CN104834921A - Electrocardio normality/abnormality big-data processing method and device - Google Patents
Electrocardio normality/abnormality big-data processing method and device Download PDFInfo
- Publication number
- CN104834921A CN104834921A CN201510268861.2A CN201510268861A CN104834921A CN 104834921 A CN104834921 A CN 104834921A CN 201510268861 A CN201510268861 A CN 201510268861A CN 104834921 A CN104834921 A CN 104834921A
- Authority
- CN
- China
- Prior art keywords
- data
- waveform
- electrocardio
- waveform data
- electrocardiogram
- 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.)
- Granted
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 13
- 230000005856 abnormality Effects 0.000 title abstract description 4
- 238000012545 processing Methods 0.000 claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 22
- 238000010606 normalization Methods 0.000 claims abstract description 9
- 230000002159 abnormal effect Effects 0.000 claims description 52
- 238000000605 extraction Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 9
- 230000011218 segmentation Effects 0.000 claims description 9
- 230000000747 cardiac effect Effects 0.000 abstract description 5
- 238000012216 screening Methods 0.000 abstract description 4
- 238000000638 solvent extraction Methods 0.000 abstract 2
- 238000004364 calculation method Methods 0.000 description 3
- 208000000418 Premature Cardiac Complexes Diseases 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 206010015856 Extrasystoles Diseases 0.000 description 1
- 206010020880 Hypertrophy Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 206010003119 arrhythmia Diseases 0.000 description 1
- 230000006793 arrhythmia Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002565 electrocardiography Methods 0.000 description 1
- 239000003792 electrolyte Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 208000010125 myocardial infarction Diseases 0.000 description 1
- 208000031225 myocardial ischemia Diseases 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000000718 qrs complex Methods 0.000 description 1
- 230000002861 ventricular Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
- G06F2218/16—Classification; Matching by matching signal segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/03—Recognition of patterns in medical or anatomical images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses an electrocardio normality/abnormality big-data processing method and device, and the method employs a normal electrocardio database, and comprises the following steps: S1, partitioning to-be-classified electrocardio data according to cardiac beat, and then carrying out the normalization processing of length and amplitude, thereby forming a plurality of pieces of beat-wave-shaped data; S2, extracting the index data of to-be-classified electrocardio data; S3, determining a confidence interval according to the index data stored in the database, comparing the index data of the extracted to-be-classified electrocardio data with the confidence interval, and outputting a comparison result; S4, calculating the similarity of the plurality of pieces of beat-wave-shaped data obtained through the partitioning of the to-be-classified electrocardio data with waveform data, corresponding to cardiac beat, in the electrocardio data stored in the database, comparing the similarity with a similarity threshold value, and outputting the comparison results. The device comprises the normal electrocardio database, and a plurality of modules which are used for achieving the above steps. The device can achieve the reliable classification and screening of the electrocardiogram which is not diagnosed, and avoids false negative.
Description
Technical Field
The invention relates to the field of pattern recognition, big data analysis and medical signal processing, in particular to a method and a device for processing electrocardio positive/abnormal big data.
Background
The electrocardiogram is a clinical examination, and can help to diagnose arrhythmia, cardiac ischemia, myocardial infarction, atrioventricular hypertrophy, conduction block, premature beat and other abnormalities by recording the electrical activity of the heart in each period, and can also be used for judging the influence of electrolytes and medicines on the heart.
At present, the popularization of the electrocardiogram screening is limited to the number of cardiologists who can read the electrocardiogram to a great extent. Many research institutes have therefore focused on the development of automated diagnostic systems for electrocardiography. Aiming at realizing automatic classification of electrocardio by computers at home and abroad, a plurality of research methods are adopted, such as an artificial neural network, fuzzy clustering, a logic discrimination tree and template matching … …, and when the electrocardio is classified, some abnormal electrocardio is classified into a plurality of types according to different diseases; some classify them for certain specific abnormalities, such as ventricular premature beats. There is no mature method to support electrocardiographic screening by identifying whether an electrocardiogram is abnormal or not.
Disclosure of Invention
The invention aims to provide a method and a device for processing electrocardiographic positive/abnormal big data, which can reliably classify and screen undiagnosed electrocardiograms (electrocardiographic data to be classified) through a computer, avoid false negatives (abnormal electrocardiograms are judged to be normal electrocardiograms), enable doctors to only need to diagnose the electrocardiograms which are judged to be abnormal again, and reduce the workload of the doctors.
