition, if research discovery c, d value is excessive, although can guarantee that all callout box can cover the complete of entire electrocardiogram
(i.e. each callout box image also contains between adjacent cardiac image whole information other than comprising its corresponding heartbeat image
Gap), but may cause the callout box image large area overlapping of two neighboring heartbeat, so that electrocardio pattern recognition device be caused to exist
Data processing more consumes resource and efficiency and reduces (especially for being integrated with the portable of embedded system or micro-system
For electrocardiogram identification device, treatment effeciency and resource occupation are especially valuable), if in addition c, d value are excessive may also make currently
The image information of adjacent cardiac is included in the callout box of heartbeat image, so that the accuracy rate and reliability to machine algorithm identification are made
At adverse effect.Therefore, the value of the present invention preferred c, d, which is respectively less than, is equal to 2/3.
Amid all these factors, the present invention is it is even more preferred that enable 1/2≤c≤2/3, and 1/2≤d≤2/3.It invents simultaneously
People can obtain the percentage of head rice of good heartbeat message interception and will not introduce substantially it has furthermore been found that as c=d=1/2
The image information of adjacent cardiac, at the same can under the premise of ensuring that all callout box cover the complete information of entire electrocardiogram,
Will not consumption resource to electrocardio pattern recognition device and efficiency cause additional effect.Therefore, using c=d=1/2 as the present invention
Highly preferred set-up mode.
(3) research that electrocardiogram completely covers
Different from traditional view think callout box interception only need to make it includes the information of a heartbeat can (such as P, Q, R,
S, T characteristic wave).It is a discovery of the invention that if all callout box can be enable to cover the complete information of entire electrocardiogram, even often
One callout box image also contains the gap between adjacent cardiac image, two phases other than comprising its corresponding heartbeat image
No longer there is gap, the sensitivity that can identify in machine algorithm, specificity, accuracy, recall rate, F1 between adjacent callout box image
In value, accuracy index parameter, acquisition more preferably promotes result.
It is illustrated below using the comparative example of c=d=1/3 as the preferred solution of the invention c=d=1/2.
Comparative example and the preferred solution of the invention using identical electrocardiogram raw data set, and are based in addition to c, d value
The identification device of same hardware configuration and identical recognizer.Test result is as follows for it shown in table 1,2:
Table 1
Table 2
Wherein, NSR, VT, IVR, VFL, Fusion, LBBBB, RBBBB, SDHB, PR, APB, AFL, AFIB, SVTA, WPW,
PVC, Bigeminy, Trigeminy indicate respectively 17 kinds of electrocardiograms common in this field.
In the electrocardiogram of above-mentioned 17 classification, except in two class of RBBBB, PVC the preferred solution of the invention on 6 index parameters
It is inferior in other 14 class electrocardiograms of comparative example in various degree due to comparative example, furthermore performance is consistent both in SDHB.Cause
And in terms of the electrocardiogram identification of most of common class, 6 index parameter of the preferred solution of the invention in machine recognition
On, it to integrate whole better than comparative example.
Above-mentioned 6 index parameters are machine learning field general parameter, are machine about its specific assignment and conversion process
Device learning areas is existing and common expression, and repeats no more herein, is only represented meaning below and shown:
(1) sensitivity: Sensitivity=TP/ (TP+FN)
(2) specific: Specificity=TN/ (FP+TN)
(3) accurate rate: Precision=TP/ (TP+FP)
(4) recall rate: Recall=TP/ (TP+FN)
(5) F1 score: F1Score=2 × Precision × recall/ (Precision+recall)
(6) accuracy rate: Accuracy=(TP+TN)/(TP+FP+FN+TN)
Wherein:
TP indicates true positives (True Positive), represents recognition result and when truth belongs to certain identification types
The case where；TN indicates true negative (True Negative), represents recognition result and when truth is not admitted to certain identification types
The case where；FP indicates false positive (False Positive), represents that recognition result belongs to certain identification types but truth does not belong to
The case where when Mr. Yu's identification types；FN indicates false negative (False Negative), represents recognition result and is not belonging to certain identification class
The case where when type but truth belong to certain identification types.
(4) application of mask method
The present invention also provides a kind of electrocardiogram recognition training methods, successively include the raw data set for obtaining electrocardiogram, figure
The step of being assessed as mark, model training and test, it is characterised in that: described image annotation step uses above-mentioned mark side
Method.
Further, the training pattern that the model training step uses is Faster R-CNN model.
