KR102640218B1 - A method and an apparatus for estimating respiratory rate through rPPG signals - Google Patents

Abstract

A method of estimating a respiratory rate through a non-contact photoplethysmographic signal according to one aspect includes extracting a luminance component from a skin image; removing noise from the luminance component; acquiring first data by applying a filter to the luminance component from which the noise has been removed; and converting the first data into a first frequency and estimating the breathing rate using the frequency with the highest power among the first frequencies.

Description

{A method and an apparatus for estimating respiratory rate through rPPG signals}

The present invention relates to a method and device for estimating respiratory rate through a non-contact photoplethysmographic signal.

Recently, with the development of imaging technology, it has become possible to obtain high frame rate images, and a technology has been developed to extract non-contact photoplethysmographic (rPPG) signals to estimate various biological signals of people in the image. At this time, the biosignals that can be estimated include respiratory rate.

Breathing is one of the four vital signs, which are important signals indicating the body's condition and function, and is deeply related to heart rate variability (HRV). Respiratory signal measurement can not only be used for various treatments by determining health status, but can also be presented as a meaningful physiological variable to confirm response.
Prior art literature: Domestic Patent Publication No. 10-2358325 (2022.01.27)

The technical problem to be achieved by the present invention is to provide a method for estimating the respiratory rate after extracting a non-contact photoplethysmographic signal from a skin image captured through a remote camera without any additional equipment, and a device for implementing the method. .

In order to solve the above-described problem of the present invention, a method of estimating a respiratory rate through a non-contact photoplethysmographic signal according to an embodiment of the present invention includes the steps of extracting a luminance component from a skin image; removing noise from the luminance component; acquiring first data by applying a filter to the luminance component from which the noise has been removed; and converting the first data into a first frequency and estimating the breathing rate using the frequency with the highest power among the first frequencies.

According to another aspect, an apparatus for estimating a respiratory rate through a non-contact photoplethysmographic signal includes at least one memory; and at least one processor, wherein the at least one processor extracts a luminance component from a skin image, removes noise from the luminance component, and applies a filter to the luminance component from which the noise has been removed to generate first data. Obtains, converts the first data into a first frequency, and controls a device that estimates the respiratory rate using the frequency with the highest power among the first frequencies.

On the other hand, another computer-readable recording medium includes a recording medium on which a program for executing the above-described method on a computer is recorded.

According to one embodiment of the present invention, accuracy can be improved through the present invention, which estimates the respiratory rate through changes in brightness due to body movement when breathing, rather than the conventional technology of estimating the respiratory rate through color changes due to blood flow. There is a possible effect.

In addition, data can be obtained through facial images taken with a camera that is easy to install at home, work, or medical facilities without the need to attach a separate device to the body, such as a respiration measurement sensor (chest belt type, etc.).

In addition, it has the effect of being able to determine the health status of the body by estimating the user's breathing rate.

Additionally, non-contact photoplethysmography (rPPG) signals can be used to extract various biological signals.

1 is an exemplary diagram illustrating an example of a device for estimating respiratory rate through a non-contact photoplethysmographic signal.
Figure 2 is an exemplary diagram illustrating a method of estimating a respiratory rate through a non-contact photoplethysmographic signal.
Figure 3a shows a non-contact photoplethysmographic signal with an adjusted time window and a non-contact photoplethysmographic signal with a filter applied.
FIG. 3B is a power distribution diagram for each frequency band of the non-contact photoplethysmography signal of FIG. 3A.
Figure 4a is a diagram showing non-contact photoplethysmographic signals extracted from four regions divided using principal component analysis.
Figure 4b is a non-contact photoplethysmographic signal for the first main component and a power distribution diagram for each frequency band.
Figure 4c is a non-contact photoplethysmographic signal for the second main component and a power distribution diagram for each frequency band.
Figure 5a is a non-contact photoplethysmographic signal measured for a predetermined time and a power distribution chart for each frequency band.
Figure 5b is a respiration signal measured using a contact respiration measurement device and a power distribution chart for each frequency band.