The specific technical scheme of the invention is as follows:
a big data processing method of electrocardio normal/abnormal, the processing method includes the normal electrocardio database, the database stores as much normal electrocardio data as possible, each normal electrocardio data includes index data and several beat waveform data obtained by segmenting the electrocardio data according to heart beat; the processing method comprises the following steps:
s1, segmenting the electrocardiogram data to be classified according to the heartbeat, and then respectively carrying out normalization processing on the length and the amplitude to form a plurality of beat waveform data;
s2, extracting index data of the electrocardiogram data to be classified;
s3, determining a confidence interval according to the index data stored in the database, comparing the extracted index data of the electrocardiogram data to be classified with the confidence interval, and outputting a comparison result; and
s4, calculating the similarity between a plurality of beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database, and comparing the similarity with a similarity threshold value to output a comparison result;
the indicator data includes at least one of QRS band length, PR interval, QT interval, and RR interval.
In the above method for processing electrocardiographic positive/abnormal big data, preferably, in the normal electrocardiographic database, each electrocardiographic data corresponding to each cardiac beat includes a plurality of segments of waveform data, the segments of waveform data have equal length and are mostly overlapped, and the central positions of the segments of waveform data are respectively located at the waveform peak and have only a plurality of data points different from each other before and after the waveform peak; the step S4 includes the steps of:
s41, calculating waveform data of a beat obtained by dividing the electrocardio data to be classified with the corresponding waveform data of the heart beat in the electrocardio data stored in the database respectively to obtain a plurality of similarities relative to the waveform data of the plurality of segments;
s42, selecting the minimum value from the obtained multiple similarities relative to the multiple pieces of waveform data as the similarity between the waveform data of the beat divided from the electrocardio data to be classified and the waveform data of the corresponding beat in the electrocardio data stored in the database;
s43, circularly executing the steps S41 and S42, and calculating the similarity between the waveform data of the beat divided from the electrocardio data to be classified and the waveform data of the corresponding beat in other electrocardio data stored in the database; and
and S44, circularly executing the steps S41, S42 and S43, and acquiring the similarity between other beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database.
In the above method for processing electrocardiographic positive/abnormal big data, preferably, the processing method further includes a step of preprocessing electrocardiographic data to be classified before segmentation and index extraction.
In the above method for processing electrocardiographic positive/abnormal big data, the comparison result output in step S4 preferably includes: and when the output comparison result is abnormal electrocardiogram data, the output comparison result also comprises an abnormal period, wherein the abnormal period refers to a heartbeat period corresponding to waveform data, of which the similarity with corresponding heartbeat waveform data in all the electrocardiogram data stored in the database is greater than the similarity threshold value, in the plurality of heartbeat waveform data of the electrocardiogram data to be classified.
In the above method for processing electrocardiographic positive/abnormal large data, the step S4 preferably includes: calculating difference values of corresponding points of two sections of waveform data to be compared one by one; and taking the absolute value of each difference value and then summing the absolute values, and taking the sum as the similarity of the two sections of waveform data.
An electrocardio positive/abnormal big data processing device, which comprises:
the normal electrocardiogram data base stores normal electrocardiogram data as much as possible, and each piece of normal electrocardiogram data comprises index data and a plurality of beat waveform data obtained by segmenting the electrocardiogram data according to heart beats;
the segmentation module is used for segmenting the electrocardiogram data to be classified according to the heartbeat, and then respectively carrying out normalization processing on the length and the amplitude to form a plurality of beat waveform data;
the index extraction module is used for extracting index data of the electrocardiogram data to be classified;
the index matching module is used for determining a confidence interval according to the index data stored in the database, comparing the extracted index data of the electrocardiogram data to be classified with the confidence interval and outputting a comparison result; and
the waveform matching module is used for calculating the similarity between a plurality of beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database, and comparing the similarity with a similarity threshold value to output a comparison result;
the indicator data includes at least one of QRS band length, PR interval, QT interval, and RR interval.
In the above apparatus for processing electrocardiographic positive/abnormal big data, preferably, in the normal electrocardiographic database, each electrocardiographic data corresponding to each cardiac beat includes a plurality of segments of waveform data, the segments of waveform data have equal length and are mostly overlapped, and the central positions of the segments of waveform data are respectively located at the waveform peak and have only a plurality of data points different from each other before and after the waveform peak; the waveform matching module includes:
the first module is used for calculating waveform data of one beat obtained by dividing the electrocardio data to be classified with the waveform data of the corresponding beat in the electrocardio data stored in the database to obtain a plurality of similarities relative to the waveform data of the multiple beats;
a second module, configured to select a minimum value from the obtained multiple similarities with respect to the multiple pieces of waveform data, as a similarity between the waveform data of one beat segmented from the electrocardiographic data to be classified and waveform data of a corresponding beat in the piece of electrocardiographic data stored in the database;
the third module is used for calling the first module and the second module in a circulating mode and calculating the similarity between the waveform data of the beat segmented from the electrocardio data to be classified and the waveform data of the corresponding beat in other electrocardio data stored in the database; and
and the fourth module is used for calling the first module, the second module and the third module in a circulating manner to acquire the similarity between other beat waveform data segmented from the electrocardio data to be classified and corresponding beat waveform data in the electrocardio data stored in the database.
