JP7491466B2 - Heart rate detection method and heart rate detection device - Google Patents

Description

The present invention relates to a heart rate detection method and a heart rate detection device for detecting a heart rate (R wave) from an electrocardiogram waveform.

An ECG (Electrocardiogram) waveform is a measurement of the electrical activity of the heart, and consists of a series of heartbeat waveforms. A single heartbeat waveform is made up of components such as P waves, Q waves, R waves, S waves, and T waves, which reflect the activity of the atria and ventricles. Of these, the R wave is associated with ventricular contraction and has a large amplitude, so heartbeats are often detected based on the R wave.

A simple method for detecting heartbeats (R waves) from an ECG waveform is to detect peaks in a time series. That is, a threshold is set for the time series of data, and when the ECG waveform exceeds the threshold, it is determined to be an R wave.

Figure 5 is a flow chart explaining a conventional heartbeat detection method. When an ECG waveform value is input (step S100 in Figure 5), the heartbeat detection device calculates an index value based on the ECG waveform value and the previous ECG waveform value (step S101 in Figure 5). Next, the heartbeat detection device compares the index value with a threshold value (step S102 in Figure 5), and when the index value exceeds the threshold value, it determines that there is a heartbeat (step S103 in Figure 5), and when the index value is equal to or less than the threshold value, it does not determine that there is a heartbeat. Then, the heartbeat detection device updates the threshold value using the index value and the previous index value (step S104 in Figure 5). By adaptively updating the threshold value, it is possible to perform heartbeat detection that reflects changes in signal level due to the state of attachment of the electrodes to the subject, etc.

One such conventional heartbeat detection method is the method disclosed in Patent Document 1. In the method disclosed in Patent Document 1, if the index value is unlikely to be derived from the ECG waveform, the index value is not reflected in the threshold. This has the advantage of preventing the threshold from jumping up due to large noise mixed into the ECG waveform, allowing for accurate heartbeat detection.

However, the method disclosed in Patent Document 1 had the potential to erroneously recognize waveforms that had small amplitudes and were considered inappropriate for R waves as R waves.

Patent No. 6652655

The present invention has been made to solve the above problem, and aims to provide a heartbeat detection method and a heartbeat detection device that do not mistakenly recognize peaks with too small amplitude as heartbeats.

The heartbeat detection method of the present invention includes a first step of calculating an index value for each sampling time from a sampling data sequence of an electrocardiogram waveform of a living body, a second step of comparing a threshold value for heartbeat detection with a lower limit value of a plausible range for the threshold value derived from the electrocardiogram waveform of the living body, and a third step of determining in the second step that the threshold value is equal to or greater than the lower limit value, and when a peak of the index value exceeding the threshold value is detected, setting the sampling time of this peak as the heartbeat time, and is characterized in that when the threshold value is determined to be less than the lower limit value in the second step, heartbeat detection is not performed by comparing the threshold value with the index value.

The heartbeat detection device of the present invention further comprises an index value calculation unit configured to calculate an index value for each sampling time from a sampling data sequence of an electrocardiogram waveform of a living body; a threshold determination unit configured to compare a threshold for heartbeat detection with a lower limit value of a plausible range for the threshold derived from the electrocardiogram waveform of the living body; and a heartbeat time determination unit configured to determine that the threshold is equal to or greater than the lower limit value by the threshold determination unit and that when a peak of the index value exceeding the threshold value is detected, the sampling time of this peak is set as the heartbeat time, and the heartbeat time determination unit is characterized in that when the threshold value is determined to be less than the lower limit value, it does not detect a heartbeat by comparing the threshold value with the index value.

According to the present invention, when the threshold value is less than the lower limit, the heartbeat is not detected even if the index value exceeds the threshold value, so that peaks of index values with too small amplitude can be excluded, enabling accurate heartbeat detection.