The present invention also provides a kind of electrocardiogram recognition methods, if including carrying out image labeling formation to electrocardiogram to be identified
The step of dry callout box image, and the step of machine recognition is executed to several callout box images；It is characterized by: utilizing
Above-mentioned mask method carries out image labeling to electrocardiogram to be identified to form several callout box images.It specifically can be with
Include:
(1) raw data set of electrocardiogram is obtained, usually mat formatted data, then raw data set is converted into JPEG
The graphic file of format；
(2) graphic file of jpeg format is divided into training set, verifying collection, test set, wherein collect to training set and verifying
Figure takes image labeling method of the invention to carry out image labeling；
(3) image labeling result input training pattern is trained, the training pattern uses Faster R-CNN mould
Type constructs image recognition test model, test set figure input test model is obtained test result；
(4) finally test result is assessed using assessment, evaluation index includes accuracy, sensitivity and specificity etc.
Index parameter.
It is described the present invention also provides a kind of electrocardiogram identification device, including electrocardiogram labeling module and picture recognition module
Electrocardiogram labeling module is used to carry out image labeling to electrocardiogram to be identified to form several callout box images；Described image identification
Module is used to execute machine recognition to several callout box images；It is characterized by: the electrocardiogram labeling module is using upper
The mask method stated carries out image labeling to electrocardiogram to be identified to form several callout box images.
Further, the electrocardiogram identification device further includes image training module, and described image training module utilizes upper
Electrocardiogram recognition training method is stated, recognition training is carried out to described image identification module.
Above-mentioned specific embodiment is the preferable embodiment of the present invention, but embodiments of the present invention are not by above-mentioned reality
Apply the limitation of example, it is other it is any without departing from the spirit and principles of the present invention made by change, modification, substitution, combination,
Simplify, should be equivalent substitute mode, be included within the scope of the present invention.
Claims (9)
1. a kind of mask method for electrocardiogram identification, delimit mark at the corresponding picture position of each heartbeat of electrocardiogram
Frame, characterized by comprising:
(1) the peak value high point position of current heartbeat R wave, previous heartbeat R wave and latter heartbeat R wave is positioned；
(2) it centered on current heartbeat R crest value high point position, is arrived with it horizontal between previous heartbeat R crest value high point position
Left margin at the c of distance as the callout box；
(3) it centered on current heartbeat R crest value high point position, is arrived with it horizontal between latter heartbeat R crest value high point position
Right margin at the d of distance as the callout box；
Described c, d meet:
1/3 < c≤2/3, and 1/3 < d≤2/3.
2. mask method according to claim 1, it is characterised in that: 1/2≤c≤2/3, and 1/2≤d≤2/3.
3. mask method according to claim 2, it is characterised in that: c=d=1/2.
4. mask method according to claim 1 to 3, which is characterized in that the delimitation of the callout box further include it is upper,
The delimitation step of lower boundary:
(1) using horizontal line locating for current heartbeat R crest value high point position as the coboundary of the callout box；
(2) with current heartbeat Q crest value low dot location and S crest value low dot location, water locating for the lower one in position in the two
Lower boundary of the horizontal line as the callout box.
5. a kind of electrocardiogram recognition training method successively includes raw data set, image labeling, the model training for obtaining electrocardiogram
The step of being assessed with test, it is characterised in that: described image annotation step is using the mark side as described in claim 1-4 is any
Method.
6. training method according to claim 5, it is characterised in that: the training pattern that the model training step uses for
Faster R-CNN model.
7. a kind of electrocardiogram recognition methods forms several callout box images including carrying out image labeling to electrocardiogram to be identified
Step, and the step of machine recognition is executed to several callout box images；It is characterized by: being appointed using claim 1-4
Mask method described in one carries out image labeling to electrocardiogram to be identified to form several callout box images.
8. a kind of electrocardiogram identification device, including electrocardiogram labeling module and picture recognition module, the electrocardiogram labeling module
Several callout box images are formed for carrying out image labeling to electrocardiogram to be identified；Described image identification module is used for described
Several callout box images execute machine recognition；It is characterized by: the electrocardiogram labeling module utilizes any institute of claim 1-4
The mask method stated carries out image labeling to electrocardiogram to be identified to form several callout box images.