In the above apparatus for processing electrocardiographic positive/abnormal big data, preferably, the apparatus further comprises a preprocessing module for preprocessing electrocardiographic data to be classified before segmentation and index extraction.
In the above-mentioned electrocardiographic positive/abnormal large data processing device, preferably, the comparison result output by the waveform matching module includes: and when the output comparison result is abnormal electrocardiogram data, the output comparison result also comprises an abnormal period, wherein the abnormal period refers to a heartbeat period corresponding to waveform data, of which the similarity with corresponding heartbeat waveform data in all the electrocardiogram data stored in the database is greater than the similarity threshold value, in the plurality of heartbeat waveform data of the electrocardiogram data to be classified.
In the above-described electrocardiographic positive/abnormal large data processing device, preferably, the waveform matching module may further include: calculating difference values of corresponding points of two sections of waveform data to be compared one by one; and taking the absolute value of each difference value and then summing the absolute values, and taking the sum as the similarity of the two sections of waveform data.
According to the method, by mining and analyzing the rules of the normal electrocardiogram database and realizing the classification of the electrocardiogram data based on the matching algorithm of the characteristics and the waveforms, the normal electrocardiogram data can be reliably identified, and false negatives (abnormal electrocardio is judged as normal electrocardio) are avoided, so that a doctor can be assisted in carrying out electrocardiogram screening, and the doctor only needs to carry out re-diagnosis on the abnormal electrocardio, and the workload of the doctor can be greatly reduced.
Drawings
FIG. 1 is a flow chart of some embodiments of a method for processing electrocardiographic positive/abnormal big data according to the present invention;
fig. 2 to 6 show the matching situation between a beat of waveform segmented from the electrocardiographic data to be classified and five segments of waveforms corresponding to the beats in a piece of electrocardiographic data stored in the database.
Detailed Description
The invention is further illustrated with reference to the following figures and examples. These more detailed descriptions are intended to aid in understanding the invention and should not be used to limit the invention. It will be apparent to one skilled in the art in light of the present disclosure that the present invention may be practiced without some or all of these specific details. In other instances, well known operations have not been described in detail in order not to unnecessarily obscure the present invention.
As shown in fig. 1, the electrocardiographic positive/abnormal big data processing method according to some embodiments includes a normal electrocardiographic database storing as many normal electrocardiographic data as possible, each piece of normal electrocardiographic data including index data and a plurality of beat waveform data obtained by segmenting electrocardiographic data by heart beat; wherein the indicator data includes QRS band length, PR interval, QT interval and RR interval.
Generally speaking, a normal electrocardiogram database can be established by collecting and sorting tens of thousands of normal electrocardiogram data, and is used for summarizing and analyzing the regularity of the normal electrocardiogram data. The normal electrocardiogram database is specifically established as follows:
and S100, intercepting each period of each piece of normal electrocardio data through a peak detection method (namely segmenting the normal electrocardio data according to heart beat). Assuming that data points intercepted in a certain period are the kth to the kth + N points in one piece of data, the kth to the kth + N points, the kth-2 to the kth + N-2 points, the kth-1 to the kth + N-1 points, the kth +1 to the kth + N +1 points and the kth +2 to the kth + N +2 points are respectively stored into a database, each intercepted section is subjected to length normalization by methods such as interpolation before storage, and the amplitude of different sections of data is subjected to normalization. As can be seen from the above, each piece of electrocardiographic data includes a plurality of (in this example, five) pieces of waveform data corresponding to each heart beat, the plurality of pieces of waveform data have equal lengths and are mostly overlapped, and the center positions of the plurality of pieces of waveform data are respectively located at the waveform peak and have only a plurality of data points different from each other before and after the waveform peak.
And S200, storing indexes such as QRS wave band length, PR interval, QT interval, RR interval and the like of each piece of normal electrocardio data into a database.
The electrocardio positive/abnormal big data processing method of some embodiments comprises the following steps:
and step S1, segmenting the electrocardiogram data to be classified according to the heartbeat, and then respectively carrying out normalization processing on the length and the amplitude to form a plurality of beat waveform data. Specifically, the waveform per beat may be segmented according to the result of QRS complex identification. During the division, a double-threshold method and the like can be adopted to determine the starting point and the end point of the division in each period. Assume that a beat waveform data (certain period data) divided from electrocardiographic data to be classified is a ═ a1，a2，...，anIn order to make the length of the waveform data consistent with the length of each section of waveform data in the database, the data is subjected to interpolation and other operations, and the processed data is
And S2, extracting index data of the electrocardiogram data to be classified. The index data includes QRS band length, PR interval, QT interval and RR interval.