FIG. 1 is a block diagram showing the configuration of a heart rate detection device according to an embodiment of the present invention. FIG. 2 is a flow chart illustrating a heartbeat detection method according to an embodiment of the present invention. FIG. 3 is a flowchart illustrating the operation of the threshold setting unit of the heart rate detection device according to the embodiment of the present invention. FIG. 4 is a block diagram showing an example of the configuration of a computer that realizes a heartbeat detection device according to an embodiment of the present invention. FIG. 5 is a flowchart illustrating a conventional heartbeat detection method.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Fig. 1 is a block diagram showing the configuration of a heart rate detection device according to an embodiment of the present invention, and Fig. 2 is a flowchart explaining a heart rate detection method according to an embodiment of the present invention. the maximum value detection unit 4 detects, for each sampling time, the maximum value of the positive/negative inversion value in a certain time range before the sampling time to be processed and the positive/negative inversion value in a certain time range after the sampling time to be processed; an index value calculation unit 5 calculates, for each sampling time, an index value obtained by subtracting the maximum value from the positive/negative inversion value in a sampling time to be processed; a threshold determination unit 6 compares a threshold for heartbeat detection with a lower limit value of a plausible range as a threshold derived from the ECG waveform of a living body; a heartbeat time determination unit 7 which, when the threshold is determined to be equal to or greater than the lower limit value and a peak of the index value exceeding the threshold is detected, sets the sampling time of the peak as the heartbeat time; and a threshold setting unit 8 which updates the threshold based on the peak of the index value exceeding the current threshold.

The maximum value detection unit 4 is composed of a FIFO buffer (First In, First Out) 40 which receives as input the time difference positive/negative inversion value calculated by the positive/negative inversion value calculation unit 3, a FIFO buffer 41 which receives as input the output value of the FIFO buffer 40, a FIFO buffer 42 which receives as input the output value of the FIFO buffer 41, and a detection processing unit 43 which detects the maximum value of the time difference positive/negative inversion value stored in the FIFO buffer 40 and the time difference positive/negative inversion value stored in the FIFO buffer 42 for each sampling time.

The index value calculation unit 5 is composed of a FIFO buffer 50 which receives as input the time difference positive/negative inverted value calculated by the positive/negative inverted value calculation unit 3, and a subtraction processing unit 51 which calculates an index value for each sampling time by subtracting the maximum value detected by the maximum value detection unit 4 from the output value of the FIFO buffer 50.

The heartbeat detection method of this embodiment will be described below. Here, the procedure for detecting one heartbeat and obtaining the heartbeat time will be described. By repeating this calculation of the heartbeat time over the period of the ECG waveform data, time series data of the heartbeat time can be obtained.

In this embodiment, the data sequence obtained by sampling the ECG waveform is designated as D(i). i (i=1, 2, ...) is a number assigned to one sample of data. It goes without saying that the larger the number i, the later the sampling time.

The electrocardiograph 1 measures the ECG waveform of a living body (human body) (not shown) and outputs a sampling data string D(i) of the ECG waveform. At this time, the electrocardiograph 1 adds information on the sampling time to each sampling data and outputs it. Note that a specific method for measuring the ECG waveform is a well-known technique, so a detailed description will be omitted.
The storage unit 2 stores the sampling data string D(i) of the ECG waveform output from the electrocardiograph 1 and information on the sampling time.

In order to calculate the time difference positive/negative inversion value Y(i) of the sampling data D(i), the positive/negative inversion value calculation unit 3 acquires the data D(i+1) after one sampling of the sampling data D(i) and the data D(i-1) before one sampling from the storage unit 2 (step S1 in FIG. 2). Then, the positive/negative inversion value calculation unit 3 calculates the time difference positive/negative inversion value Y(i) of the sampling data D(i) for each sampling time as shown in the following equation (step S2 in FIG. 2).
Y(i)=-{D(i+1)-D(i-1)}...(1)

The positive/negative inversion value calculation unit 3 inputs the calculated time difference positive/negative inversion value Y(i) to the FIFO buffer 50 for each sampling time (step S3 in FIG. 2). The input value is held in the FIFO buffer 50 and is used for subtraction processing after a time period equivalent to the size of the FIFO buffer 50 (the delay time from when the time difference positive/negative inversion value is input to the FIFO buffer 50 until it is output).

The positive/negative inversion value calculation unit 3 also inputs the calculated time difference positive/negative inversion value Y(i) to the FIFO buffer 40 for each sampling time (step S4 in FIG. 2). The output of the FIFO buffer 40 is input to the FIFO buffer 41 (step S5 in FIG. 2), and the output of the FIFO buffer 41 is input to the FIFO buffer 42 (step S6 in FIG. 2). The FIFO buffers 40 to 42 are used to find the maximum value of the time difference positive/negative inversion value within a certain time range.

The time interval L3 (the delay time from when the time difference positive/negative inversion value is input to the FIFO buffer 41 until it is output) corresponding to the length of the FIFO buffer 41 must be sufficiently wide relative to the width of the peak derived from the R wave (approximately 10 ms), and is preferably about 50 ms. In addition, the time interval L2 (the delay time from when the time difference positive/negative inversion value is input to the FIFO buffer 40 until it is output) corresponding to the length of the FIFO buffer 40 and the time interval L4 (the delay time from when the time difference positive/negative inversion value is input to the FIFO buffer 42 until it is output, L2 = L4) corresponding to the length of the FIFO buffer 42 are appropriately about 100 ms. In addition, the time interval L1 corresponding to the length of the FIFO buffer 50 may be set to L1 = L2 + L3 / 2. Therefore, in the above numerical example, L1 is 125 ms. By setting L1 = L2 + L3/2 and L2 = L4, it is possible to find the maximum value M in the ranges of -(L2 + L3/2) to -(L3/2) and (L3/2) to (L2 + L3/2) for the time of output value a of the FIFO buffer 50 (the sampling time to be processed), and it is possible to subtract the maximum value M from the output value a.

The detection processing unit 43 detects the maximum value M of the positive/negative inverted time difference values stored in the FIFO buffer 40 and the positive/negative inverted time difference values stored in the FIFO buffer 42 for each sampling time (step S7 in FIG. 2).
The subtraction processing unit 51 calculates an index value b=a−M by subtracting the maximum value M from the output value a of the FIFO buffer 50 for each sampling time (step S8 in FIG. 2).

Next, the threshold determination unit 6 compares the current threshold value Th for heartbeat detection with the lower limit value Thmin of the plausible range of the threshold value Th derived from the ECG waveform of the living body (step S9 in FIG. 2). The plausible range of the threshold value Th derived from the ECG waveform of the living body can be determined based on past measurement results. The initial value of the threshold value Th may be set in advance to any value within the plausible range of the threshold value Th.

When the threshold determination unit 6 determines that the current threshold Th is equal to or greater than the lower limit value Thmin (YES in step S9) and detects a peak of the index value b(i) that exceeds the threshold Th (YES in step S10 in Figure 2), the heartbeat time determination unit 7 determines the sampling time of this peak as the heartbeat time (step S11 in Figure 2).

The sampling time of the index value b(i) refers to the sampling time of the time difference positive/negative inversion value Y(i) on which the index value b(i) is based (the sampling time of the data D(i)). The sampling time of the index value b(i) can be obtained from the memory unit 2.

In addition, if the threshold determination unit 6 determines that the current threshold Th is less than the lower limit value Thmin (NO in step S9), or if it has not detected a peak of the index value b(i) exceeding the threshold Th (NO in step S10), the heartbeat time determination unit 7 does not determine that it is a heartbeat (R wave) and does not determine the heartbeat time.

Next, the threshold setting unit 8 updates the threshold Th based on the peak of the index value b(i) that exceeds the current threshold Th (step S12 in FIG. 2). FIG. 3 is a flowchart illustrating the operation of the threshold setting unit 8 (step S12).

First, the threshold setting unit 8 determines whether a peak of the index value b(i) exceeding the current threshold Th has been detected (step S20 in FIG. 3). If a predetermined number V (V = 5 in this embodiment) or more of peaks have been detected consecutively (YES in step S21 in FIG. 3), the average value bave of the latest predetermined number V of peaks among these peaks is calculated, and the value obtained by multiplying this average value bave by a predetermined coefficient α (for example, α = 0.4) is set as a candidate threshold value Thc (step S22 in FIG. 3).
Thc = α × bave ... (2)

The threshold setting unit 8 compares the calculated candidate threshold Thc with a threshold limit value L based on a difference limit value that can be said to be a plausible time difference positive/negative inversion value of the ECG waveform of the living body and the current threshold Th, and if the candidate threshold Thc exceeds the threshold limit value L (YES in step S23 in FIG. 3), the threshold Th is not updated, and if the candidate threshold Thc is equal to or less than the threshold limit value L (NO in step S23), the candidate Thc is set as the new threshold Th (step S24 in FIG. 3). The threshold limit value L will be described later.

The threshold setting unit 8 performs the above process for each sampling time. If a peak of the index value b(i) exceeding the threshold Th is continuously detected, the process of step S22 is repeatedly executed to sequentially calculate the threshold candidate Thc, and if the threshold candidate Thc is equal to or less than the threshold limit value L, the threshold Th is updated.

Here, if the threshold setting unit 8 cannot detect a peak of the index value b(i) exceeding the threshold Th for 3 seconds (YES in step S25 in FIG. 3), it resets the peak search operation and redoes the setting of the threshold Th (step S26 in FIG. 3). Specifically, the threshold setting unit 8 uses the data of the index value b(i) during the period of 2 seconds from the point in time when the peak search operation is reset to determine its maximum value bmax, and sets the value obtained by multiplying the maximum value bmax by a predetermined coefficient β (for example, β=0.4) as the new threshold Th.
Th = β × bmax (3)

Thus, in this embodiment, time series data of heartbeat time is obtained by repeating the processing of steps S1 to S12 in Figure 2 for each sampling time.

Here, we consider what the maximum time difference positive/negative inversion value of the ECG waveform of a living body (human body) is, or conversely, what value is unlikely to be the time difference positive/negative inversion value of a human ECG waveform. The QRS interval (from the start of the Q wave to the end of the S wave) is 0.06 to 0.1 seconds.

The time from the maximum R wave to the minimum S wave can be considered to be about 15 to 25 ms, which is 1/4 of the QRS interval. Furthermore, the change is steepest near the center between the maximum R wave and the minimum S wave. If this time range is roughly half the time from the maximum R wave to the minimum S wave, it can be estimated to be about 7.5 to 12.5 ms. If there is a change in the cardiac potential equivalent to the QRS amplitude within this time, the rate of change of the cardiac potential is thought to be a maximum of about 350 μV/ms. If this rate of change of the cardiac potential is converted to the time difference positive/negative inversion value calculated using equation (1), it becomes 700 μV when the sampling interval is 1 ms, and is rounded off to 1000 μV to allow for some margin.

From the above, when the sampling interval is 1 ms, a time difference positive/negative inversion value exceeding 1000 μV can be said to be unlikely as a time difference positive/negative inversion value of a human ECG waveform. This guideline is 5000 μV when the sampling interval is 5 ms. In other words, when the sampling interval is dT, the difference limit value X that can be said to be plausible as a time difference positive/negative inversion value of a human ECG waveform can be expressed as follows.
X = dT [ms] × 1000 [μV] ... (4)

Index value b represents the height of the clearance of the time difference value in the time domain surrounding the peak derived from the R wave, and when the time difference positive/negative inversion value of the ECG waveform contains large noise and the noise presents a sharp unimodal peak, in the worst case scenario, the value itself will be reflected in index value b.

If the new threshold value Th' is the average value bave of the peaks of a certain number V (V = 5 in this embodiment) that exceed the current threshold value Th multiplied by a coefficient α = 0.4, and the peak value of the standard index value b is p, then the updated threshold value Th' during normal times will be approximately Th' = p x α = p x 0.4. If abnormal noise with a peak value of X occurs once, the next threshold value Th' will increase by (X-p) ÷ 5 x 0.4 = 0.08X-0.2Th. If X = 5000 μV, it will be 400-0.2Th [μV]. In other words, if the sampling interval is 5 ms, when the increase in the next threshold value Th' from the current threshold value Th exceeds 400-0.2Th [μV], there is a high probability that it is affected by noise.

Therefore, in this embodiment, when the candidate threshold Thc calculated in step S22 exceeds the threshold limit value L=400+0.8Th [μV] (=Th+400-0.2Th) with respect to the current threshold Th (YES in step S23), the threshold Th is not updated. The general formula for the threshold limit value L is as follows:
L = Th + (α/V)X - (1/V) Th ... (5)

As described above, the threshold Th is set based on the value of the peak to be detected multiplied by a coefficient α that is equal to or less than 1. In the method of adaptively updating the threshold Th, when the input signal level decreases and a series of peaks of the index value b having small amplitude occurs, the threshold Th decreases. In a situation in which the threshold Th has decreased to an extent that it is unlikely that the threshold Th is derived from a human ECG waveform, even if a peak of the index value b exists, the probability that this peak is derived from an R wave is low.

Although the method of limiting the threshold value Th so that it does not fall below the lower limit can avoid detecting small amplitude peaks in the index value b, it still reacts to index values b that exceed the threshold value Th. As a result, even peaks in the index value b that are due to noise added to the ECG waveform will be detected as heartbeats.

