Deep learning-based blood pressure estimation system using a PPG signal sensing ring

Abstract

A deep learning-based blood pressure estimation system using a photoplethysmogram (PPG) signal sensing ring is provided. The system includes a server, the server including a signal feature extraction component configured to extract PPG features from a PPG signal using a first deep learning model, a user feature extraction component configured to extract user features using a second deep learning model from a first test PPG signal measured using the PPG signal sensing ring, a systolic test blood pressure and a diastolic test blood pressure measured simultaneously with the first test PPG signal using a conventional blood pressure meter, and user information, and a blood pressure estimation component configured to estimate systolic blood pressure and diastolic blood pressure from the PPG feature and the user feature using a third deep learning model, the PPG signal being measured using the PPG signal sensing ring, and the server may receive the PPG signal from the PPG signal sensing ring via a terminal device.

Description

[Technical Field]

[0001] The present invention relates to a blood pressure estimation system, and more particularly to a deep learning-based blood pressure estimation system using a photoplethysmography (PPG) signal sensing ring. [Background technology]

[0002] Blood pressure is generated by the pumping action of the heart and is the pressure that blood exerts on the walls of blood vessels. A commonly used method for measuring blood pressure is the indirect method using Krotkoff sounds. First, a cuff is wrapped around the upper arm and a Korotkoff sound detector is placed. Next, the cuff pressure is increased above the systolic blood pressure, and then gradually reduced until Korotkoff sounds begin to be heard. This pressure is the systolic blood pressure. Further reduction of the cuff pressure causes the Korotkoff sounds to decrease and disappear. This pressure is the diastolic blood pressure. This conventional blood pressure measurement method may cause discomfort to the user due to the compression of the arm. Furthermore, this conventional blood pressure measurement method is unable to measure blood pressure in real time. Summary of the Invention [Problem to be solved by the invention]

[0003] The problem to be solved by the present disclosure is to provide a deep learning-based blood pressure estimation system using a PPG (photoplethysmography) signal sensing ring. [Means for solving the problem]

[0004] In order to solve the above-mentioned problems, according to one embodiment of the present disclosure, there is provided a deep learning-based blood pressure estimation system using a PPG signal sensing ring, including a server, the server including a signal feature extraction component configured to extract PPG features from a PPG signal using a first deep learning model; a user feature extraction component configured to extract user features using a second deep learning model from a first test PPG signal measured using the PPG signal sensing ring, a systolic test blood pressure and a diastolic test blood pressure measured simultaneously with the first test PPG signal using a conventional sphygmomanometer, and user information; and a blood pressure estimation component configured to estimate systolic blood pressure and diastolic blood pressure from the PPG features and the user features using a third deep learning model, wherein the PPG signal is measured using the PPG signal sensing ring, and the server may receive the PPG signal from the PPG signal sensing ring via a terminal.

[0005] In some embodiments, the user information also includes at least one of the user's age, weight, height, and gender.

[0006] In some embodiments, the test PPG signal, the systolic and diastolic test blood pressures, and the user information may be updated periodically.

[0007] In some embodiments, the PPG signal sensing ring includes a plurality of sensors configured to simultaneously measure a plurality of PPG signals at different locations, each of the sensors including a light source and a photoelectric conversion device.

[0008] In some embodiments, the terminal device also includes a light source control component configured to control the light source of each of the plurality of sensors so that a DC (direct current) component of each of the second test PPG signals measured using the plurality of sensors is within a predetermined range.

[0009] In some embodiments, the terminal device further includes a sensor selection component configured to select, among the plurality of sensors, a sensor that measures a second test PPG signal having the highest signal quality among the plurality of second test PPG signals as the sensor for measuring the PPG signal.

[0010] In some embodiments, the signal quality of the plurality of second test PPG signals may be evaluated by at least one of acceleration signal magnitude, signal to noise ratio (SNR), and alternating current (AC) component magnitude to DC component magnitude ratio.