Step S3, determining a confidence interval according to the index data stored in the database, comparing the extracted index data of the electrocardiogram data to be classified with the confidence interval, and outputting a comparison result. Firstly, the normality of the electrocardiogram data to be classified is preliminarily judged through index matching. Specifically, a confidence interval can be determined according to index ranges, mean values and variances of the QRS wave band length, the PR interval, the QT interval, the RR interval and the like of the normal electrocardiogram data in the database, and is used for making a preliminary judgment on whether the electrocardiogram data to be detected is normal or not.
And step S4, calculating the similarity between a plurality of beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database, and comparing the similarity with a similarity threshold value to output a comparison result. I.e. a waveform matching step.
In the waveform matching step, whether the electrocardiogram (to-be-classified electrocardiogram data) is normal or not is judged mainly by comparing the similarity between the waveform of the undiagnosed electrocardiogram and the known normal electrocardiogram.
Considering that the similarity analysis is closely related to the relative positions of the two compared waveforms, when the similarity calculation is carried out, a plurality of similarities obtained when the peaks of the two waveforms are aligned and when the peaks are moved by a plurality of data points relative to the positions are simultaneously calculated, and the minimum value is taken as the final similarity of the two waveforms. Therefore, the step S4 preferably includes the steps of:
and S41, calculating the waveform data of one beat obtained by dividing the electrocardio data to be classified with the waveform data of the corresponding beat in the electrocardio data stored in the database respectively to obtain a plurality of similarities relative to the waveform data of the plurality of beats. Specifically, it can be expressed as the following formula
Wherein,representing a beat of waveform data segmented from said electrocardiographic data to be classified, bi，j，k，wSpecifically, 5 pieces of data (k-2 to k + N-2 points, k-1 to k + N-1, k to k + N, k +1 to k + N +1 points, and k +2 to k + N +2 points) stored in the database for each cycle (assuming that the starting point is the kth point) are stored.
As can be seen from the above, in this embodiment, the similarity calculation includes: calculating difference values of corresponding points of two sections of waveform data to be compared one by one; and taking the absolute value of each difference value and then summing the absolute values, and taking the sum as the similarity of the two sections of waveform data. However, the present invention is not limited to this, and for example, the absolute values of the differences may be averaged, and the average may be used as the similarity between the two pieces of waveform data, or other similar methods may be used.
Fig. 2 to 6 show the matching of a beat waveform 1 segmented from the electrocardiographic data to be classified with five segments of waveforms 2, 3, 4, 5 and 6 of a corresponding beat in a piece of electrocardiographic data stored in the database.
And S42, selecting the minimum value from the obtained multiple similarities relative to the multiple pieces of waveform data as the similarity between the waveform data of the beat divided from the electrocardio data to be classified and the waveform data of the corresponding beat in the electrocardio data stored in the database. Can be expressed as the following formula
And S43, executing the steps S41 and S42 in a circulating way, and calculating the similarity between the waveform data of the beat segmented from the electrocardio data to be classified and the waveform data of the corresponding beat in other electrocardio data stored in the database. I.e. traverse i, j in the above formula. Similarity Δ to be calculatedi，jWhen there is a, compared to the similarity threshold thetai，jWhen the waveform is larger than theta, judging the waveform of the segmentThe method is a normal cardiac cycle, wherein the similarity threshold theta can be set according to an empirical value.
And S44, circularly executing the steps S41, S42 and S43, and acquiring the similarity between other beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database. If the similarity of all the segments is smaller than the similarity threshold theta, the waveform matching output result is that the electrocardiogram data is normal electrocardiogram data, otherwise, the electrocardiogram data is output as abnormal electrocardiogram data, and an abnormal period is output.
And when the index matching result and the waveform matching result are normal, judging that the electrocardiogram data to be classified is normal, otherwise, judging that the electrocardiogram data to be classified is abnormal.
Further, in some embodiments, the method further includes a step S1' of preprocessing the electrocardiographic data to be classified before the segmentation and index extraction. These pretreatments may include: fast classifying through basic electrocardio indexes, and if the heart rate is abnormal, directly judging that the electrocardio data is abnormal; for an electrocardiogram with normal basic electrocardiogram indexes, if baseline drift or noise is serious, denoising the electrocardiogram by means of filtering and the like.
It should be noted that the steps in the present invention are numbered for convenience of description only and do not represent a specific order relationship, for example, steps S1, S2, S3 and S4 are not necessarily performed in the order of the sequence in the method described above, the order of steps S1 and S2 may be reversed, step S1 may be placed after step S3, and so on.