The advantages and features of the present invention and methods for achieving them will become clear by referring to the embodiments described in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments presented below, but may be implemented in various different forms, and should be understood to include all conversions, equivalents, and substitutes included in the spirit and technical scope of the present invention. . The embodiments presented below are provided to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention. In describing the present invention, if it is determined that a detailed description of related known technologies may obscure the gist of the present invention, the detailed description will be omitted.

The terms used in this application are only used to describe specific embodiments and are not intended to limit the invention. Phrases such as “in some embodiments” or “in one embodiment” that appear in various places in this specification do not necessarily all refer to the same embodiment. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, terms such as “comprise” or “have” are intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, but are not intended to indicate the presence of one or more other features. It should be understood that this does not exclude in advance the possibility of the existence or addition of elements, numbers, steps, operations, components, parts, or combinations thereof.

Additionally, connection lines or connection members between components shown in the drawings merely exemplify functional connections and/or physical or circuit connections. In an actual device, connections between components may be represented by various replaceable or additional functional connections, physical connections, or circuit connections.

Since these embodiments can be subject to various changes and have various forms, some embodiments will be illustrated in the drawings and described in detail. However, this is not intended to limit the present embodiments to a specific disclosure form, and should be understood to include all changes, equivalents, and substitutes included in the spirit and technical scope of the present embodiments. The terms used in this specification are merely used to describe the embodiments and are not intended to limit the embodiments.

Unless otherwise defined, the terms used in the present embodiments have the same meaning as generally understood by those skilled in the art to which the present embodiments belong. Terms such as those defined in commonly used dictionaries should be interpreted as having meanings consistent with the meanings they have in the context of the related technology, and unless clearly defined in the present embodiments, they should not be used in an ideal or excessively formal sense. It should not be interpreted.

The detailed description of the present invention described below refers to the accompanying drawings, which show by way of example specific embodiments in which the present invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that the various embodiments of the present invention are different from one another but are not necessarily mutually exclusive. For example, specific shapes, structures and characteristics described herein may be implemented with changes from one embodiment to another without departing from the spirit and scope of the invention. Additionally, it should be understood that the location or arrangement of individual components within each embodiment may be changed without departing from the spirit and scope of the present invention. Accordingly, the detailed description described below is not to be taken in a limiting sense, and the scope of the present invention should be taken to encompass the scope claimed by the claims and all equivalents thereof. Like reference numbers in the drawings indicate identical or similar elements throughout various aspects.

Hereinafter, various embodiments of the present invention will be described in detail with reference to the attached drawings in order to enable those skilled in the art to easily practice the present invention.

1 is an exemplary diagram illustrating an example of a device for estimating respiratory rate through a non-contact photoplethysmographic signal.

Referring to FIG. 1 , a device 100 (hereinafter referred to as “device”) that estimates respiratory rate through a non-contact photoplethysmographic signal may include a memory 110 and a processor 120.

The device 100 shown in FIG. 1 shows only components related to the present embodiments, and it is obvious to those skilled in the art that other general-purpose elements may be included in addition to the components shown in FIG. 1. do.

Accordingly, the device 100 is a mobile device including a laptop PC, desktop PC, laptop, tablet computer, and smart phone, and a server device. It can be implemented with various types of devices such as devices and embedded devices. Specific examples include smartphones, tablet devices, AR (Augmented Reality) devices, IoT (Internet of Things) devices, self-driving cars, robotics, medical devices, etc. that perform voice recognition, image recognition, and image classification using artificial intelligence. It may be, but is not limited to this. Furthermore, the device 100 may correspond to a dedicated hardware accelerator (HW accelerator) mounted on the above device, and the device 100 may include a neural processing unit (NPU) and a tensor processing unit (TPU), which are dedicated modules for driving artificial intelligence. It may be a hardware accelerator such as Unit), Neural Engine, etc., but is not limited thereto.