[0011] In some embodiments, the light source control by the light source control component and the sensor selection by the sensor selection component may be performed sequentially, and the light source control by the light source control component and the sensor selection by the sensor selection component may be performed cyclically.

[0012] In some embodiments, the server also includes a signal quality classification component configured to classify the signal quality of the PPG signal as one of good and bad.

[0013] In some embodiments, the server further includes a blood pressure index calculation component configured to calculate a blood pressure index, wherein the blood pressure index may be defined as a ratio of a time during which the signal quality classification component classifies the PPG signal as having good quality and the systolic blood pressure falls outside a first normal range or the diastolic blood pressure falls outside a second normal range to a time during which the signal quality classification component classifies the PPG signal as having good quality. [Effects of the Invention]

[0014] A deep learning-based blood pressure estimation system using a photoplethysmography (PPG) signal sensing ring is provided. According to the present invention, there is no need to compress the arm to measure blood pressure, so users can measure blood pressure conveniently. Furthermore, according to the present invention, blood pressure can be measured in real time. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a three-dimensional view of a PPG signal sensing ring according to one embodiment of the present disclosure. [Figure 2] FIG. 1 is an exploded three-dimensional view of a PPG signal sensing ring according to one embodiment of the present disclosure. [Figure 3] 1 is a block diagram of a deep learning-based blood pressure estimation system using a PPG signal sensing ring according to an embodiment of the present disclosure. [Figure 4A] 1 is a graph showing a PPG signal before preprocessing. [Figure 4B] 1 is a graph showing a PPG signal before preprocessing. [Figure 4C] 4B is a graph showing the PPG signal of FIG. 4A after being pre-processed. [Figure 4D] 4C is a graph showing the PPG signal of FIG. 4B after being preprocessed. [Figure 5A] 10 is a graph showing PPG signals classified as good quality. [Figure 5B] 10 is a graph showing PPG signals classified as good quality. [Figure 5C] 1 is a graph showing PPG signals classified as low quality. [Figure 5D] 1 is a graph showing PPG signals classified as low quality. [Figure 6] FIG. 10 is a block diagram illustrating the signal feature extraction component, the user feature extraction component, and the blood pressure estimation component. [Figure 7] 1 is a flowchart showing a deep learning-based blood pressure estimation method using a PPG signal sensing ring. [Figure 8]10 is a flowchart showing a method for setting a sensor of a PPG signal sensing ring. DETAILED DESCRIPTION OF THE INVENTION

[0016] Figure 1 is a three-dimensional drawing of a photoplethysmography (PPG) signal sensing ring 100 according to one embodiment of the present disclosure. Figure 2 is an exploded three-dimensional drawing of a PPG signal sensing ring 100 according to one embodiment of the present disclosure.

[0017] Referring to FIGS. 1 and 2, the PPG signal sensing ring 100 also includes an outer electrode 130 , an inner electrode 140 , an insulating unit 150 , a top cover 110 , an operation indicating unit 120 , and a plurality of sensors 160 .

[0018] The external electrode 130 may have an arc shape. The external electrode 130 may be made of a conductor and may function as an electrode for measuring an electrocardiogram (ECG). The external electrode 130 also forms the exterior of the PPG signal sensing ring 100, and the external electrode 130 may be in contact with the user's body.

[0019] The internal electrode 140 may have a ring shape and may have multiple openings 145 for multiple sensors 160. The internal electrode 140 may be made of a conductor and may function as an electrode for measuring an electrocardiogram. The internal electrode 140 also forms the interior of the PPG signal sensing ring 100, and the internal electrode 140 may be in contact with a user's finger.

[0020] The insulating unit 150 may be disposed between the outer electrode 130 and the inner electrode 140. The insulating unit 150 may enable electrical insulation between the outer electrode 130 and the inner electrode 140.

[0021] The top cover 110 may have an arc shape and may form a ring shape together with the external electrode 130. The top cover 110 may form the outer appearance of the PPG signal sensing ring 100.