The electrocardio normal/abnormal big data processing device corresponding to the electrocardio normal/abnormal big data processing method of some embodiments comprises the following steps:
the normal electrocardiogram data base stores normal electrocardiogram data as much as possible, and each piece of normal electrocardiogram data comprises index data and a plurality of beat waveform data obtained by segmenting the electrocardiogram data according to heart beats;
the segmentation module is used for segmenting the electrocardiogram data to be classified according to the heartbeat, and then respectively carrying out normalization processing on the length and the amplitude to form a plurality of beat waveform data;
the index extraction module is used for extracting index data of the electrocardiogram data to be classified;
the index matching module is used for determining a confidence interval according to the index data stored in the database, comparing the extracted index data of the electrocardiogram data to be classified with the confidence interval and outputting a comparison result; and
the waveform matching module is used for calculating the similarity between a plurality of beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database, and comparing the similarity with a similarity threshold value to output a comparison result;
the indicator data includes at least one of QRS band length, PR interval, QT interval, and RR interval.
In the normal electrocardiogram database, each electrocardiogram data corresponding to each heartbeat comprises a plurality of sections of waveform data, the lengths of the plurality of sections of waveform data are equal, most of the plurality of sections of waveform data are overlapped, and the central positions of the plurality of sections of waveform data are respectively positioned at the waveform peak value and only have a plurality of data points which are different from each other before and after the waveform peak value; the waveform matching module includes:
the first module is used for calculating waveform data of one beat obtained by dividing the electrocardio data to be classified with the waveform data of the corresponding beat in the electrocardio data stored in the database to obtain a plurality of similarities relative to the waveform data of the multiple beats;
a second module, configured to select a minimum value from the obtained multiple similarities with respect to the multiple pieces of waveform data, as a similarity between the waveform data of one beat segmented from the electrocardiographic data to be classified and waveform data of a corresponding beat in the piece of electrocardiographic data stored in the database;
the third module is used for calling the first module and the second module in a circulating mode and calculating the similarity between the waveform data of the beat segmented from the electrocardio data to be classified and the waveform data of the corresponding beat in other electrocardio data stored in the database; and
and the fourth module is used for calling the first module, the second module and the third module in a circulating manner to acquire the similarity between other beat waveform data segmented from the electrocardio data to be classified and corresponding beat waveform data in the electrocardio data stored in the database.
The processing device also comprises a preprocessing module which is used for preprocessing the electrocardiogram data to be classified before segmentation and index extraction.
The comparison result output by the waveform matching module comprises: and when the output comparison result is abnormal electrocardiogram data, the output comparison result also comprises an abnormal period, wherein the abnormal period refers to a heartbeat period corresponding to waveform data, of which the similarity with corresponding heartbeat waveform data in all the electrocardiogram data stored in the database is greater than the similarity threshold value, in the plurality of heartbeat waveform data of the electrocardiogram data to be classified.
In the waveform matching module, the similarity calculation includes: calculating difference values of corresponding points of two sections of waveform data to be compared one by one; and taking the absolute value of each difference value and then summing the absolute values, and taking the sum as the similarity of the two sections of waveform data.
In the embodiment, the normal electrocardiogram data base is mined and analyzed according to rules, and classification of electrocardiogram data is realized based on a matching algorithm of features and waveforms, so that the normal electrocardiogram data can be reliably identified, and false negative is avoided. Therefore, a doctor can be assisted in diagnosing a large amount of electrocardio data, namely, before the doctor diagnoses the large amount of electrocardio data, the doctor firstly uses the computer to automatically screen through the method or the device, and the doctor only needs to diagnose the screened abnormal electrocardio data.
Claims (10)
1. The method for processing the big electrocardio positive/abnormal data is characterized by comprising a normal electrocardio database, wherein the database stores as much normal electrocardio data as possible, and each piece of normal electrocardio data comprises index data and a plurality of beat waveform data obtained by segmenting the electrocardio data according to heart beats; the processing method comprises the following steps:
s1, segmenting the electrocardiogram data to be classified according to the heartbeat, and then respectively carrying out normalization processing on the length and the amplitude to form a plurality of beat waveform data;
s2, extracting index data of the electrocardiogram data to be classified;
s3, determining a confidence interval according to the index data stored in the database, comparing the extracted index data of the electrocardiogram data to be classified with the confidence interval, and outputting a comparison result; and
s4, calculating the similarity between a plurality of beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database, and comparing the similarity with a similarity threshold value to output a comparison result;
the indicator data includes at least one of QRS band length, PR interval, QT interval, and RR interval.