Additionally, device 100 may further include a user interface. The user interface may refer to a means for inputting data to control the device 100. For example, the user interface includes a key pad, dome switch, and touch pad (contact capacitive type, pressure resistive type, infrared detection type, surface ultrasonic conduction type, and integral tension measurement type). , piezo effect method, etc.), jog wheels, jog switches, etc., but are not limited to these.

The memory 110 is hardware that stores various data processed within the device 100, and may include a computer-readable recording medium. For example, memory 110 may store data processed and data to be processed in device 100. Additionally, the memory 110 may store applications, drivers, etc. to be run by the device 100. The memory 110 may include at least one of volatile memory or nonvolatile memory. Volatile memory includes dynamic random access memory (DRAM), static random access memory (SRAM), synchronous dynamic random access memory (SDRAM), phase-change random access memory (PRAM), magnetic random access memory (MRAM), and resistive random access memory (RRAM). access memory), FeRAM (ferroelectric random access memory), etc. Non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), electrically programmable read-only memory (EPROM), and electrically erasable programmable read-only memory (EEPROM). In embodiments, memory 110 may include magnetic memory, CD-ROM, Blu-ray or other optical disk storage, hard disk drive (HDD), solid state drive (SSD), compact flash (CF), or secure digital (SD). , Micro-SD (micro secure digital), Mini-SD (mini secure digital), xD (extreme digital), or Memory Stick. Additionally, the memory 110 may store an operating system and at least one program code (code to be executed by the processor 120 operating with reference to FIGS. 2 to 5B).

Referring to FIGS. 2 to 5B , the processor 120 may serve to control overall functions for executing the device 100 for estimating the respiratory rate through a non-contact photoplethysmographic signal. For example, the processor 110 executes software (e.g., a program) stored in the memory 110 in the device 100, thereby executing at least one other component (e.g., hardware or software) of the electronic device connected to the processor 120. components) can be controlled, and the device 100 can be generally controlled by performing various data processing or operations.

For example, the processor 120 may extract a luminance component from a skin image. More specifically, the processor 120 may extract a luminance component from a skin image captured with an RGB camera or an infrared camera. Additionally, the processor 120 may extract the luminance component by converting the skin image into a YCbCr channel or a YCgCo channel. Additionally, the processor 120 can adjust the length of the skin image by considering the heart rate per minute and extract the luminance component from the length-adjusted skin image.

For example, the processor 120 may remove noise from the luminance component. More specifically, the processor 120 can remove noise caused by movement by applying a principal component analysis method to the luminance component.

For example, the processor 120 may obtain first data by applying a filter to the luminance component from which noise has been removed. More specifically, the first data in a frequency range considering the breathing rate per minute can be obtained by applying BPF (Bandpass filtering) to the luminance component from which noise has been removed.

For example, the processor 120 may convert the first data into a first frequency and estimate the breathing rate using the frequency with the highest power among the first frequencies. More specifically, the processor 120 may apply a Fast Fourier Transform (FFT) to the first data and convert it into frequency. Additionally, the processor 120 may correct the respiratory rate using second data extracted through color components from the skin image. More specifically, the processor 120 converts the second data to the second frequency, and when the frequency of the highest power among the second frequencies matches the frequency of the highest power among the first frequencies, the processor 120 converts the second data to the second frequency. The breathing rate can be estimated using the frequency, and if they do not match, the breathing rate can be estimated using the frequency with the highest power among the second frequencies.

According to one embodiment, the processor 120 includes a central processing unit (CPU), a graphics processing unit (GPU), an application processor (AP), and a neural network provided in the device 100. It may be implemented as a neural processing unit (NPU), an image signal processor (ISP), a sensor hub processor, or a communication processor (CP), but is not limited thereto.