On the other hand, in this embodiment, when the threshold value Th is less than the lower limit value Thmin, no heartbeat determination (step S10) is performed, so that noise added to the ECG waveform is not mistakenly detected as a heartbeat, and appropriate heartbeat detection can be achieved.
The method of updating the threshold value described with reference to FIG. 3 is just one example, and the threshold value may be updated by another method.

The memory unit 2, positive/negative inversion value calculation unit 3, maximum value detection unit 4, index value calculation unit 5, threshold determination unit 6, heartbeat time determination unit 7, and threshold setting unit 8 of the heartbeat detection device described in this embodiment can be realized by a computer equipped with a CPU (Central Processing Unit), a storage device, and an interface, and a program that controls these hardware resources. An example of the configuration of this computer is shown in Figure 4.

The computer includes a CPU 200, a storage device 201, and an interface device (I/F) 202. An electrocardiograph 1 and the like are connected to the I/F 202. A program for implementing the heart rate detection method of the present invention is stored in the storage device 201. The CPU 200 executes the processing described in this embodiment in accordance with the program stored in the storage device 201.

The present invention can be applied to technology for detecting the heart rate of a living body.

1...electrocardiograph, 2...memory unit, 3...positive/negative inversion value calculation unit, 4...maximum value detection unit, 5...index value calculation unit, 6...threshold determination unit, 7...heartbeat time determination unit, 8...threshold setting unit, 40-42, 50...FIFO buffer, 43...detection processing unit, 51...subtraction processing unit.

Claims (6)

A first step of calculating an index value for each sampling time from a sampling data sequence of an electrocardiogram waveform of a living body;
a second step of comparing a threshold value for detecting a heartbeat with a lower limit value of a plausible range of the threshold value derived from an electrocardiogram waveform of the living body;
a third step of determining in the second step that the threshold value is equal to or greater than the lower limit value, and when a peak of the index value exceeding the threshold value is detected, determining a sampling time of the peak as a heartbeat time,
a step of detecting a heartbeat by comparing the index value with the threshold value when the threshold value is determined to be less than the lower limit value in the second step; 2. The method for detecting a heartbeat according to claim 1,
a fourth step of calculating a positive/negative inversion value of a time difference of sampled data from the sampled data string of the electrocardiogram waveform for each sampling time;
a fifth step of detecting, for each sampling time, a maximum value of the positive/negative inverted value in a certain time range before the sampling time to be processed and a maximum value of the positive/negative inverted value in a certain time range after the sampling time to be processed,
The heart rate detection method according to the present invention, wherein the first step includes a step of calculating, for each sampling time, a subtraction value obtained by subtracting the maximum value from the positive/negative inverted value at the sampling time to be processed as the index value. 3. The heartbeat detection method according to claim 1, further comprising:
A method for detecting heartbeats, comprising the further step of updating the threshold based on peaks of the index value that exceed a current threshold. an index value calculation unit configured to calculate an index value for each sampling time from a sampling data sequence of an electrocardiogram waveform of a living body;
a threshold value determination unit configured to compare a threshold value for detecting a heartbeat with a lower limit value of a plausible range of the threshold value derived from an electrocardiogram waveform of a living body;
a heartbeat time determination unit configured to determine, when the threshold determination unit determines that the threshold is equal to or greater than the lower limit and detects a peak of the index value that exceeds the threshold, a sampling time of the peak as a heartbeat time;
The heartbeat detection device, wherein the heartbeat time determination unit does not detect a heartbeat by comparing the threshold value with the index value when the threshold value is determined to be less than the lower limit value. 5. The heartbeat detection device according to claim 4,
a positive/negative inversion value calculation unit configured to calculate a positive/negative inversion value of a time difference of sampled data from the sampled data string of the electrocardiogram waveform for each sampling time;
a maximum value detection unit configured to detect, for each sampling time, a maximum value of the positive/negative inverted value in a certain time range before a sampling time to be processed and a maximum value of the positive/negative inverted value in a certain time range after the sampling time to be processed,
The heart rate detection device, wherein the index value calculation unit calculates, for each sampling time, a subtraction value obtained by subtracting the maximum value from the positive/negative inverted value at the sampling time to be processed, as the index value. 6. The heartbeat detection device according to claim 4,
The heartbeat detection device further comprises a threshold setting unit configured to update the threshold based on a peak of the index value that exceeds a current threshold.

Images