KR20170086740A

KR20170086740A - Method for processing heartbeat sound in smart phone

Info

Publication number: KR20170086740A
Application number: KR1020160005906A
Authority: KR
Inventors: 조정권
Original assignee: 주식회사 시그테크
Priority date: 2016-01-18
Filing date: 2016-01-18
Publication date: 2017-07-27

Abstract

There is provided a method of amplifying the heartbeat sound to an appropriate level regardless of the type of the smartphone, measuring the heartbeat and listening to the heartbeat sound. In the present invention, an impulse of a heartbeat is detected from an input signal, and a relatively high gain is multiplied to an impulse detected section, and a relatively low gain is multiplied to a section in which an impulse is not detected. It is possible to perform a smoothing process after multiplying the gain, and the clipping can be prevented by compressing the input signal and the smoothed output signal before detecting the impulse. According to the present invention, clipping does not occur even when the input signal is large depending on the type of the input signal because the input signal is compressed, and the signal compression is performed with different slopes according to the input level, so that no problem occurs even when the input signal is small . Since the impulse-less section is less amplified than the impulse section, a low-pitched heartbeat can be obtained.

Description

[0001] The present invention relates to a method for processing a heartbeat sound in a smartphone,

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a method of processing heartbeat sounds in a smartphone, and more particularly, to a method of processing a heartbeat sound by amplifying a heartbeat sound to an appropriate level regardless of a type of smartphone, .

As the need for telemedicine becomes more important, more and more attempts are being made to listen to heartbeat using smartphones.

For example, in Patent Publication No. 10-2014-0040186, a digital stethoscope digitally converts a recorded stutter sound from a patient to a smartphone, and the smartphone transmits the stutter sound data to the server, In addition, the smartphone is capable of performing functions such as outputting a chart of a stethoscope waveform and playing back a stethoscope sound.

Instead of using a separate digital stethoscope as described above, the utility model registration No. 20-0389343 allows a mobile phone to be used as a stethoscope by attaching a stethoscope microphone to the mobile phone. Alternatively, a stethoscope may be connected to an earphone terminal or a separate dedicated terminal of a mobile phone as in Patent Registration No. 10-1557793 or Utility Model Registration No. 20-0384551.

When the stethoscope is connected to the smartphone and the analog stethoscope input through the stethoscope is converted into digital data from the smartphone, the input signal to the smartphone is very small, so the input signal must be multiplied by a large gain It can be used as a level of data and there is a problem that the basic noise of the hardware or the environmental noise is also amplified during amplification of the stethoscope sound. If the noise is amplified together, a clear heartbeat can not be obtained, and the probability of error in measuring the heart rate from the heartbeat sound increases.

Also, since the tuning state of the microphone mounting mode or the codec is different for each smartphone, the size of the input of the stethoscope sound to the smartphone may be several hundred times to several thousand times different depending on the model. Therefore, in the case of a model in which a small stethoscope is input, there is a problem that even if the volume button is raised to the maximum, it can not be heard well. If the stethoscope is designed to fit a small input type, Cursor clipping problems may occur.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and it is an object of the present invention to provide a heartbeat sound processing method in a smartphone capable of obtaining a heartbeat sound with little noise regardless of a model. It is another object of the present invention to provide a method for processing heartbeat sounds in a smartphone capable of increasing the accuracy of heart rate measurement irrespective of the model.

In the present invention, the impulse of the heartbeat is detected from the input signal (impulse detecting step), and the impulse detected section is multiplied by a relatively high gain, and the section where the impulse is not detected is multiplied by a relatively low gain step).

The impulse detection process, when referred avx (n) for an input signal x (n) referred to as the noise floor (floor noise), when avx (n) is less than or equal to x (n) = avx (n) α * avx (n) + (1- α ) * x (n) to calculate and, if avx (n) is greater than x (n) avx (n) = β * avx (n) + (1- β) * x (n), where 0.995 < α <1, 0 < β <0.5. Also, savx (n) = a savx (n) β * savx ( n) + (1- β) * x (n) savx (n) when d> * ν satisfies avx (n) (ν> 1 ) It is determined that the lower impulse is detected.