[0022] The operation indicator unit 120 may be coupled to the top cover 110. The operation indicator unit 120 may also include a plurality of LEDs (light-emitting diodes), for example, green and red LEDs. The operation indicator unit 120 may use the plurality of LEDs to indicate the operation of the PPG signal sensing ring 100. For example, the green LED may be turned on for two seconds to indicate the start of measurement, and the green LED may be turned on for one second to indicate the end of measurement. Furthermore, the red LED may flash repeatedly at one-second intervals to indicate a malfunction.

[0023] The sensors 160 may be arranged to contact a user's finger. The sensors 160 may be located within the openings 145 of the internal electrode 140 and protrude from the surface of the internal electrode 140. The sensors 160 may be configured to acquire different PPG signals at different positions. Each sensor 160 may also include a light source and a photoelectric conversion device.

[0024] In some embodiments, although not shown in FIGS. 1 and 2, the PPG signal sensing ring 100 also includes an acceleration sensor disposed between the outer electrode 130 and the inner electrode 140.

[0025] FIG. 3 is a block diagram of a deep learning-based blood pressure estimation system 1000 using a PPG signal sensing ring according to one embodiment of the present disclosure. FIGS. 4A and 4B are graphs showing PPG signals before preprocessing. FIGS. 4C and 4D are graphs showing the PPG signals of FIGS. 4A and 4B after preprocessing, respectively. FIGS. 5A and 5B are graphs showing PPG signals classified as good quality. FIGS. 5C and 5B are graphs showing PPG signals classified as low quality. FIG. 6 is a block diagram showing a signal feature extraction component 330, a user feature extraction component 340, and a blood pressure estimation component 350.

[0026] Referring to Figures 3, 4A to 4D, 5A to 5D, and 6, a deep learning-based blood pressure estimation system 1000 using a PPG signal sensing ring also includes a PPG signal sensing ring 100, a first terminal 200, a server 300, and a second terminal 400.

[0027] The PPG signal sensing ring 100 also includes multiple sensors, for example, a first sensor 160A and a second sensor 160B. Although Fig. 3 illustrates the PPG signal sensing ring 100 as including two sensors, the number of sensors included in the PPG signal sensing ring 100 is not limited to two. In some embodiments, the PPG signal sensing ring 100 also includes an acceleration sensor, although not shown in Fig. 3.

[0028] The first sensor 160A also includes a first light source 161A and a first photoelectric conversion device 164A. The first sensor 160A can sense a first PPG signal using the first light source 161A and the first photoelectric conversion device 164A. The first light source 161A includes, for example, a green LED, a red LED, or an infrared LED, and the first photoelectric conversion device 164A also includes a photodiode.

[0029] The second sensor 160B may also include a second light source 161B and a second photoelectric conversion device 164B. The second sensor 160B may sense a second PPG signal using the second light source 161B and the second photoelectric conversion device 164B. The second light source 161B may also include, for example, a green LED, a red LED, or an infrared LED. The second light source 161B may emit light of the same wavelength as the first light source 161A. The second photoelectric conversion device 164B may also include a photodiode.

[0030] The acceleration sensor may sense the movement of the user by measuring the acceleration of the PPG signal sensing ring 100 .

[0031] The first terminal 200 may be used by a user. The first terminal 200 may include, for example, a smartphone or a tablet PC (personal computer). The first terminal 200 may be connected to the PPG signal sensing ring 100 by wire or wirelessly. For example, the first terminal 200 may be connected to the PPG signal sensing ring 100 using Bluetooth or Wi-Fi. The first terminal 200 may be connected to the server 300 by wire or wirelessly. For example, the first terminal 200 may be connected to the server 300 by Wi-Fi (wireless fidelity) or, for example, 3G (3G). rd generation), LTE (long term evolution), 5G (5 th The server 300 may be connected to the server 300 via a mobile telecommunication technology such as a mobile phone generation (MPU).