2. The method for processing big electrocardiographic/abnormal data according to claim 1,
in the normal electrocardiogram database, each electrocardiogram data corresponding to each heartbeat comprises a plurality of sections of waveform data, the lengths of the plurality of sections of waveform data are equal, most of the plurality of sections of waveform data are overlapped, and the central positions of the plurality of sections of waveform data are respectively positioned at the waveform peak value and only have a plurality of data points which are different from each other before and after the waveform peak value;
the step S4 includes the steps of:
s41, calculating waveform data of a beat obtained by dividing the electrocardio data to be classified with the corresponding waveform data of the heart beat in the electrocardio data stored in the database respectively to obtain a plurality of similarities relative to the waveform data of the plurality of segments;
s42, selecting the minimum value from the obtained multiple similarities relative to the multiple pieces of waveform data as the similarity between the waveform data of the beat divided from the electrocardio data to be classified and the waveform data of the corresponding beat in the electrocardio data stored in the database;
s43, circularly executing the steps S41 and S42, and calculating the similarity between the waveform data of the beat divided from the electrocardio data to be classified and the waveform data of the corresponding beat in other electrocardio data stored in the database; and
and S44, circularly executing the steps S41, S42 and S43, and acquiring the similarity between other beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database.
3. The method for processing the electrocardiographic positive/abnormal big data according to claim 1, wherein the processing method further comprises the step of preprocessing the electrocardiographic data to be classified before segmentation and index extraction.
4. The method for processing big electrocardiographic/abnormal data according to claim 1, wherein the comparison result output in step S4 includes: and when the output comparison result is abnormal electrocardiogram data, the output comparison result also comprises an abnormal period, wherein the abnormal period refers to a heartbeat period corresponding to waveform data, of which the similarity with corresponding heartbeat waveform data in all the electrocardiogram data stored in the database is greater than the similarity threshold value, in the plurality of heartbeat waveform data of the electrocardiogram data to be classified.
5. The electrocardiographic positive/abnormal big data processing method according to claim 1, wherein the calculating of the degree of similarity in step S4 includes: calculating difference values of corresponding points of two sections of waveform data to be compared one by one; and taking the absolute value of each difference value and then summing the absolute values, and taking the sum as the similarity of the two sections of waveform data.
6. An electrocardio positive/abnormal big data processing device is characterized by comprising:
the normal electrocardiogram data base stores normal electrocardiogram data as much as possible, and each piece of normal electrocardiogram data comprises index data and a plurality of beat waveform data obtained by segmenting the electrocardiogram data according to heart beats;
the segmentation module is used for segmenting the electrocardiogram data to be classified according to the heartbeat, and then respectively carrying out normalization processing on the length and the amplitude to form a plurality of beat waveform data;
the index extraction module is used for extracting index data of the electrocardiogram data to be classified;
the index matching module is used for determining a confidence interval according to the index data stored in the database, comparing the extracted index data of the electrocardiogram data to be classified with the confidence interval and outputting a comparison result; and
the waveform matching module is used for calculating the similarity between a plurality of beat waveform data segmented from the electrocardiogram data to be classified and waveform data of corresponding heartbeats in the electrocardiogram data stored in the database, and comparing the similarity with a similarity threshold value to output a comparison result;
the indicator data includes at least one of QRS band length, PR interval, QT interval, and RR interval.
7. The electrocardiographic positive/abnormal big data processing device according to claim 6,
in the normal electrocardiogram database, each electrocardiogram data corresponding to each heartbeat comprises a plurality of sections of waveform data, the lengths of the plurality of sections of waveform data are equal, most of the plurality of sections of waveform data are overlapped, and the central positions of the plurality of sections of waveform data are respectively positioned at the waveform peak value and only have a plurality of data points which are different from each other before and after the waveform peak value;
the waveform matching module includes:
the first module is used for calculating waveform data of one beat obtained by dividing the electrocardio data to be classified with the waveform data of the corresponding beat in the electrocardio data stored in the database to obtain a plurality of similarities relative to the waveform data of the multiple beats;
a second module, configured to select a minimum value from the obtained multiple similarities with respect to the multiple pieces of waveform data, as a similarity between the waveform data of one beat segmented from the electrocardiographic data to be classified and waveform data of a corresponding beat in the piece of electrocardiographic data stored in the database;
the third module is used for calling the first module and the second module in a circulating mode and calculating the similarity between the waveform data of the beat segmented from the electrocardio data to be classified and the waveform data of the corresponding beat in other electrocardio data stored in the database; and
and the fourth module is used for calling the first module, the second module and the third module in a circulating manner to acquire the similarity between other beat waveform data segmented from the electrocardio data to be classified and corresponding beat waveform data in the electrocardio data stored in the database.
8. The apparatus as claimed in claim 6, further comprising a preprocessing module for preprocessing the electrocardiographic data to be classified before segmentation and index extraction.
