JPS6156633A - Apparatus for detecting abnormality of electrocardiograph - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Field of Industrial Application This invention relates to an electrocardiographic abnormality detection device that can automatically detect abnormalities in an electrocardiogram.

(b) Conventional technology In general, a normal electrocardiogram consists of P, Q,
It consists of RSS and T parts. Therefore, if the electrocardiogram is observed in time sequence and a waveform as shown in Figure 3 is obtained, waveforms A and B are normal waveforms, but
Waveform C becomes an abnormal waveform.

In recent years, electrocardiograms have been stored time-sequentially on storage means such as magnetic tape, and after recording they can be processed and analyzed at high speed, or played back on a CRT screen to macroscopically check whether there are any abnormalities. Electronic monitoring devices are emerging.

(C) Problems to be Solved by the Invention None of the above-mentioned conventional electrocardiogram monitoring devices is capable of determining the presence or absence of an abnormality in real-time processing while observing an electrocardiogram.
In order to determine the presence or absence of abnormalities, so-called batch processing was performed after collecting a large amount of data. Therefore, even if an abnormality occurs, it cannot be reported in real time, and treatment may be delayed depending on the patient's symptoms.

An object of the present invention is to provide an electrocardiogram abnormality detection device capable of detecting electrocardiogram abnormalities in real time.

(d) Means and operation for solving the problems As shown in FIG. An electrocardiographic waveform storage means 2 that updates and stores the shape of the electrocardiographic waveform in time order, a standard electrocardiographic waveform storage means 3 that stores a standard electrocardiographic waveform, and a detected electrocardiographic waveform stored in the electrocardiographic waveform storage means and the standard electrocardiographic waveform. The electrocardiographic waveforms are compared to determine whether they are similar or not, and if they are dissimilar, the detected electrocardiographic waveforms are determined to be abnormal waveforms.

This electrocardiographic abnormality detection device detects an electrocardiographic waveform, compares the standard electrocardiographic waveform and the detected electrocardiographic waveform in real time, and determines the presence or absence of an abnormal waveform.

(e) Example 11 The present invention will be explained in more detail with reference to Examples below.

FIG. 4 is a block diagram of an electrocardiographic abnormality detection device showing one embodiment of the present invention. In the figure, reference numerals 11a and llb are biological electrodes for detecting electrocardiogram signals generated from the human body. It is added to the amplifier 13. The high-pass filter 12 is
It consists of a capacitor and a resistor, and is provided to remove baseline fluctuations that are unrelated to the electrocardiogram signal (that is, to remove low frequency components).

The differential amplifier 13 amplifies the electrocardiogram signal from which low frequency components have been cut, and inputs the amplified electrocardiogram signal to the A/D converter 15 through the low-pass filter 14 . Normally, an electrocardiogram signal has a level of only about 1 mV, so if the input impedance of the device is small, it is easily affected by noise.

Therefore, the input impedance of the differential amplifier 13 is set to be sufficiently large.

From the input of the A/D converter 15, frequency components of two or more sampling frequencies must be removed in accordance with the sampling theorem when performing A/D conversion. To remove frequency components above each of this sampling frequency,
A low-pass filter 14 is provided on the input side of the A/D converter 15.

The sampling frequency of the A/D converter 15 is selected to be 200H2 or more based on the sampling theorem since the frequency component of an electrocardiogram is approximately 100H2 at most. Also, although 8-bit resolution is used, 12-bit resolution may also be used.

The electrocardiogram signal converted into a digital signal by the A/D converter 15 is stored in a RAM 17, which is a semiconductor memory, under the control of a CPU (microcomputer) 16.

The electrocardiogram data storage area of the RAM 17 is configured in an endless loop with a ring-like structure. In addition to this, the RAM 17 has many storage areas that will be described later. An abnormality detection program is stored in the ROM 18, which is a semiconductor memory, and the CPU 6 performs abnormality detection processing on the electrocardiogram waveform according to this program. The buzzer 20 is designed to issue a sound alarm. In addition, if necessary, the required information can be transmitted to the voice synthesis circuit 21 and the speaker 22.
It is output after passing through.

Next, the abnormality detection processing operation of the above-mentioned embodiment device will be explained with reference to flowcharts shown in FIGS. 5 to 9.

The flowcharts in FIGS. 5 to 9 show the abnormality detection algorithm. In parallel with this abnormality detection processing, the device executes a data import program and
The electrocardiogram data storage area 7 always stores the latest electrocardiogram signal data.

When the operation of the apparatus starts, as shown in FIG. 5, first, parameters such as the end time of electrocardiogram monitoring are set [step ST (hereinafter abbreviated as ST) 1]. and,
Another data import program is started, and i starts storing electrocardiogram data xi (storage area of RAM 17) (Sr2).

Next, pass 1 and pass 2 processing (Sr1, ST4) is performed as preprocessing to search for the position of the QRS complex (hereinafter referred to as QR3) in the electrocardiogram for abnormality detection.