[0032] The first terminal 200 also includes a light source control component 210, a sensor selection component 220, a measurement control component 230, and a display component 240. An application may be installed on the first terminal 200. The light source control component 210, the sensor selection component 220, the measurement control component 230, and the display component 240 may be implemented by the application. Alternatively, the first terminal 200 may connect to a website, and the light source control component 210, the sensor selection component 220, the measurement control component 230, and the display component 240 may be implemented on the website.

[0033] The light source control component 210 may control the first light source 161A and the second light source 161B so that the DC components of the two test PPG signals received from the first sensor 160A and the second sensor 160B respectively satisfy a predetermined range.

[0034] After the light source control component 210 controls the first light source 161A and the second light source 161B, the sensor selection component 220 may select one of the first sensor 160A and the second sensor 160B as the sensor for measuring a PPG signal. The sensor selection component 220 may select one of the first sensor 160A and the second sensor 160B that measured the test PPG signal with the highest signal quality among the two test PPG signals from the first sensor 160A and the second sensor 160B. The signal quality may be evaluated by at least one of the acceleration signal magnitude, the signal-to-noise ratio (SNR), and the ratio of the AC component magnitude to the DC component magnitude. The smaller the acceleration signal magnitude, the higher the SNR, and the higher the ratio of the AC component magnitude to the DC component magnitude, the higher the signal quality. The PPG signal may then be measured using the sensor selected by the sensor selection component 220.

[0035] The control of the first light source 161A and the second light source 161B by the light source control component 210 and the selection of a sensor by the sensor selection component 220 may be performed sequentially. That is, the light source control component 210 may control the first light source 161A and the second light source 161B, and then the sensor selection component 220 may select a sensor. The control of the first light source 161A and the second light source 161B by the light source control component 210 and the selection of a sensor by the sensor selection component 220 may be performed periodically. Because the PPG signal sensing ring 100 may move or rotate with respect to the finger, the light source control component 210 may periodically control the light sources, and the sensor selection component 220 may periodically select a sensor.

[0036] The measurement control component 230 may select a measurement mode or start or end a measurement based on user input. For example, a user may use the measurement control component 230 to select a measurement mode between a self-check mode and a background mode. In the self-check mode, the user may use the measurement control component 230 to start and end a measurement. In the background mode, a measurement begins and continues regardless of user input. In the background mode, the user may set or change the measurement period. In some embodiments, the measurement control component 230 may be included in the PPG signal sensing ring 100. That is, a user may use the measurement control component 230 of the PPG signal sensing ring 100 to select a measurement mode or start or end a measurement. In some embodiments, the measurement control component 230 may also be included in the server 300. That is, a service provider may use the measurement control component 230 of the server 300 to select a measurement mode or start or end a measurement.

[0037] The display component 240 may display at least one of the PPG signal measured by the PPG signal sensing ring 100 and stored in the PPG signal storage component 381 of the server 300, the measurement date and time of the PPG signal stored in the PPG signal storage component 381, the signal quality classification result classified by the signal quality classification component 320 of the server 300 and stored in the blood pressure storage component 382 of the server 300, the blood pressure estimated by the blood pressure estimation component 350 of the server 300 and stored in the blood pressure storage component 382 of the server 300, and the blood pressure index calculated by the blood pressure index calculation component 360 of the server 300 and stored in the blood pressure index storage component 383 of the server 300.

[0038] The server 300 also includes a PPG signal pre-processing component 310, a signal quality classification component 320, a signal feature extraction component 330, a user feature extraction component 340, a blood pressure estimation component 350, a PPG signal storage component 381, and a blood pressure storage component 382. In some embodiments, the server 300 also includes a blood pressure index calculation component 360 and a blood pressure index storage component 383. In some embodiments, the server 300 also includes an alarm component 370.