9. The apparatus as claimed in claim 6, wherein the comparison result outputted from the waveform matching module comprises: and when the output comparison result is abnormal electrocardiogram data, the output comparison result also comprises an abnormal period, wherein the abnormal period refers to a heartbeat period corresponding to waveform data, of which the similarity with corresponding heartbeat waveform data in all the electrocardiogram data stored in the database is greater than the similarity threshold value, in the plurality of heartbeat waveform data of the electrocardiogram data to be classified.
10. The apparatus as claimed in claim 6, wherein the waveform matching module comprises: calculating difference values of corresponding points of two sections of waveform data to be compared one by one; and taking the absolute value of each difference value and then summing the absolute values, and taking the sum as the similarity of the two sections of waveform data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510268861.2A CN104834921B (en) | 2015-05-22 | 2015-05-22 | Electrocardio just/abnormal big data processing method and processing device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510268861.2A CN104834921B (en) | 2015-05-22 | 2015-05-22 | Electrocardio just/abnormal big data processing method and processing device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104834921A true CN104834921A (en) | 2015-08-12 |
| CN104834921B CN104834921B (en) | 2018-01-16 |
Family
ID=53812799
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510268861.2A Expired - Fee Related CN104834921B (en) | 2015-05-22 | 2015-05-22 | Electrocardio just/abnormal big data processing method and processing device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104834921B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108090509A (en) * | 2017-12-13 | 2018-05-29 | 四川大学 | A kind of adaptive electrocardiogram sorting technique of data length |
| CN108227930A (en) * | 2018-01-18 | 2018-06-29 | 四川斐讯信息技术有限公司 | The gestural control method and wearable device of a kind of wearable device |
| WO2018166136A1 (en) * | 2017-03-14 | 2018-09-20 | 华为技术有限公司 | Method and apparatus for processing waveform signal |
| CN108903956A (en) * | 2018-05-22 | 2018-11-30 | 上海掌门科技有限公司 | It detects a lie device and method |
| CN109754877A (en) * | 2018-01-23 | 2019-05-14 | 上海移视网络科技有限公司 | A kind of 12 lead standard cardioelectric figure acute myocardial infarction AMI intelligent distinguishing system based on artificial intelligence |
| WO2020031105A3 (en) * | 2018-08-07 | 2020-06-11 | Goldtech Sino Ltd | Noninvasive systems and methods for continuous hemodynamic monitoring |
| CN113349791A (en) * | 2021-05-31 | 2021-09-07 | 平安科技（深圳）有限公司 | Abnormal electrocardiosignal detection method, device, equipment and medium |
| CN113576422A (en) * | 2021-10-08 | 2021-11-02 | 隽智生物医学研究实验室(佛山)有限公司 | Method, system and equipment for simplifying processing and quantizing physiological signal waveform |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030073917A1 (en) * | 2001-10-12 | 2003-04-17 | Neuropace, Inc. | Patient-specific parameter selection for neurological event detection |
| CN103110417A (en) * | 2013-02-28 | 2013-05-22 | 华东师范大学 | Automatic electrocardiogram recognition system |
| CN104102915A (en) * | 2014-07-01 | 2014-10-15 | 清华大学深圳研究生院 | Multiple-template matching identity recognition method based on ECG (Electrocardiogram) under electrocardiogram abnormality state |
| CN104473633A (en) * | 2014-12-31 | 2015-04-01 | 广州视源电子科技股份有限公司 | Abnormal electrocardiogram data judgment method and device |
| CN104523265A (en) * | 2014-12-31 | 2015-04-22 | 广州视源电子科技股份有限公司 | Abnormal electrocardiogram data judgment method and device |
-
2015
- 2015-05-22 CN CN201510268861.2A patent/CN104834921B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030073917A1 (en) * | 2001-10-12 | 2003-04-17 | Neuropace, Inc. | Patient-specific parameter selection for neurological event detection |
| CN103110417A (en) * | 2013-02-28 | 2013-05-22 | 华东师范大学 | Automatic electrocardiogram recognition system |
| CN104102915A (en) * | 2014-07-01 | 2014-10-15 | 清华大学深圳研究生院 | Multiple-template matching identity recognition method based on ECG (Electrocardiogram) under electrocardiogram abnormality state |
| CN104473633A (en) * | 2014-12-31 | 2015-04-01 | 广州视源电子科技股份有限公司 | Abnormal electrocardiogram data judgment method and device |
| CN104523265A (en) * | 2014-12-31 | 2015-04-22 | 广州视源电子科技股份有限公司 | Abnormal electrocardiogram data judgment method and device |
Non-Patent Citations (1)
| Title |
|---|
| 聂桥桥: "智能处理在ECG检测中的应用研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11321561B2 (en) | 2017-03-14 | 2022-05-03 | Huawei Technologies Co., Ltd. | Electrocardiogram waveform signal processing method and apparatus |