Pass 1 is a routine for determining the difference value of the electrocardiogram data xi, and its specific processing procedure is shown in FIG. Pass 2 is a routine to find the maximum difference value, and its specific processing procedure is shown in FIG.

When entering pass 1, HOB is stored in counter i (Sr31), as shown in FIG. Note that HOB here indicates the beginning position of the data to be detected, and TO
B indicates the last position of the data to be detected.

Next, Sr12 calculates the difference value Δxi between electrocardiogram data xi+l and xi specified by counter i, stores this in RAM 17, and then increments counter i by 1 (Sr33). The process of calculating the difference value Δxi between the data at adjacent positions i+l and i is repeated until the content becomes TOB (Sr34). 5T34, i
When =TOB, it means that the detection of the difference value in the detection target section has been completed, pass 1 is removed, and the process moves to pass 2.

When entering pass 2, as shown in FIG. 7, HOB is stored in the counter i again, and the area of the RAM 17 that stores the maximum value Δx max of the difference value ΔX is set to 0 (Sr41). Then, the magnitude relationship between the stored maximum value Δx max and the difference value Δxi is compared (Sr42).

If Δxi is larger as a result of comparison, that Δxi is stored as a new maximum difference value Δx max (Sr4
3), if Δxi is not larger than the maximum value ΔXmax up to that point, 5T43 is skipped and counter i is incremented by +1 (Sr44), and 5T42 to 5T44 are executed until i reaches T OB (Sr45). The process is repeated to find the maximum value Δx max of the difference values in the section from HOB to TOB. As a result, the largest change within the target section is determined.

When the value of counter i reaches TOB, “1=
The judgment of “Is it ros?” becomes YES, and the obtained Δxm
A value TL obtained by multiplying ax by 0.6, that is, a 60% value of the maximum value of the difference value is obtained (Sr46).

This will allow you to pass pass 2. This TL is used as a trigger level for finding the electrocardiogram waveform, ie, QR5, and 0.6 is determined empirically, and may be set to other values as the case requires.

Returning again to the main routine of FIG. 5, following pass 2, it is determined in Sr1 whether or not the set time has elapsed.

This set time is for performing abnormality detection every set time, and here the set time is updated every second. In other words, the determination of Sr1 becomes YES every second, and the next Sr1
QR3 detection processing is performed in r6.

As shown in FIG. 8, the QR3 detection routine first detects HO.
The sum of B and the offset value Os is set as i, and the counter j (the count value is the number of QR3s detected in the target section) is set as O (Sr61).

Then, it is determined whether the difference value Δxi is greater than or equal to TL (5T62), and if the difference value Δxi is smaller than TL, QR
3 is not detected, +1 processing is performed on i (Sr63
), “Is i = TOB?” Sr64), i is T
If it is not OB, it returns to 5T62 and Δxf≧TL, but it continues from 5T62 to 5T64 until i = TOB.
Repeat the process.

When a difference value Δxi greater than or equal to TL appears, the determination of 5T62 becomes YES, which means the detection of QR3, and with that point 1 as the QR3 detection position, yj(yo ), the counter j is incremented by +1, ARP is added to i at that point, and the result is stored as a new i (Sr65). Here, ARP takes into consideration the absolute refractory period, and since the contraction does not occur even if stimulation is applied during this period, QR3 detection processing is skipped. Following 5T55, “i≧T OB
5T66), if i is not ToB, 5T
Returns to 61 and again depending on the judgment situation Sr61~5
Perform T66 processing. If 1=TOB,
When Δxi≧TL and the second QR3 is detected,
The judgment of 5T62 becomes YES, and i at that point becomes Yj=(
Yl) is stored as yl, and j advances +1 step (
j=2), and ARP is further added to i at that point.

As described above, the QRS position is stored in Yj.

Note that when 1=TOB is determined in 5T64, the offset value Os is set to 0 (Sr67), but in 5T66, i
When ≧TOB is determined (j4-1+ARP, i>
TOB), the amount by which i exceeds TOB, i.e., i - TOB, is stored as an offset value O3 (Sr68).

When the QRS detection process is completed, in the main routine shown in FIG. 5, it is subsequently determined whether j=0 (Sr1).
. If QRS is detected in this process, since j is not 0, the determination of Sr1 is No, and QRS classification processing is then performed (Sr1). This QR8 classification process is performed to classify which waveform the detected electrocardiographic waveform, ie, QRS, belongs to.

As shown in Fig. 9, the QRS classification routine first sets template i' to 1 (5T81), then compares the current input data (detected QRS) j and template i' to 5T.
82), and determine whether they are similar (Sr83). Template i° is a standard electrocardiographic waveform to be compared with the detected electrocardiographic waveform, and its waveform data is registered in the template storage area of the RAM 17. At the beginning of operation, i'
= 0, but as the operation progresses, new electrocardiographic waveforms will be detected and registered, so the template i' will increase.

For input data j, if j=1, the waveform is the template i, and if j-2, the second detected electrocardiogram waveform is the template i.
” will be compared.

Comparison of template i'' and input data j is performed by determining the correlation coefficient of both waveforms.

