Blood pressure estimation apparatus and method utilizing pseudo calibration

Abstract

The present disclosure relates to a blood pressure estimation apparatus utilizing pseudo calibration and a method therefor. The blood pressure estimation apparatus utilizing pseudo calibration includes: a signal measurement device for measuring bio-signals of a user; a blood pressure estimation algorithm for estimating blood pressure using the bio-signals; a pseudo-labeling unit for assigning and labeling with a predetermined blood pressure value in response to the bio-signals obtained from the signal measurement device; a high-quality signal selection unit for selecting a target bio-signal satisfying a predetermined signal quality criterion from among the bio-signals labeled by the pseudo-labeling unit; and a control unit for pseudo calibration of the blood pressure estimation algorithm using the target bio-signal and the blood pressure value with which the target bio-signal is labeled.

Description

Blood pressure estimation device and method using water calibration

[0001] The present invention relates to a blood pressure estimation device and method using pseudo calibration, and more particularly, to a blood pressure estimation device and method using pseudo calibration, which performs pseudo labeling for assigning a blood pressure value to a bio-signal obtained from a user, applies pseudo calibration for calibrating a blood pressure estimation algorithm using the pseudo-labeled bio-signal, and uses pseudo-labeled bio-signals used in the pseudo-calibration, which are bio-signals obtained from the user in different states and are of high quality, thereby improving the accuracy of the blood pressure estimation value.

[0002] Electrocardiogram (ECG) monitoring devices are widely used in the clinical diagnosis of heart disease and can be used to diagnose conditions such as myocardial infarction and pulmonary embolism. In particular, people with or suspected of having heart disease need to be able to detect or receive early warning of sudden cardiac abnormalities.

[0003] Electrocardiogram (ECG) devices attach electrodes to the body to detect the minute electrical signals generated when the heart beats. However, conventional ECG devices suffer from the problem of being uncomfortable for the examiner to wear due to the size of the electrodes.

[0004] Meanwhile, wearable devices are being used in a variety of ways to detect biosignals such as blood pressure. Most conventional wearable devices are worn on the wrist. However, this type of wearable device has limitations, such as low reliability of biosignal data, as the wearable device does not come into contact with any part of the body other than the wearer.

[0005] To address the low reliability of biosignal data in conventional wearable devices, a ring-shaped sensing device is being developed that can improve the reliability of biosignal data. Conventional ring-shaped sensing devices may include an inner ring, an outer ring, and a top cover, and can be manufactured by assembling various components.

[0006] A ring-type sensing device can obtain a PPG signal (Photoplethysmography) from a user, derive blood vessel elasticity information using the characteristic values ​​of the PPG signal, and estimate blood pressure using the blood vessel elasticity information. Estimating blood pressure using the PPG signal in this way allows for simpler and easier blood pressure estimation compared to the conventional method using a cuff blood pressure monitor. However, since the method of estimating blood pressure using the PPG signal using the ring-type sensing device is not a direct measurement method like a cuff, there is a need to improve the accuracy of the measurement, and the present invention proposes an invention that can improve the accuracy of the estimated value by correcting the blood pressure estimation value.

[0007] The present invention is intended to solve the above-described problem, and more specifically, to provide a blood pressure estimation device and method using pseudo calibration, which performs pseudo labeling for assigning a blood pressure value to a bio-signal obtained from a user, and calibrates a blood pressure estimation algorithm using the pseudo-labeled bio-signal, and improves the accuracy of the blood pressure estimation value by using a high-quality signal while the pseudo-labeled bio-signal used in the pseudo calibration is a bio-signal obtained from the user in different states.

[0008] A blood pressure estimation device applying pseudo calibration according to one aspect of the present invention is characterized by including: a signal measuring device for measuring a user's bio-signal; a blood pressure estimation algorithm for estimating blood pressure using the bio-signal; a pseudo-labeling unit for labeling the bio-signal obtained from the signal measuring device by assigning a predetermined blood pressure value corresponding to the bio-signal; a high-quality signal selection unit for selecting a target bio-signal satisfying a predetermined signal quality standard among the bio-signals labeled by the pseudo-labeling unit; and a control unit for pseudo-calibrating the blood pressure estimation algorithm using the target bio-signal and the blood pressure value labeled in the target bio-signal.

[0009] In addition, it is preferable that the signal measuring device include a first sensor unit that collects a user's photoplethysmography signal (PPG signal).

[0010] In addition, it is preferable that the signal measuring device obtains the bio-signals measured in at least two different states of the user, and that the different states are states in which the user's cardiac output or peripheral resistance is in different environments.

[0011] In addition, it is preferable that the signal measuring device include a second sensor unit that detects the user's movement.