[0039] The PPG signal preprocessing component 310 may preprocess the PPG signal received by the server 300 via the first terminal 200 from a selected sensor of the PPG signal sensing ring 100. For example, a low-pass filter, a high-pass filter, and normalization may be used to preprocess the PPG signal. FIGS. 4A and 4B show the PPG signals before preprocessing, and FIGS. 4C and 4D show the PPG signals of FIGS. 4A and 4B, respectively, after preprocessing. In some embodiments, the PPG signal preprocessing component 310 may preprocess the acceleration signal received by the server 300 via the first terminal 200 from an acceleration sensor of the PPG signal sensing ring 100.

[0040] The signal quality classification component 320 may classify the signal quality of the PPG signal as good or bad. Figures 5A and 5B show PPG signals classified by the signal quality classification component 320 as having good quality. Figures 5C and 5B show PPG signals classified by the signal quality classification component 320 as having poor quality. In some embodiments, the signal quality classification component 320 may reference the acceleration signal to classify the signal quality of the PPG signal as either good or bad.

[0041] A blood pressure estimated from a PPG signal classified as having poor quality is determined to be unreliable and is not displayed by the display component 240 of the first terminal 200. A blood pressure estimated from a PPG signal classified as having good quality is determined to be reliable and may be displayed by the display component 240 of the first terminal 200. In an alternative embodiment, the blood pressure estimated from the PPG signal may be displayed by the display component 240 of the first terminal 200 together with the signal quality classification result, regardless of the signal quality classification result. Furthermore, the blood pressure estimated from the PPG signal may be displayed by the display component 420 of the second terminal 400 together with the signal quality classification result, regardless of the signal quality classification result.

[0042] In some embodiments, the signal quality classification component 320 can use a deep learning model. The deep learning model used in the signal quality classification component 320 can include a convolution neural network (CNN), a long short-term memory (LSTM), a fully connected network (FCN), an encoder, a decoder, or a combination thereof. The method for classifying the signal quality is not limited to the methods described herein.

[0043] As shown in FIG. 6 , the signal feature extraction component 330 may use a first deep learning model to extract PPG features from the PPG signal. The user feature extraction component 340 may use a second deep learning model to extract user features from the test PPG signal, the systolic and diastolic test blood pressures, and user information. The systolic and diastolic test blood pressures may be measured simultaneously with the test PPG signal using a conventional sphygmomanometer. The user information may also include at least one of the user's age, weight, height, and gender. The blood pressure estimation component 350 may use a third deep learning model to estimate the systolic and diastolic blood pressures from the PPG features and the user features. The test PPG signal, the systolic and diastolic test blood pressures, and the user information may be periodically updated.

[0044] The first deep learning model, the second deep learning model, and the third deep learning model may each include a convolutional neural network (CNN), a long short-term memory (LSTM), a fully connected neural network (FCN), an encoder, a decoder, or a combination thereof. The first deep learning model, the second deep learning model, and the third deep learning model may be trained using supervised learning. Data augmentation may be used to ensure a sufficient data set for the supervised learning.

[0045] The blood pressure index calculation component 360 may calculate a blood pressure index. The blood pressure index may be defined based on the time during which the signal quality classification component 320 classifies the PPG signal as having good quality and the time during which the signal quality classification component 320 classifies the PPG signal as having good quality and the systolic blood pressure falls outside a first normal range or the diastolic blood pressure falls outside a second normal range. For example, the blood pressure index may be defined as the ratio of the time during which the signal quality classification component 320 classifies the PPG signal as having good quality and the systolic blood pressure falls outside a first normal range or the diastolic blood pressure falls outside a second normal range to the time during which the signal quality classification component 320 classifies the PPG signal as having good quality. In other words, the blood pressure index may represent the proportion of time during which abnormal blood pressure is indicated among the time during which a reliable blood pressure estimation can be obtained.

[0046] The alarm component 370 can transmit an alarm to at least one of the first terminal 200 and the second terminal 400 if at least one of the blood pressure estimated by the blood pressure estimation component 350 and the blood pressure index calculated by the blood pressure index calculation component 360 satisfies the alarm condition set by the alarm condition setting component 410 of the second terminal 400.

[0047] The PPG signal storage component 381 may store the PPG signal received by the server 300 via the first terminal 200 from a selected sensor of the PPG signal sensing ring 100. The PPG signal storage component 381 may further store the measurement date and time of the PPG signal. The PPG signal storage component 381 may further store the PPG signal preprocessed by the PPG signal preprocessing component 310. The PPG signal storage component 381 may further store the acceleration signal received by the server 300 via the first terminal 200 from the acceleration sensor of the PPG signal sensing ring 100. The PPG signal storage component 381 may further store the acceleration signal preprocessed by the PPG signal preprocessing component 310. The blood pressure storage component 382 may store the blood pressure estimated by the blood pressure estimation component 350. The blood pressure storage component 382 may further store the signal quality classification result classified by the signal quality classification component 320. The blood pressure index storage component 383 may store the blood pressure index calculated by the blood pressure index calculation component 360 .

[0048] The second terminal 400 may be connected to the server 300 via a wired or wireless connection. For example, the second terminal 400 may be connected to the server 300 via Wi-Fi (wireless fidelity) or a mobile communication technology such as 3G, LTE, or 5G. The second terminal 400 may be used by a doctor. The second terminal 400 may include, for example, a smartphone, a tablet PC, a computer, or a laptop.

[0049] The second terminal 400 also includes an alarm condition setting component 410 and a display component 420. The second terminal 400 may be connected to a website or have an application installed thereon. The alarm condition setting component 410 and the display component 420 may be implemented using the website or the application.

[0050] Using the alarm condition setting component 410, a physician can set alarm conditions, such as a systolic blood pressure above 140 mmHg or below 90 mmHg, or a diastolic blood pressure above 90 mmHg or below 60 mmHg, sustained for 10 minutes or more.

[0051] The display component 420 may display at least one of the PPG signal from the PPG signal storage component 381, the PPG signal measurement date and time from the PPG signal storage component 381, the pre-processed PPG signal stored in the PPG signal storage component 381, the signal quality classification result from the blood pressure storage component 382, ​​the blood pressure from the blood pressure storage component 382, ​​and the blood pressure index from the blood pressure index storage component 383.

[0052] FIG. 7 is a flowchart showing a deep learning-based blood pressure estimation method (2000) using a PPG signal sensing ring.

[0053] 3 and 7, the server 300 may receive a PPG signal from a selected sensor of the PPG signal sensing ring 100 via the first terminal 200 (S2050). The PPG signal pre-processing component 310 may pre-process the PPG signal (S2100). For example, a low-pass filter, a high-pass filter, and normalization may be used to pre-process the PPG signal. FIGS. 4A and 4B show the PPG signals before pre-processing, and FIGS. 4C and 4D show the PPG signals of FIGS. 4A and 4B, respectively, after pre-processing. In some embodiments, the PPG signal pre-processing component 310 may further pre-process the acceleration signal received by the server 300 via the first terminal 200 from the acceleration sensor of the PPG signal sensing ring 100.

[0054] Next, the signal quality classification component 320 may classify the signal quality of the PPG signal as good or bad (S2200). The signal quality classification component 320 may refer to acceleration information to classify the signal quality of the PPG signal as good or bad. FIGS. 5A and 5B show PPG signals classified as good quality by the signal quality classification component 320. FIGS. 5C and 5B show PPG signals classified as poor quality by the signal quality classification component 320. A blood pressure estimated from a PPG signal classified as poor quality is determined to be unreliable and is not displayed by the display component 240 of the first terminal 200. A blood pressure estimated from a PPG signal classified as good quality is determined to be reliable and may be displayed by the display component 240 of the first terminal 200.

[0055] In some embodiments, a deep learning model may be used in the signal quality classification component 320. The deep learning model used in the signal quality classification component 320 may include a convolutional neural network (CNN), a long short-term memory (LSTM), a fully connected neural network (FCN), an encoder, a decoder, or a combination thereof. The method for classifying signal quality is not limited to the methods described herein.