9. electrocardiogram identification device according to claim 8, it is characterised in that: further include image training module, the figure
As training module is using electrocardiogram recognition training method described in claim 5 or 6, described image identification module is identified
Training.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910327442.XA CN110141214A (en) | 2019-04-23 | 2019-04-23 | A kind of mask method of electrocardiogram identification and its application |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910327442.XA CN110141214A (en) | 2019-04-23 | 2019-04-23 | A kind of mask method of electrocardiogram identification and its application |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110141214A true CN110141214A (en) | 2019-08-20 |
Family
ID=67593851
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910327442.XA Pending CN110141214A (en) | 2019-04-23 | 2019-04-23 | A kind of mask method of electrocardiogram identification and its application |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110141214A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110811608A (en) * | 2019-11-19 | 2020-02-21 | 中电健康云科技有限公司 | Atrial fibrillation monitoring method based on ECG (ECG) signals |
| TWI758039B (en) * | 2020-12-29 | 2022-03-11 | 財團法人國家衛生研究院 | Electronic device and method for selecting feature of electrocardiogram |
Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030135097A1 (en) * | 2001-06-25 | 2003-07-17 | Science Applications International Corporation | Identification by analysis of physiometric variation |
| CN105748063A (en) * | 2016-04-25 | 2016-07-13 | 山东大学齐鲁医院 | Intelligent arrhythmia diagnosis method based on multiple-lead and convolutional neural network |
| CN106108880A (en) * | 2016-06-28 | 2016-11-16 | 吉林大学 | A kind of heart claps automatic identifying method and system |
| CN106725428A (en) * | 2016-12-19 | 2017-05-31 | 中国科学院深圳先进技术研究院 | A kind of electrocardiosignal sorting technique and device |
| CN106778685A (en) * | 2017-01-12 | 2017-05-31 | 司马大大（北京）智能系统有限公司 | Electrocardiogram image-recognizing method, device and service terminal |
| CN106805965A (en) * | 2016-12-19 | 2017-06-09 | 深圳先进技术研究院 | A kind of electrocardiosignal sorting technique and device |
| CN107137072A (en) * | 2017-04-28 | 2017-09-08 | 北京科技大学 | A kind of ventricular ectopic beating detection method based on 1D convolutional neural networks |
| CN107239684A (en) * | 2017-05-22 | 2017-10-10 | 吉林大学 | A kind of feature learning method and system for ECG identifications |
| CN108464827A (en) * | 2018-03-08 | 2018-08-31 | 四川大学 | It is a kind of it is Weakly supervised under electrocardio image-recognizing method |
| CN108596142A (en) * | 2018-05-09 | 2018-09-28 | 吉林大学 | A kind of cardioelectric characteristic extracting process based on PCANet |
| CN108932452A (en) * | 2017-05-22 | 2018-12-04 | 中国科学院半导体研究所 | Arrhythmia cordis beat classification method based on multiple dimensioned convolutional neural networks |
| CN109044347A (en) * | 2018-07-11 | 2018-12-21 | 上海夏先机电科技发展有限公司 | Identify electrocardio wave image junctional escape beat method, apparatus, system and electronic equipment |
| CN109117729A (en) * | 2018-07-11 | 2019-01-01 | 上海夏先机电科技发展有限公司 | Electrocardiogram room escape real-time judge method, apparatus, system and storage medium |
| CN109124620A (en) * | 2018-06-07 | 2019-01-04 | 深圳市太空科技南方研究院 | A kind of atrial fibrillation detection method, device and equipment |
| CN109165556A (en) * | 2018-07-24 | 2019-01-08 | 吉林大学 | One kind being based on GRNN personal identification method |
| CN109303559A (en) * | 2018-11-01 | 2019-02-05 | 杭州质子科技有限公司 | A kind of dynamic ECG beat classification method promoting decision tree based on gradient |
| CN109497986A (en) * | 2018-11-22 | 2019-03-22 | 杭州脉流科技有限公司 | Electrocardiogram intelligent analysis method, device, computer equipment and system based on deep neural network |
| CN109620211A (en) * | 2018-11-01 | 2019-04-16 | 吉林大学珠海学院 | A kind of intelligent abnormal electrocardiogram aided diagnosis method based on deep learning |
-
2019
- 2019-04-23 CN CN201910327442.XA patent/CN110141214A/en active Pending
Patent Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030135097A1 (en) * | 2001-06-25 | 2003-07-17 | Science Applications International Corporation | Identification by analysis of physiometric variation |
| CN105748063A (en) * | 2016-04-25 | 2016-07-13 | 山东大学齐鲁医院 | Intelligent arrhythmia diagnosis method based on multiple-lead and convolutional neural network |
| CN106108880A (en) * | 2016-06-28 | 2016-11-16 | 吉林大学 | A kind of heart claps automatic identifying method and system |
| CN106725428A (en) * | 2016-12-19 | 2017-05-31 | 中国科学院深圳先进技术研究院 | A kind of electrocardiosignal sorting technique and device |
| CN106805965A (en) * | 2016-12-19 | 2017-06-09 | 深圳先进技术研究院 | A kind of electrocardiosignal sorting technique and device |
| CN106778685A (en) * | 2017-01-12 | 2017-05-31 | 司马大大（北京）智能系统有限公司 | Electrocardiogram image-recognizing method, device and service terminal |