If the correlation coefficient is 0.9 or more, it is determined that the two waveforms have a correlation and are similar; otherwise, they are determined to be dissimilar.

However, if the two waveforms are at different positions even if they are the same electrocardiogram waveform, the correlation coefficient will be 0.9 or less and it will be determined that they are dissimilar, so if a NO determination is made at 5T83, the contents of the counter will be +1 step until k reaches the predetermined value R+1 (
Sr85), and repeats the similarity determination of 5T83. That is, the detected electrocardiographic waveform is shifted with respect to the template i' up to a maximum of R+1, and the correlation coefficient is determined at each shifted position. If the highest value of the correlation coefficient is 0.9 or more during -R+1 processing, both waveforms are said to be similar. Since the similarity is determined by relatively shifting both waveforms in this way, it is possible to perform a similarity determination with higher precision than when comparing only once.

If it is determined in 5T83 that both waveforms are similar, the counter (counting the number of each template) corresponding to template i° is incremented by 1 and updated (Sr86).
, both waveforms are similar, 5T83 judgment is No, 5T8
5 (7) If the judgment is YES, add 1 to i' (S
r15), it is determined whether there are any more templates i° to be compared (Sr88). If there are still templates to be compared, the determination in 5T88 is YES, the process returns to 5T82, the input data j and the new template i° are compared, and it is determined whether they are similar (Sr83.5T85) , here, if both data are similar, +1 is added to the counter corresponding to that template (Sr86), but if both are not similar and there is no template to compare (Sr83.578).
8) registers the human data j as a new template (Sr89). Then, the buzzer 20 is sounded (Sr90). The sound of the buzzer 20 means the appearance of a new waveform, and if the already registered template is a normal waveform (standard waveform), the sound of the buzzer 20 indicates the detection of an abnormal waveform. .

When the QRS classification is completed, the main routine moves to Sr9, processes j=j-1, and returns to Sr1. Then, a determination is made as to whether "j=0". If two or more QRSs are detected in the current QR5 detection, since j=0 is not the case, classification processing is performed on the remaining QRSs until j=0.

If the determination of "j = 0" is YES in ST7, it is determined whether the end time has been reached (5TIO), and unless it is the end time, the process returns to ST5 and the QR3 in the next 1 second is determined.
Detection and QR3 classification processing is performed, and thereafter, the processing is continued until the end time is reached.

Through the above series of processes, the QR9s obtained in time order are classified into similar waveforms. Once this classification has progressed to a certain extent, if a waveform that rarely appears appears, it may be determined that this is an abnormal waveform.

Furthermore, in the case of a patient who has undergone several rounds of determination, a normal waveform may be stored in the ROM 18 in advance as a template, and if a QR3 that differs from this template appears, it may be determined to be abnormal.

(f) Effects of the Invention According to the electrocardiographic abnormality detection device of the present invention, abnormal waveforms are detected by monitoring electrocardiographic waveforms in time sequence, comparing detected electrocardiographic waveforms with standard waveforms, and determining similarity. Because it is a real-time device, anomalies can be detected in real time.

can.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the general structure of the electrocardiographic abnormality detection device of the present invention, FIG. 2 is a diagram showing a normal electrocardiographic waveform, and FIG. 3 is a diagram showing a normal electrocardiographic waveform.
The figure shows an example of an electrocardiographic waveform including an abnormal waveform, FIG. 4 is a block diagram of an electrocardiographic abnormality detection device showing one embodiment of the present invention, and FIGS. 5 to 9 show a concentric electrocardiographic abnormality detection device. This is an operation flowchart, and FIG. 5 is a flowchart showing the main routine. Figure 6 is a flowchart showing the details of pass 1 of the main routine, Figure 7 is a flowchart showing the details of pass 2 of the main routine.
8 is a flowchart showing details of the QR3 detection routine of the main routine,
FIG. 9 is a flowchart showing details of the QR5 classification routine of the main routine. 2: Electrocardiogram waveform storage means; 3: Standard electrocardiogram waveform storage means; 4: Comparison means Patent applicant Tateishi Electric Co., Ltd. agent
Patent Attorney Shigeru Nakamura Figure 2 Figure 3 Figure 4 Figure 6 Figure 7 Figure 8

Claims (2)

[Claims] (1) Electrocardiographic waveform detection means for detecting electrocardiographic waveforms, electrocardiographic waveform storage means for updating and storing detected electrocardiographic waveforms in time sequence, and standard electrocardiographic waveform storage means for storing standard electrocardiographic waveforms. and comparing the detected electrocardiographic waveform stored in the electrocardiographic waveform storage means with the standard electrocardiographic waveform to determine similarity, and in the case of dissimilarity, the detected electrocardiographic waveform is determined to be an abnormal waveform. An electrocardiographic abnormality detection device comprising means. (2) The comparison means includes means for calculating a correlation coefficient between the detected electrocardiogram waveform and the standard electrocardiogram waveform, and determines similarity based on whether the correlation coefficient is greater than or equal to a predetermined value. An electrocardiographic abnormality detection device according to claim 1.