[0056] The signal feature extraction component 330 may extract PPG features from the PPG signal using a first deep learning model (S2300). The user feature extraction component 340 may extract user features from the test PPG signal, the systolic and diastolic test blood pressures, and user information using a second deep learning model (S2400). The systolic and diastolic test blood pressures may be measured simultaneously with the test PPG signal using a conventional sphygmomanometer. The user information may also include at least one of the user's age, weight, height, and gender. The blood pressure estimation component 350 may estimate the systolic and diastolic blood pressures from the PPG features and user features using a third deep learning model (S2500). The test PPG signal, the systolic and diastolic test blood pressures, and the user information may be periodically updated.

[0057] The first deep learning model, the second deep learning model, and the third deep learning model may each include a convolutional neural network (CNN), a long short-term memory (LSTM), a fully connected neural network (FCN), an encoder, a decoder, or a combination thereof. The first deep learning model, the second deep learning model, and the third deep learning model may be trained using supervised learning. Data augmentation may be used to ensure a sufficient data set for the supervised learning.

[0058] The PPG signal, the signal quality classification result, and the blood pressure may then be stored (S2570). The PPG signal may be stored in a PPG signal storage component 381. In some embodiments, the pre-processed PPG signal may be further stored in the PPG signal storage component 381. In some embodiments, the acceleration signal may be further stored in the PPG signal storage component 381. In some embodiments, the pre-processed acceleration signal may be further stored in the PPG signal storage component 381. The signal quality classification result and the blood pressure may be stored in a blood pressure storage component 382.

[0059] Next, the steps of receiving a PPG signal (S2050), preprocessing the PPG signal (S2100), classifying the quality of the PPG signal (S2200), extracting PPG features (S2300), extracting user features (S2400), estimating systolic and diastolic blood pressure (S2500), and storing the PPG signal, signal quality classification results, and blood pressure (S2570) may be repeated.

[0060] Next, the blood pressure index calculation component 360 may calculate a blood pressure index (S2600). The blood pressure index may be defined based on the time during which the signal quality classification component 320 classifies the PPG signal as having good quality and the time during which the signal quality classification component 320 classifies the PPG signal as having good quality and the systolic blood pressure falls outside a first normal range or the diastolic blood pressure falls outside a second normal range. For example, the blood pressure index may be defined as a ratio of the time during which the signal quality classification component 320 classifies the PPG signal as having good quality and the systolic blood pressure falls outside a first normal range or the diastolic blood pressure falls outside a second normal range to the time during which the signal quality classification component 320 classifies the PPG signal as having good quality. In other words, the blood pressure index may represent the time during which abnormal blood pressure is indicated among the time during which a reliable blood pressure estimation result can be obtained.

[0061] The blood pressure index may then be stored in the blood pressure index storage component 383 (S2700).

[0062] FIG. 8 is a flow chart illustrating a method (3000) for configuring a sensor in a PPG signal sensing ring.

[0063] 8 and 3, the light source control component 210 may control each light source so that the DC components of each test PPG signal measured using a plurality of sensors satisfy a predetermined range (S3100). This step (S3100) is for setting each sensor, for example, the first sensor 160A and the second sensor 160B, to the best state for measuring the PPG signal.

[0064] Next, the sensor selection component 220 may select one of the plurality of sensors as a sensor for measuring a PPG signal (S3200). The sensor selection component 220 may select the sensor that measured the test PPG signal with the highest signal quality from the two test PPG signals from the first sensor 160A and the second sensor 160B. The signal quality may be evaluated by at least one of the acceleration signal magnitude, the signal-to-noise ratio, and the ratio of the AC component magnitude to the DC component magnitude. The smaller the acceleration signal magnitude, the higher the signal-to-noise ratio, and the higher the ratio of the AC component magnitude to the DC component magnitude, the higher the signal quality. The PPG signal measured by the sensor selected by the sensor selection component 220 may then be transmitted to the server 300 via the first terminal 200.

[0065] The step of controlling the light source of each of the plurality of sensors (S3100) and the step of selecting a sensor (S3200) may be performed sequentially. The method of setting a sensor of a PPG signal sensing ring (3000) may also be performed periodically. As the PPG signal sensing ring 100 may move or rotate relative to the finger, the light source control component 210 may periodically control the light source (S3100), and the sensor selection component 220 may periodically select a sensor (S3200).

[0066] The embodiments disclosed in this disclosure are for illustrative purposes only and are not intended to limit the technical idea of ​​the disclosure, and such embodiments do not limit the scope of the technical idea of ​​the disclosure. The scope of protection of the disclosure should be interpreted by the following claims, and all technical ideas within the equivalent range should be interpreted as being included in the scope of rights of the disclosure.

Claims

1. A deep learning-based blood pressure estimation system using a photoplethysmogram (PPG) signal sensing ring, the blood pressure estimation system including a server, the server comprising: a signal feature extraction component including a first deep learning model configured to input a PPG signal, extract and output features of the PPG signal that can be utilized to estimate systolic and diastolic blood pressure; a user feature extraction component configured to input a first test PPG signal measured using the PPG signal sensing ring, a systolic test blood pressure and a diastolic test blood pressure measured simultaneously with the first test PPG signal using a sphygmomanometer, and user information, and to extract and output user features used to estimate the systolic blood pressure and the diastolic blood pressure, wherein the user information includes at least one of the user's age, weight, height, and sex; a blood pressure estimation component including a third deep learning model configured to input the features of the PPG signal and the features of the user, estimate and output systolic blood pressure and diastolic blood pressure; each of the first deep learning model, the second deep learning model, and the third deep learning model includes a convolutional neural network (CNN), a long short-term memory (LSTM), a fully connected neural network (FCN), an encoder, a decoder, or a combination thereof, and is trained using supervised learning; the first test PPG signal is measured using the PPG signal sensing ring, the PPG signal sensing ring transmits the first test PPG signal to a terminal through communication, and the server receives the first test PPG signal from the terminal through communication; the PPG signal sensing ring includes a plurality of sensors configured to simultaneously measure a plurality of PPG signals at different locations; each of the plurality of sensors includes a light source and a photoelectric conversion device; The terminal a light source control component configured to control the light source of each of the plurality of sensors so that a direct current (DC) component of each of the second test PPG signals measured using the plurality of sensors is within a predetermined range; a sensor selection component configured to select, from the plurality of sensors, a sensor that measured a second test PPG signal having the highest signal quality among the plurality of second test PPG signals as a sensor for measuring the first test PPG signal; Including, A deep learning-based blood pressure estimation system using a PPG signal sensing ring, characterized in that light source control by the light source control component and sensor selection by the sensor selection component are performed periodically and sequentially.

2. 2. The deep learning-based blood pressure estimation system using a PPG signal sensing ring according to claim 1, wherein the first test PPG signal, the systolic test blood pressure, the diastolic test blood pressure, and the user information are periodically updated.

3. 2. The deep learning-based blood pressure estimation system using a PPG signal sensing ring according to claim 1, wherein the signal quality of the plurality of second test PPG signals is evaluated based on at least one of an acceleration signal magnitude, a signal-to-noise ratio (SNR), and an AC component magnitude to DC component magnitude ratio.

4. 2. The deep learning-based blood pressure estimation system using a PPG signal sensing ring according to claim 1, wherein the server further comprises a signal quality classification component configured to classify the signal quality of the first test PPG signal as either good or bad.

5. the server further includes a blood pressure index calculation component configured to calculate a blood pressure index; 5. The deep learning-based blood pressure estimation system using a PPG signal sensing ring according to claim 4, wherein the blood pressure index is defined as a ratio of a time during which the signal quality classification component classifies the quality of the first test PPG signal as good and the systolic blood pressure falls outside a first normal range or the diastolic blood pressure falls outside a second normal range to a time during which the signal quality classification component classifies the quality of the first test PPG signal as good.

Images