Figure 2 is an exemplary diagram illustrating a method of estimating a respiratory rate through a non-contact photoplethysmographic signal.

Referring to FIG. 2, in step 200, the processor 120 may extract a luminance component from a skin image.

The skin image may be an image containing the user's skin captured using an RGB camera, an infrared camera, a zoom camera, etc. At this time, the image containing the user's skin may include consecutive photos at regular time intervals.

The processor 120 may convert the skin image into another color channel. For example, the processor 120 may convert the color channel of the skin image into various color channels such as YCbCr, YCgCo, etc. At this time, luminance is the brightness of light, and the luminance value can be the Y value in the YCbCr channel or YCgCo channel. The luminance component changes through motion artifacts of the upper body caused by breathing, such as inhalation or exhalation. This is different from estimating the pulse through slight changes in skin color to measure changes in blood flow per unit time flowing through arterial blood vessels. When measuring pulse rate, the luminance component is a noise component that needs to be removed, such as facial movement, changes in lighting, and skin brightness. However, when estimating the respiration rate, the change in brightness due to body movement during breathing must be checked rather than the change in color due to blood flow, so the respiration rate is estimated through the luminance component. Additionally, by extracting the luminance component rather than the color component from the skin image to estimate the breathing rate, problems with noise caused by changes in light and ambiguity of the input signal itself can be solved.

Additionally, the processor 120 may adjust the length of the skin image required for respiration rate estimation to the minimum. More specifically, the processor 120 may adjust the length of the skin image by considering the heart rate per minute. The processor 120 may calculate the measurement time for one breathing cycle using the expected respiratory rate calculated through heart rate. Additionally, the processor 120 may determine the minimum length of the image required for respiration rate estimation through the measurement time of one respiration cycle and the incoming respiration cycle when the frequency of the highest power is reflected.

A detailed description of step 200 will be described later with reference to FIGS. 3A and 3B.

In step 201, the processor 120 may remove noise from the luminance component.

The processor 120 may calculate the average value of the skin pixels of each frame at the cluster center of the luminance component. The processor 120 can reduce noise by enlarging skin color by averaging skin pixels.

Additionally, the processor 120 may apply principal component analysis to the luminance component to remove noise caused by movement. More specifically, the processor 120 may reconstruct the luminance component into a first principal component and a second principal component through principal component analysis.

Principal component analysis (PCA) is a technique that converts multidimensional data into low-dimensional data and compresses it. The principal component is a straight line that captures most of the data distribution, and can refer to a vector containing direction and magnitude. The first principal component (PC1) is a main principal component that explains the maximum variance of the data, and may mean a non-contact photoplethysmographic signal for body movement due to breathing. The second principal component (PC2) may be a vector in a direction that is perpendicular to the first principal component (PC1) and has the greatest variance, and may represent a non-contact photoplethysmographic signal for body movement excluding movement due to breathing.

A detailed description of step 201 will be described later with reference to FIGS. 4A to 4C.

In step 202, the processor 120 may obtain first data by applying a filter to the luminance component from which noise has been removed. That is, the processor 120 may apply bandpass filtering (BPF) to the first principal component to obtain first data, which is a remote photoplethysmogram (rPPG) signal in a frequency range considering the respiratory rate per minute.

Since the luminance component from which the noise has been removed contains information in all frequency bands, a process of extracting only the frequency band that takes into account the breathing rate per minute of an ordinary person is necessary. Hereinafter, for convenience of explanation, the frequency range considering the breathing rate per minute of an ordinary person is assumed to be 0.08Hz to 0.66Hz (5bpm to 40bpm), but is not limited thereto.

BPF (Bandpass filtering) is a band-pass filter, which may refer to a filter that passes only forces within a specific desired frequency band without attenuation and attenuates the remaining frequency forces.

The processor 120 may apply BPF (Bandpass filtering) to the first principal component to obtain first data in the range of 0.08 Hz to 0.66 Hz (5 bpm to 40 bpm), which is a frequency range considering the breathing rate per minute.

In step 203, the processor 120 may convert the first data into a first frequency and estimate the breathing rate using the frequency with the highest power among the first frequencies.

The processor 120 may apply Fast Fourier Transform (FFT) to the first data and convert it into frequency.

Fast Fourier Transform (FFT) is a mathematical technique that transforms a given function in the time domain into the frequency domain. The processor 120 may apply a Fast Fourier Transform (FFT) to the first data to list the spectrum of the first frequency as a time series.

The processor 120 may estimate the breathing rate using the frequency value of the highest power among the frequency regions to which the FFT is applied.

Additionally, the processor 120 may correct the breathing rate estimated in step 203 with second data extracted through color components from the skin image. More specifically, the processor 120 may extract color components by converting the skin image into another color system. In addition, by applying BPF to the extracted color component, second data, which is a non-contact photoplethysmogram (rPPG) signal in a frequency range considering the respiratory rate per minute, can be obtained. Additionally, FFT can be applied to the second data to convert it to a second frequency, and the frequency value of the highest power among the second frequencies can be compared with the frequency value of the highest power among the first frequencies. More specifically, when the frequency value of the highest power among the second frequencies matches the frequency of the highest power among the first frequencies, the processor 120 estimates the breathing rate using the frequency of the highest power among the first frequencies. You can. On the other hand, if the frequency value of the highest power among the second frequencies does not match the frequency of the highest power among the first frequencies, the processor 120 may estimate the breathing rate using the frequency of the highest power among the second frequencies. there is.

A detailed description of step 203 will be described later with reference to FIGS. 5A and 5B.

Figure 3a shows a non-contact photoplethysmographic signal with an adjusted time window and a non-contact photoplethysmographic signal with a filter applied.

Referring to FIG. 3A, the processor 120 may determine the minimum length of the image required to estimate the breathing rate by considering the heart rate. The processor 120 may calculate the expected respiratory rate through heart rate per minute * R. According to existing research, it was confirmed through experiments that the heart rate/respiratory rate value is about 4 in a state of no exercise. This could mean that the average adult's breathing rate to heart rate ratio is about 1:4, meaning that the heart beats four times for each breath. Therefore, the value of R can be 1/4=0.25, but this may vary depending on the situation or experiment.

If you calculate the expected respiratory rate using your heart rate, you can find the measurement time for one breathing cycle. More specifically, the processor 120 may calculate the minimum video time required for respiration rate estimation when there are n incoming respiration cycles when the frequency of the highest power is reflected. In other words, the minimum video time can be 1 breathing cycle measurement time * n.

The processor 120 may adjust the time window of the non-contact photoplethysmographic (rPPG) signal 301 extracted from the skin image to the minimum time of the image required to estimate respiratory rate. According to the graph in drawing number 301, since there are 5 breathing cycles, the minimum time of the image can be 20 seconds depending on the measurement time of 1 breathing cycle * n.

Additionally, the processor 120 may apply a Savitzky-Golay (SG) filter to the non-contact photoplethysmography (rPPG) signal with an adjusted time window (302). The Savitzky-Golay (SG) filter is a polynomial filter or least squares filter, and can refer to a filter that removes noise and smoothes data. The processor 120 can reduce noise while maintaining the peak shape and height by applying a Savitzky-Golay (SG) filter (302) to the non-contact photoplethysmography (rPPG) signal with an adjusted time window.

FIG. 3B is a power distribution diagram for each frequency band of the non-contact photoplethysmography signal of FIG. 3A.

The processor 120 may apply FFT to the non-contact photoplethysmography (rPPG) signal 302 to which lowpass filtering has been applied and convert it into frequency. Referring to FIGS. 3A and 3B, the processor 120 can convert the rPPG signal extracted from a 20-second skin image into a frequency and check the value of the frequency of the highest power. According to Figure 3b, the power is highest at 0.25Hz, so the estimated breathing rate can be 15bpm.

Figure 4a is a diagram showing non-contact photoplethysmographic signals extracted from four regions divided using principal component analysis.

Referring to FIG. 4A, the processor 120 may extract luminance components from the face 401, right cheek 402, left cheek 403, and chin 404 portions of the skin image.

Figure number 410 is a non-contact photoplethysmographic (rPPG) signal extracted for 1 minute for 4 areas including the face 401, right cheek 402, left cheek 403, and chin 404. The processor can apply the PCA technique to the non-contact photoplethysmography signal 410 to reconstruct it into two main components 411. At this time, the first principal component (PC1) may refer to a non-contact photoblood flow signal for body movement due to respiration, and the second main component (PC2) may refer to a non-contact photoblood flow signal for body movement excluding movement due to respiration. It can mean.

Figure 4b is a non-contact photoplethysmographic signal for the first main component and a power distribution diagram for each frequency band.

Referring to FIG. 4B, the processor 120 may express (420) the non-contact photoplethysmographic (rPPG) signal for the first principal component (PC1) of figure number 411 of FIG. 4A as a time series. The first principal component (PC1) is a non-contact photoblood flow signal for body movement due to respiration, and may be in a state in which noise for body movement other than respiration has been removed. Additionally, the processor 120 may apply FFT to the non-contact photoplethysmographic time series signal 420 for the first principal component (PC1) and convert it to frequency (421). Looking at the power distribution for each frequency band in drawing number 421, the highest power is at 0.2Hz, so the processor 120 can estimate the breathing rate at 12bpm.

Figure 4c is a non-contact photoplethysmographic signal for the second main component and a power distribution diagram for each frequency band.

Referring to FIG. 4C, the processor 120 may express (430) the non-contact photoplethysmographic (rPPG) signal for the second principal component (PC2) of figure number 411 of FIG. 4A as a time series. The second principal component (PC2) is a non-contact photoblood flow signal for body movement excluding movement due to breathing, and may refer to the non-contact photoblood flow signal for body movement removed in FIG. 4B. Additionally, the processor 120 may apply FFT to the non-contact photoplethysmographic time series signal 430 for the second principal component (PC2) and convert it to frequency (431).

Figure 5a is a non-contact photoplethysmographic signal measured for a predetermined time and a power distribution chart for each frequency band.

Referring to FIG. 5A, the processor 120 may extract (510) a non-contact photoplethysmographic (rPPG) signal for a predetermined time from the skin image. More specifically, the processor 120 may extract the luminance component from the skin image and obtain a non-contact photoplethysmographic signal with noise removed (510). Additionally, the processor 120 can apply FFT to the non-contact photoplethysmographic signal 510 from which noise has been removed and convert it into a frequency (520), and estimate the respiratory rate using the frequency with the highest power. Looking at the power distribution 520 for each frequency band, the highest power is at 0.15 Hz, so the processor 120 can estimate the breathing rate at 9 bpm.

Figure 5b is a respiration signal measured using a contact respiration measurement device and a power distribution chart for each frequency band.

Referring to Figure 5b, drawing number 530 applies a Savitzky-Golay (SG) filter and a moving average filter to the non-contact photoplethysmography (rPPG) time series signal 510 to reduce noise. This is the removed graph.

Drawing number 540 is a respiration signal measured using a respiration measurement sensor (chest belt type). The respiratory rate per minute measured with a respiration sensor is approximately 9 bpm.

Drawing number 520 is a power distribution diagram obtained by extracting the luminance component from a skin image and converting it to frequency by applying FFT to the non-contact photoplethysmographic signal 510 from which noise has been removed. Since it has the highest power at 0.15Hz, the processor 120 can estimate the breathing rate at 9bpm. Since the respiratory rate per minute measured using a respiration sensor is the same as the respiratory rate per minute estimated using a non-contact photoplethysmographic signal extracted from a skin image, it can be seen that the accuracy of the respiratory rate estimated through the luminance component is high.

The description of the present specification described above is for illustrative purposes, and those skilled in the art will understand that the present invention can be easily modified into other specific forms without changing the technical idea or essential features of the present invention. You will be able to. Therefore, the embodiments described above should be understood in all respects as illustrative and not restrictive. For example, each component described as unitary may be implemented in a distributed manner, and similarly, components described as distributed may also be implemented in a combined form.

Unless there is an explicit order or statement to the contrary regarding the steps constituting the method according to the invention, the steps may be performed in any suitable order. The present invention is not necessarily limited by the order of description of the above steps. The use of any examples or illustrative terms (e.g., etc.) in the present invention is merely to describe the present invention in detail, and unless limited by the claims, the scope of the present invention is limited by the examples or illustrative terms. It is not limited. Additionally, those skilled in the art will recognize that various modifications, combinations and changes may be made according to design conditions and factors within the scope of the appended claims or their equivalents.

Therefore, the spirit of the present invention should not be limited to the above-described embodiments, and the scope of the patent claims described below as well as all scopes equivalent to or equivalently changed from the scope of the claims are within the scope of the spirit of the present invention. It will be said to belong to

Claims (10)

In a method of estimating respiratory rate through a non-contact photoplethysmographic signal in a device,
Extracting a luminance component from a skin image;
removing noise from the luminance component;
acquiring first data by applying a filter to the luminance component from which the noise has been removed;
Converting the first data to a first frequency and estimating a respiratory rate using the frequency with the highest power among the first frequencies; and
Comprising: correcting the breathing rate with second data extracted through color components from the skin image,
The correction step is,
If the second data is converted to a second frequency and the frequency of the highest power among the second frequencies matches the frequency of the highest power among the first frequencies, breathing is performed using the frequency of the highest power among the first frequencies. estimating the number, and if they do not match, estimating the breathing rate using the frequency with the highest power among the second frequencies;
A method of estimating respiratory rate through a non-contact photoplethysmographic signal, including. According to paragraph 1,
The extracting step is,
Extracting luminance components from skin images captured with an RGB camera or infrared camera;
Method, including. According to paragraph 1,
The extracting step is,
converting the skin image into a YCbCr channel or YCgCo channel to extract a luminance component;
Method, including. According to paragraph 1,
The extracting step is,
adjusting the length of the skin image in consideration of heart beats per minute; and
extracting a luminance component from the length-adjusted skin image;
Method, including. According to paragraph 1,
The removing step is,
Step of removing noise caused by movement by applying principal component analysis to the luminance component:
Method, including. According to paragraph 1,
The obtaining step is,
acquiring the first data in a frequency range considering the breathing rate per minute by applying BPF (Bandpass filtering) to the luminance component from which the noise has been removed;
Method, including. According to paragraph 1,
The estimation step is,
converting the first data into frequencies by applying FFT (Fast Fourier Transform) to the first data;
Method, including. delete at least one memory; and
At least one processor;
The at least one processor,
Extract the luminance component from the skin image,
Remove noise from the luminance component,
Obtaining first data by applying a filter to the luminance component from which the noise has been removed,
Converting the first data to a first frequency and estimating the respiratory rate using the frequency with the highest power among the first frequencies,
Correcting the respiratory rate with second data extracted through color components from the skin image,
The at least one processor,
If the second data is converted to a second frequency and the frequency of the highest power among the second frequencies matches the frequency of the highest power among the first frequencies, breathing is performed using the frequency of the highest power among the first frequencies. A device for estimating the respiratory rate through non-contact photoplethysmography, which estimates the number and, if they do not match, estimates the respiratory rate using the frequency with the highest power among the second frequencies. A computer-readable recording medium recording a program for executing the method according to claim 1 on a computer.