| CN107137072A (en) * | 2017-04-28 | 2017-09-08 | 北京科技大学 | A kind of ventricular ectopic beating detection method based on 1D convolutional neural networks |
| CN108932452A (en) * | 2017-05-22 | 2018-12-04 | 中国科学院半导体研究所 | Arrhythmia cordis beat classification method based on multiple dimensioned convolutional neural networks |
| CN107239684A (en) * | 2017-05-22 | 2017-10-10 | 吉林大学 | A kind of feature learning method and system for ECG identifications |
| CN108464827A (en) * | 2018-03-08 | 2018-08-31 | 四川大学 | It is a kind of it is Weakly supervised under electrocardio image-recognizing method |
| CN108596142A (en) * | 2018-05-09 | 2018-09-28 | 吉林大学 | A kind of cardioelectric characteristic extracting process based on PCANet |
| CN109124620A (en) * | 2018-06-07 | 2019-01-04 | 深圳市太空科技南方研究院 | A kind of atrial fibrillation detection method, device and equipment |
| CN109044347A (en) * | 2018-07-11 | 2018-12-21 | 上海夏先机电科技发展有限公司 | Identify electrocardio wave image junctional escape beat method, apparatus, system and electronic equipment |
| CN109117729A (en) * | 2018-07-11 | 2019-01-01 | 上海夏先机电科技发展有限公司 | Electrocardiogram room escape real-time judge method, apparatus, system and storage medium |
| CN109165556A (en) * | 2018-07-24 | 2019-01-08 | 吉林大学 | One kind being based on GRNN personal identification method |
| CN109303559A (en) * | 2018-11-01 | 2019-02-05 | 杭州质子科技有限公司 | A kind of dynamic ECG beat classification method promoting decision tree based on gradient |
| CN109620211A (en) * | 2018-11-01 | 2019-04-16 | 吉林大学珠海学院 | A kind of intelligent abnormal electrocardiogram aided diagnosis method based on deep learning |
| CN109497986A (en) * | 2018-11-22 | 2019-03-22 | 杭州脉流科技有限公司 | Electrocardiogram intelligent analysis method, device, computer equipment and system based on deep neural network |
Non-Patent Citations (2)
| Title |
|---|
| JIAPU PAN等: "《A Real-Time QRS Detection Algorithm》", 《TRANSACTIONS ON BIOMEDICAL ENGINEERING》 * |
| 马玉润: "《ECG预处理与QRS波群检测技术研究》", 《信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110811608A (en) * | 2019-11-19 | 2020-02-21 | 中电健康云科技有限公司 | Atrial fibrillation monitoring method based on ECG (ECG) signals |
| TWI758039B (en) * | 2020-12-29 | 2022-03-11 | 財團法人國家衛生研究院 | Electronic device and method for selecting feature of electrocardiogram |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Oresko et al. | A wearable smartphone-based platform for real-time cardiovascular disease detection via electrocardiogram processing | |
| Mousavi et al. | ECGNET: Learning where to attend for detection of atrial fibrillation with deep visual attention | |
| JP4386235B2 (en) | Method and apparatus for sequential comparison of electrocardiograms | |
| CN109171712A (en) | Auricular fibrillation recognition methods, device, equipment and computer readable storage medium | |
| CN107103182A (en) | A kind of heart disease Warning System and method based on deep learning algorithm | |
| EP3698709A1 (en) | Electrocardiogram information processing method and electrocardiogram workstation system | |
| CN108647565A (en) | A kind of data preprocessing method classified to electrocardiosignal based on deep learning model | |
| EP3698708A1 (en) | Ecg information processing method and ecg workstation | |
| CN102551701A (en) | System and method of detecting abnormal movement of a physical object | |
| CN109288515B (en) | Periodicity monitoring method and device based on premature beat signal in wearable electrocardiosignal | |
| CN102869303A (en) | Visualization of myocardial infarct size in diagnostic ECG | |
| CN112826513B (en) | Fetal heart rate detection system based on deep learning and specificity correction on FECG | |
| CN106073760A (en) | 18 Lead ambulatory electrocardiograms based on wilson's lead2s analyze method and system | |
| CN112932498B (en) | T waveform state classification system with generalization capability based on deep learning | |
| Zhou et al. | ECG quality assessment using 1D-convolutional neural network | |
| CN110141214A (en) | A kind of mask method of electrocardiogram identification and its application | |
| CN105962926A (en) | Cardiac beat feature analysis method and equipment based on electrocardiogram | |
| CN105943020B (en) | A kind of 6 lead misconnection detecting system of cardiac electric axis and method | |
| CN112971800A (en) | Deep learning-based electrocardio premature beat signal processing method, module and device | |
| CN107530020A (en) | Method and system for the heart ischemia detection based on ECG | |
| CN110507311A (en) | A kind of ecg analysis method, apparatus, equipment and medium based on image information | |
| Srinivas et al. | Adaptive learning based heartbeat classification | |
| JPH0366358A (en) | Biosignal measuring instrument | |
| Bian et al. | Direct feature extraction and diagnosis of ECG signal in the compressed domain | |
| CN109394206B (en) | Real-time monitoring method and device based on premature beat signal in wearable electrocardiosignal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190820 |