| WO2018166136A1 (en) * | 2017-03-14 | 2018-09-20 | 华为技术有限公司 | Method and apparatus for processing waveform signal |
| CN108701219A (en) * | 2017-03-14 | 2018-10-23 | 华为技术有限公司 | The method and device of waveform signal processing |
| CN108090509A (en) * | 2017-12-13 | 2018-05-29 | 四川大学 | A kind of adaptive electrocardiogram sorting technique of data length |
| CN108090509B (en) * | 2017-12-13 | 2021-10-08 | 四川大学 | Data length self-adaptive electrocardiogram classification method |
| CN108227930A (en) * | 2018-01-18 | 2018-06-29 | 四川斐讯信息技术有限公司 | The gestural control method and wearable device of a kind of wearable device |
| CN109754877A (en) * | 2018-01-23 | 2019-05-14 | 上海移视网络科技有限公司 | A kind of 12 lead standard cardioelectric figure acute myocardial infarction AMI intelligent distinguishing system based on artificial intelligence |
| CN109754877B (en) * | 2018-01-23 | 2023-12-22 | 上海移视网络科技有限公司 | Twelve-lead standard electrocardiogram acute myocardial infarction intelligent discrimination system based on artificial intelligence |
| CN108903956A (en) * | 2018-05-22 | 2018-11-30 | 上海掌门科技有限公司 | It detects a lie device and method |
| WO2020031105A3 (en) * | 2018-08-07 | 2020-06-11 | Goldtech Sino Ltd | Noninvasive systems and methods for continuous hemodynamic monitoring |
| CN113349791A (en) * | 2021-05-31 | 2021-09-07 | 平安科技（深圳）有限公司 | Abnormal electrocardiosignal detection method, device, equipment and medium |
| CN113576422B (en) * | 2021-10-08 | 2021-12-24 | 隽智生物医学研究实验室(佛山)有限公司 | Method, system and equipment for simplifying processing and quantizing physiological signal waveform |
| CN113576422A (en) * | 2021-10-08 | 2021-11-02 | 隽智生物医学研究实验室(佛山)有限公司 | Method, system and equipment for simplifying processing and quantizing physiological signal waveform |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104834921B (en) | 2018-01-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104834921B (en) | Electrocardio just/abnormal big data processing method and processing device | |
| CN109171712B (en) | Atrial fibrillation identification method, atrial fibrillation identification device, atrial fibrillation identification equipment and computer readable storage medium | |
| CN103110417B (en) | Automatic electrocardiogram recognition system | |
| US7769434B2 (en) | Method of physiological data analysis and measurement quality check using principal component analysis | |
| Ye et al. | Combining general multi-class and specific two-class classifiers for improved customized ECG heartbeat classification | |
| CN108596142B (en) | PCANet-based electrocardiogram feature extraction method | |
| US20220265199A1 (en) | Automatic method to delineate or categorize an electrocardiogram | |
| CN108511055B (en) | Ventricular premature beat recognition system and method based on classifier fusion and diagnosis rules | |
| De Lannoy et al. | Weighted SVMs and feature relevance assessment in supervised heart beat classification | |
| Kaya et al. | Feature selection using genetic algorithms for premature ventricular contraction classification | |
| Emanet | ECG beat classification by using discrete wavelet transform and Random Forest algorithm | |
| CN107622259B (en) | T wave detection method, electrocardiogram data analysis method and device | |
| Luz et al. | How the choice of samples for building arrhythmia classifiers impact their performances | |
| CN112932498A (en) | T wave morphology classification system with strong generalization capability based on deep learning | |
| CN109893119A (en) | A kind of P wave recognition positioning method based on multi-lead clustering | |
| CN115486855A (en) | Electrocardiogram heart beat classification method based on QRS wave group indefinite cycle superposition | |
| CN105877739A (en) | Clinical examination method of electrocardio intelligent analyzing system | |
| De Giovanni et al. | A patient-specific methodology for prediction of paroxysmal atrial fibrillation onset | |
| Villa et al. | Variational mode decomposition features for heartbeat classification | |
| Kovalchuk et al. | A novel feature vector for ECG classification using deep learning | |
| Imtiaz et al. | Cross-Database and Cross-Channel ECG Arrhythmia Heartbeat Classification Based on Unsupervised Domain Adaptation | |
| bin Sinal et al. | Early abnormal heartbeat multistage classification by using decision tree and K-nearest neighbor | |
| Issa et al. | Fully automatic detection of premature ventricular contractions: A new approach based on unsupervised learning | |
| De Chazal | Heartbeat classification system using adaptive learning from selected beats | |
| Kumar et al. | Computerized detection & classification of ECG signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180116 Termination date: 20210522 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |