KR0133428B1 - Data processing method of cardiogram time series - Google Patents

Abstract

The present invention relates to a method for processing electrocardiogram time series data, and in the related art, it is necessary to stabilize the baseline and then reconfigure the attractor when measuring the electrocardiogram of the examinee. It takes a long time until the diagnosis takes a long time. Therefore, the present invention measures the electrocardiogram of the human body, extracts the ECG time series, finds the maximum point from the extracted ECG time series, finds the left and right local maximum points based on the maximum point, and then between the maximum point and the left and right local maximum points. Adjust the points of the EKG to adjust the baseline of the ECG time series, and when the adjustment is complete, reconstruct the ECG attractor from the adjusted time series to check for health and accurate diagnosis.

Description

ECG Time Series Data Processing Method

1 is a block diagram of an ECG time series data processing apparatus of the present invention.

2 is an electrocardiogram time series and attractor reconstruction diagram.

(a) is a reconstruction diagram of ECG time series and attractor before filtering,

(b) is a reconstruction diagram of the ECG time series and the attractor after filtering.

3 is a timing diagram of the ECG time series in FIG.

(a) is a timing chart before baseline filtering,

(b) is a timing chart after baseline filtering.

4 is an electrocardiogram baseline filtering adjustment diagram according to FIG.

(a) is an adjustment of the maximum point and the local maximum point adjacent to the left and right,

(b) is the degree of adjustment between the maximum point and the adjacent local maximum point.

5 is a signal flow diagram for the ECG time series data processing method of the present invention.

6 is an attractor configuration.

* Explanation of symbols for main parts of the drawings

10: electrocardiogram 20: data acquisition board

30: personal computer 40: time series extraction unit

50: baseline adjustment unit 60: adjusted time series extraction unit

70: reconstruction section 80: diagnostic section

90: database 100: real time time series display

The present invention relates to electrocardiogram time series data processing, and more particularly, to a method for processing electrocardiogram time series data for baseline filtering of electrocardiogram time series data for reconstructing an accurate electrocardiogram attractor of an examinee.

ECG time series oscillates the baseline up and down, so ECG attractor reconstruction is not an accurate attractor of examinee without stabilizing the baseline horizontally with the centerline. It may be.

As a result, it is necessary to reconstruct the attractor after stabilizing the baseline, but it takes a long time to stabilize the examinee's baseline until it becomes a certain physical and psychological state of calm, so that the diagnosis takes a long time. In addition, even if the baseline is stabilized to some extent, some degree of vibration is inevitable.

Therefore, an object of the present invention is to receive the electrocardiogram of the examinee in real-time (real-time) ECG time series baseline signal processing that makes the baseline horizontal to the center line to enable accurate and rapid diagnosis of ECG time series data processing In providing a method.

It is another object of the present invention to measure the electrocardiogram of a human body using an electrocardiogram and reconstruct an ECG attractor from a examinee's heartbeat time series in real time under the control of a personal computer, thereby enabling the examinee's health diagnosis An electrocardiogram time series data processing method is provided.

The present invention for achieving the above object is to obtain an electrocardiogram time series by measuring the electrocardiogram of the human body, finding the maximum point of the electrocardiogram time series obtained in the step, and the left and right by using the maximum point found in the step Finding a local maximum point, adjusting a point between the maximum point and the local maximum point in the step to adjust the baseline of the ECG series, and reconstructing the ECG attractor from the adjusted time series Consists of steps.

The configuration for realizing the above method is a digital electrocardiogram (10) for detecting the heartbeat sound of the human body as shown in Figure 1 and transmits a signal accordingly, and the digital electrocardiogram signal transmitted from the electrocardiogram (10) A data acquisition board 20 for converting a signal into a memory and storing it in a memory, a personal computer 30 for controlling each unit so as to read and display data from the data acquisition board 20 and to extract time series; Under the control of the personal computer 30, the time series extraction unit 40 converts the data stored in the data acquisition board 20 according to the format of the program, and the time series extracted by the time series extraction unit 40 according to the center line. A baseline adjuster 50 for adjusting, an adjusted time series extractor 60 for extracting a time series adjusted through the baseline adjuster 50, and the time series weight A reconstruction unit 70 for reconstructing the ECG attractor with the time series extracted by the unit 60 and an ECG attractor configured by the reconstruction unit 70 based on the ECG attractor stored in the database 90. And a diagnosis unit 80 for determining health status, and a real time time series display unit 100 displaying a time series of an electrocardiogram in real time under the control of the personal computer 30.

The present invention configured as described above will be described in detail with respect to the operation and effect.

The electrocardiogram, or pulsating sound, of the human body is a chaos signal, which is seemingly random at first, but is actually a system determined by complex laws of dynamics. When shown, it forms a unique topology.

The idea of applying ECG attractors to human health diagnosis is that although there are some differences between people, there is a clear difference between the normal and the abnormal.

Therefore, when the ECG attractor is applied to the human health diagnosis, there is a slight difference for each person, but there is a clear difference between the normal person and the abnormal person.

In order to detect the heartbeat sound of the examinee, the two arms and the ankles are attached to the input unit of the ECG, and the CPI 104 of the ECG 10 that detects the heartbeat of the examinee controls the light emitting diode in the sensor unit 101. When an infrared ray having a wavelength of 940 mm is generated from the light emitting diode and reaches the capillary at the fingertip part of the human body, the reflected light represented by the amount of oxidized hemoglobin due to the volume change of the blood vessel is received by the phototransistor.

Then, when the phototransistor outputs the time series of which the received ECG is changed to the voltage to the amplifier 102, the phototransistor is sufficiently amplified by the amplifier 102 and then output to the data acquisition board 20.

Accordingly, the analog / digital converter 201 of the data acquisition board 20 converts the analog time series into digital data and outputs the digital data to the personal computer 30, and outputs the same to the time series acquisition control unit 202. The electrocardiogram time series is visualized through the real-time time series display unit 100 in real time so that the examiner can easily understand, and the electrocardiogram time series acquisition control unit 202 acquires the electrocardiogram time series and stores it in the memory 203.

Here, the sampling frequency on the data acquisition board 20 is fixed at 700KHz, and the user can select from 60 to 900Hz by software. The ECG time series data is 100-250Hz and the number of data is 15,000.

Then, the time series extractor 40 extracts the time series data of the electrocardiogram from the memory 203 of the data acquisition board 100 and transmits the time series data to the baseline adjuster 50.

At this time, the EKG time series of the examinee displayed on the real-time time series display unit 100 is operated up and down around the baseline (L _B ) as shown in (a) of FIG. 2. It is known to be caused by the state of motion or psychological factors. When reconstructing this by ECG time series, the criteria of each point in the phase space are different so that the trajectory is missed as shown in (a) of FIG. It is difficult to see the wrong person and can mistakenly mistake a normal person's attraction for being an abnormal person.

Accordingly, the baseline adjusting unit 50 performs baseline filtering on the time series data input from the adjusted time series extracting unit 60, which is performed in the following manner based on FIG. 4.

First, find the maximum point (a) of the acquired ECG time series.

Second, the maximum point found is used as a reference point for adjustment.

Third, the local maximum point (c, d) is found from the left and right as shown in (a) of FIG. 4 or around the reference point as shown in (b) of FIG.

Fourth, find the difference between the maximum point (a) and the local maximum point (c, d) (a-c, a-d, ...) and the number of points (N, M, ...) obtained in the third method.

Fifth, using the values obtained in the fourth, adjust the ECG time series as follows. That is, as shown in (a) of FIG. 3, when the local maximum point (c, d) is centered on the maximum point (a), the adjustment value

Becomes

Finally, find the difference between the maximum point and the local maximum point, and then divide the difference by the number of points in the number of N points by the number of points, then add the original value to the division to get the value you want to adjust. Lose.

Further, when the local maximum point c, d is located behind the maximum point a as shown in (b) of FIG. 3, around the maximum point a, the adjustment value is obtained by adjusting as follows.

Through the above process, the ECG time series as shown in FIG. 2A stabilizes the baseline L _B horizontally with the horizontal line L _{C as} shown in FIG. When the stable electrocardiogram time series is transferred to the reconstruction unit 70 through the adjusted time series extracting unit 60, the reconstructing unit 70 performs the attractor reconstruction to derive the attractor of the examinee. Let's look at configuration and reconfiguration.

The chaos system belongs to a deterministic nonlinear dynamics system. The dynamics system is a law that shows how the system evolves over time. It is called essential information to characterize the system.

Then, the state equation of the dynamics system is almost a solution of a differential or differential equation, the solution of the dynamics system, or the trajectory in the state plane, and the place where the system's behavior ultimately settles is called an attractor.

Although it is impossible for us to derive all the state variables and parameter values of a particular chaotic system in nature and derive a state equation for that system, we can experimentally time-series the evolution of only one state variable that has a significant effect on a particular system. When measured, it was mathematically proved by Taken in 1981 that reconstructing the attractor from its time series resembles the original attractor, and the attractor reconstruction is based on the kinematic equations of a particular chaotic system. Although deriving is impossible, it is a very important and useful qualitative analysis method for experimenters.

The qualitative analysis method refers to a method of analyzing a system in the state space in the form of a path in which the trajectory moves over time and the shape of the overall trajectory. Thus, the attractor reconstruction is done with an embedding dimension and a delay with the sampled time series. That is, time series data: {X ₁ , X ₂ , X ₃ , X ₄ ,..., X _N } (where N is a sufficiently large value)

If you construct an m-dimensional vector from the measured time series data X (t)

i = 0, 1, 2, 3, ... N

m = Embedding Dimension

p = delay time (appropriate value)

{X ₁ , X ₂ , X ₃ , X ₄ ,..., X _n } {X (t), X (t + p), X (t + 2p), ..., X (t + (m-1) p}, and the embedding dimension (m) is the For example, when m is 2, when m = 3 on a two-dimensional plane, it is reconstructed in three-dimensional space, but usually on a two-dimensional plane.

For example, if m = 2, p = 3

Time series data such as {X ₁ , X ₂ , X ₃ , X ₄ ,..., X _N - ₁ , X _N }

In the case of partitioning into two-dimensional phase space space in turn, it has the same configuration as in FIG. 6 and this is the attractor reconstruction of a specific case system.

Such a thing becomes a normal person.

When the retractor 70 configures the attractor and outputs the diagnosis unit 80, the diagnosis unit 80 compares the data of the database 90 to diagnose health.

Referring again to the above-described process based on the flowchart of FIG. 5, the electrocardiogram time series is extracted by measuring the electrocardiogram of the human body, and then the maximum point is found from the extracted electrocardiogram time series, and the left and right local maximums are determined based on the maximum point. After finding the point, adjust the baseline of the ECG time series by adjusting the points between the maximum and left and right local maxima, and when the adjustment is complete, reconstruct the ECG attractor from the adjusted time series to check for health. .

As described in detail above, the present invention has an effect of making the baseline and the centerline horizontal so that baseline signal processing of the ECG time series can be accurately and quickly diagnosed.

Claims (2)

A first step of extracting an electrocardiogram time series by measuring an electrocardiogram of a human body; a second step of finding a maximum point of an electrocardiogram time series extracted in the first step; and a left and right local part based on the maximum point found in the second step. A third step of finding the maximum point, and a fourth step of adjusting the baseline of the ECG time series by adjusting the points between the maximum point and the local maximum points in the third step. Electrocardiogram time series data processing method comprising the fifth step of reconstructing. The method of claim 1, wherein the method for adjusting the points between the maximum point and the local maximum point comprises the steps of: obtaining a difference between the maximum point and the local maximum point; and obtaining the difference obtained from the step between the maximum point and the local maximum point. Obtaining a number corresponding to the number of numbers in the field, multiplying the number, dividing the value multiplied by the step, and dividing the number of points between the maximum point and the local maximum point; ECG time series data processing method characterized in that to obtain the adjustment value by adding the value of.