When the input signal is amplified, the input signal may become too large depending on the type of the signal, so clipping may occur. To prevent this, the input signal may be compressed before the impulse detection step (first signal compression step). In the first signal compression step, if the input signal is z, the maximum value of the input is E1, the first knee point is THR_E1, the second bending value is F11_E, and the bending slope is SCALE_E, then the compressed output value z '

If z < = THR_E1 then z '= z;

If z > THR_E1

z '= THR_E1 + (z-THR_E1) / SCALE_E;

If z '> = F11_E1 then z' = F11_E + (z'- F11_E) / SCALE_E;

If z '> = E1 then z' = E1

. &Lt; / RTI >

Meanwhile, according to the embodiment, the signal output in the step of emphasizing the heartbeat can be smoothed (smoothing processing step). In the smoothing processing step, if the gain is g (n) and the smoothing gain is avg (n)

avg (n) = μ * avg (n) + (1-μ) * g

Can be performed by multiplying the signal output from the step of emphasizing the heartbeat by avg (n) obtained by the following equation.

The output signal may be too large depending on the type of the signal output from the smoothing processing step so that it can be compressed again (second signal compression step). The second signal compression step comprises:

If the input signal is y, the maximum value of the input is Y1, the first knee point is THR_Y1, the second folding value is F11_Y, and the folding slope is SCALE_Y, then the compressed output value y '

If y < = THR_Y1, y '= y;

y > THR_Y1

y '= THR_Y1 + (y - THR_Y1) / SCALE_Y;

If y '> = F11_Y1, y' = F11_Y + (y'-F11_Y) / SCALE_Y;

If y '> = Y1, y' = Y1

. &Lt; / RTI >

In the step of emphasizing the heartbeat, if the gain determined experimentally in the section other than the impulse is G1, the gain G_av to be actually multiplied to the input signal is expressed by

From the non-impulse interval,

G_av =? * G_av + (1 -?) * G1

In the impulse section,

G_av =? * G_av + (1 -?);

Where 0 < G1 < 1, 0 < <1, 0 <η <1, ζ> η.

In addition, the heartbeat negative per minute can be calculated from the number of impulses per minute detected in the impulse detection step, and the mean value of the heartbeat negative per minute measured and the heartbeat negative per minute measured now can be displayed as a bar graph.

According to the present invention, since the input signal is signal-compressed, the possibility of clipping is low even when the input signal is large depending on the type of the apparatus. Since the signal compression is performed by varying the slope according to the input level, no problem occurs even when the input signal is small . In addition, by detecting the impulse while reducing the influence of noise, more accurate heart rate measurement can be performed. Since the impulse-less section is less amplified than the impulse section, a low-pitched heartbeat can be obtained.

FIG. 1 is a system configuration diagram for listening to a heartbeat sound using a smartphone equipped with a program implementing the method of the present invention.
2 is a flowchart of a program for implementing the method of the present invention.
FIG. 3 is an example of a screen for displaying the average value of the heartbeat negative per minute measured and the heartbeat negative per minute measured by the bar graph.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a system configuration diagram for listening to a heartbeat sound using a smartphone equipped with a program implementing the method of the present invention.

The user connects the stethoscope 200 to the earphone terminal of the smartphone 100 after executing the application of the method of the present invention on the smartphone 100 and then the stethoscope 200 is placed on the patient's chest, A heartbeat sound is input to the smartphone 100. The stethoscope 200 can be connected to the earphone terminal of a smartphone using, for example, an earset of Patent Registration No. 10-1557793. Or, for example, a stethoscope connected to a dedicated terminal of a smart phone or a stethoscope connected to a microphone terminal of a smartphone as illustrated in Patent Registration No. 10-1300676 may be used as shown in the registration utility model 20-0384551. The present invention is not limited to a specific type of stethoscope or a specific type of heart sound input method.

The heartbeat sound input to the smartphone 100 may be processed according to the method of the present invention and may be heard through the internal or external speaker 300 or the earphone or may be stored in the smartphone 100. In addition, according to the embodiment, the heartbeat sound data may be transmitted to the server 400 via the Internet and stored or inquired by the doctor in charge.

Next, the heartbeat sound processing method of the present invention will be described in detail with reference to FIG. 2 is a flowchart of a program for implementing the method of the present invention.

In the present invention, an impulse of a heartbeat is detected from an input signal (impulse detection step S200), and a relatively high gain is multiplied to a section in which an impulse is detected, and a relatively low gain is multiplied in a section in which an impulse is not detected (Step S300 of emphasizing the heartbeat).

In the impulse detecting step S200, when the input signal is x (n) and the floor noise is avx (n), when the floor noise avx (n) is smaller than or equal to the input signal x the noise floor avx (n) = α * avx (n) + (1- α) * x (n) to the calculation, and when the ground noise avx (n) is greater than the input signal x (n) avx the noise floor ( n) = ? * avx (n) + (1- ? ) * x (n). At this time, 0.995 < α <1, 0 < β <0.5.

Input signal x (n) = Heart rate is negative and there are various noise mixed in order to reduce the impact of such noise, the savx savx (n) after the determined noise floor (n) β * savx (n ) + (1- β ) * It is recursive as x (n). After finding savx (n), we check if savx (n)> ν * avx (n) ( ν > 1) is satisfied.

In the heartbeat emphasizing step S300, if the gain determined experimentally in the section in which the impulse is not detected is G1, the gain G_av to be actually multiplied by the input signal is

In the section where the impulse is not detected,

G_av =? * G_av + (1 -?) * G1

In the section where the impulse is detected (impulse interval), it is obtained by the following equation.

G_av =? * G_av + (1 -?);

Where 0 < G1 < 1, 0 < <1, 0 <η <1, ζ> η.

The value of G1 is a constant to make the input signal smaller by multiplying the input signal in the non-impulse interval. If it is too small, it becomes uncomfortable to listen. If it is too large, the interval which is not the impulse can not be made small. G1 can be set to a constant value that can not be disturbed by hardware, regardless of hardware.

If the input signal is amplified, the input signal may become too large depending on the type of the signal, so clipping may occur. To prevent this, the input signal may be compressed before the impulse detection step (first signal compression step S100). In the first signal compression step S100, when the input signal is z, the maximum value of the input is E1, the first knee point is THR_E1, the second folding value is F11_E, and the folding slope is SCALE_E, The

If z < = THR_E1 then z '= z;

If z > THR_E1

z '= THR_E1 + (z-THR_E1) / SCALE_E;

If z '> = F11_E1 then z' = F11_E + (z'- F11_E) / SCALE_E;

If z '> = E1 then z' = E1

. &Lt; / RTI >

Here, each variable value is experimentally designed according to the characteristics and range of the input signal. The input signal range of the audio codec is typically -1 to +1 V. If the input signal is greater than the first deflection point (THR_E1) (eg 0.8), then the portion larger than the first deflection point is slowly increased to the slope SCALE_E , And if the compressed signal is greater than the second deflection point F11_E1 (ex 0.85), then increase the portion of the second deflection larger than the second deflection point by a little more slowly to the slope SCALE_E (eg 2). If the value is still larger than E1, the signal is limited to E1. With this configuration, it is possible to prevent clipping of the signal to the utmost. This is also applied to the second signal compression step (S400) to be described later. Meanwhile, according to the embodiment, the input signal may be amplified with a predetermined amplification degree before the first signal compression step (S100).

Meanwhile, according to the embodiment, the signal output in the step S300 for emphasizing the heartbeat can be smoothed (step S400 of smoothing). In the smoothing processing step S400, if the gain is g (n) and the smoothing gain is avg (n)

avg (n) = μ * avg (n) + (1-μ) * g

Depending on the type of the signal output in the smoothing processing step S400, the output signal may become too large, so that the signal can be compressed again (second signal compression step S500). The second signal compression step (S400) is performed through the following procedure.

That is, if the input signal is y, the maximum value of the input is Y1, the first knee point is THR_Y1, the second deflection is F11_Y, and the tilting slope is SCALE_Y, then the compressed output value y '

If y < = THR_Y1, y '= y;

y > THR_Y1

y '= THR_Y1 + (y - THR_Y1) / SCALE_Y;

If y '> = F11_Y1, y' = F11_Y + (y'-F11_Y) / SCALE_Y;

If y '> = Y1, y' = Y1

. &Lt; / RTI >

On the other hand, the heartbeat negative per minute can be calculated from the number of impulses per minute detected in the impulse detecting step S200. In addition, after the measured heartbeat negative per minute was stored and the mean value of the heartbeat per minute measured was calculated, the calculated mean value of the heartbeat per minute and the heartbeat negative per minute measured now are shown in a bar graph It is easy to see how the number of beats of the heart compared to the past is different by the display.

Although the present invention has been described with reference to several examples, the present invention is not limited to the specific examples.

100 smartphones,
200 stethoscopes,
300 speakers,
400 server.

Claims

An impulse detecting step of detecting an impulse of a heartbeat from an input signal,
A step of multiplying a relatively high gain in a section where the impulse is detected and a step of multiplying a relatively low gain in a section in which the impulse is not detected
And a heartbeat sound processing method.

2. The method of claim 1, wherein the impulse detection step
Assuming that the input signal is x (n) and the floor noise is avx (n)
When avx (n) is less than the x (n) or equal to avx (n) = α * avx (n) + (1- α) * x (n),
When avx (n) is greater than x (n) is calculated by avx (n) = β * avx (n) + (1- β) * x (n),
0.995 < alpha < 1, 0 < beta < 0.5,
savx (n) the savx (n) = β * savx (n) + when La (1- β) * x (n )
savx (n)> ν * avx (n) (ν> 1) when it satisfies cardiac sound processing method characterized in that it is determined that the impulse is detected.

3. The method of claim 2,
And a first signal compression step of compressing an input signal before the impulse detection step.

4. The method of claim 3, wherein the first signal compression step comprises:
If the input signal is z, the maximum value of the input is E1, the first knee point is THR_E1, the second bending value is F11_E, and the bending slope is SCALE_E, then the compressed output value z '
If z < = THR_E1 then z '= z;
If z > THR_E1
z '= THR_E1 + (z-THR_E1) / SCALE_E;
If z '> = F11_E1 then z' = F11_E + (z'- F11_E) / SCALE_E;
If z '> = E1 then z' = E1
Wherein the heartbeat sound is obtained by the following equation.

3. The method of claim 2,
And a smoothing processing step of smoothing the signal outputted in the step of emphasizing the heartbeat sound.

6. The method of claim 5,
If the gain is g (n) and the smoothing gain is avg (n)
avg (n) = μ * avg (n) + (1-μ) * g
And multiplying the signal output from the heartbeat negative emphasis step by the avg (n) obtained by the step (b).

6. The method of claim 5,
And a second signal compression step of compressing the signal output from the smoothing processing step.

8. The method of claim 7, wherein the second signal compression step comprises:
If the input signal is y, the maximum value of the input is Y1, the first knee point is THR_Y1, the second folding value is F11_Y, and the folding slope is SCALE_Y, then the compressed output value y '
If y < = THR_Y1, y '= y;
y > THR_Y1
y '= THR_Y1 + (y - THR_Y1) / SCALE_Y;
If y '> = F11_Y1, y' = F11_Y + (y'-F11_Y) / SCALE_Y;
If y '> = Y1, y' = Y1
Wherein the heartbeat sound is obtained by the following equation.

9. The method according to any one of claims 1 to 8,
Let G1 be the experimentally determined gain in the non-impulse interval, and the gain G_av, which is actually multiplied by the input signal,
At non-impulse intervals
G_av =? * G_av + (1 -?) * G1;
In the impulse section
G_av =? * G_av + (1 -?);
Where 0 < G1 < 1, 0 < <1, 0 <η <1, ζ> η
Wherein the heartbeat sound is obtained by the following equation.

10. The method of claim 9,
And calculating a heartbeat negative per minute from the number of impuls per minute detected in the impulse detecting step.

11. The method of claim 10,
And displaying the mean value of the heartbeat negative per minute measured in the past and the heartbeat negative per minute measured now in a bar graph.