[0012] In addition, it is preferable that the second sensor unit is an acceleration sensor that measures the magnitude or amount of change in acceleration that occurs according to the movement pattern of the user's finger.

[0013] In addition, it is preferable to further include a state determination unit that determines that the user is in the first state when a predetermined first pattern is measured for a specified time by the second sensor unit, and determines that the user is in the second state when a predetermined second pattern is measured for a specified time by the second sensor unit.

[0014] In addition, the signal measuring device is preferably configured as a ring to be worn on the user's finger, and the status determination unit preferably determines whether the user is wearing the signal measuring device based on the size or shape of the signal measured by the first sensor unit.

[0015] In addition, it is preferable that the status determination unit determines whether the user is sleeping or active from the bio-signal received from the signal measurement device.

[0016] In addition, it is preferable that the labeling unit assigns and labels blood pressure values ​​to the bio-signals measured in at least two different states of the user.

[0017] In addition, it is preferable that the high-quality signal selection unit selects a target biosignal that satisfies a predetermined signal quality standard for each of the labeled biosignals of the different states, and the control unit calibrates the blood pressure estimation algorithm using the target biosignals reflecting the different states and the blood pressure values ​​labeled in the target biosignals.

[0018] In addition, the signal measuring device includes a first sensor unit that collects a photoplethysmography signal (PPG signal) of the user, and a second sensor unit that detects the movement of the user, and it is preferable that the first sensor unit continuously or periodically collects a biosignal from the user in the different states, and the second sensor unit continuously or periodically measures the acceleration of the user's finger.

[0019]

[0020] Meanwhile, a blood pressure estimation method using pseudo calibration according to another aspect of the present invention is characterized by including a signal measurement step of measuring a user's bio-signal; a blood pressure estimation step of estimating blood pressure using a blood pressure estimation algorithm that estimates blood pressure using the bio-signal; a pseudo-labeling step of labeling by assigning a predetermined blood pressure value corresponding to the bio-signal obtained from the signal measuring device; and a high-quality signal selection step of selecting a target bio-signal that satisfies a predetermined signal quality standard among the bio-signals labeled by the pseudo-labeling step; and a pseudo-calibration step of calibrating the blood pressure estimation algorithm using the target bio-signal and the blood pressure value labeled in the target bio-signal.

[0021] Here, it is preferable that the signal measurement step is performed by a first sensor unit that collects the user's photoplethysmography signal (PPG signal).

[0022] Here, the signal measuring device obtains the bio-signals measured in at least two different states of the user, and it is preferable that the different states are states in which the user's cardiac output or peripheral resistance is in different environments.

[0023] Here, it is preferable that the signal measurement step detects the user's movement by the second sensor unit.

[0024] Here, it is preferable that the second sensor unit is an acceleration sensor that measures the magnitude or amount of change in acceleration that occurs according to the movement pattern of the user's finger.

[0025] Here, it is preferable to further include a state determination step of determining that the user is in the first state when a predetermined first pattern is measured for a specified time by the second sensor unit, and determining that the user is in the second state when a predetermined second pattern is measured for a specified time by the second sensor unit.

[0026] Here, the signal measurement step is performed through a ring worn on the user's finger, and the status determination step preferably determines whether the user is wearing the ring based on the size or shape of the signal measured by the first sensor unit.

[0027] Here, it is preferable that the state determination step determines whether the user is sleeping or active from the bio-signal received from the signal measurement step.

[0028] Here, it is preferable that the labeling unit assigns and labels blood pressure values ​​to the bio-signals measured in at least two different states of the user.

[0029] Here, it is preferable that the high-quality signal selection unit selects a target biosignal that satisfies a predetermined signal quality standard for each of the labeled biosignals of the different states, and the control unit calibrates the blood pressure estimation algorithm using the target biosignals reflecting the different states and the blood pressure values ​​labeled in the target biosignals.

[0030] Here, the signal measurement step preferably collects a photoplethysmography signal (PPG signal) of the user by a first sensor unit, detects the user's movement by a second sensor unit, and the first sensor unit continuously or periodically collects a biosignal from the user in the different states, and the second sensor unit continuously or periodically measures the acceleration of the user's finger.

[0031] The present invention relates to a blood pressure estimation device and method using pseudo calibration, wherein pseudo labeling is performed to assign a blood pressure value to a bio-signal obtained from a user, and pseudo calibration is applied to calibrate a blood pressure estimation algorithm using the pseudo-labeled bio-signal, and pseudo-labeled bio-signals used in the pseudo calibration are bio-signals obtained from the user in different states, and use high-quality signals, thereby providing an effect of improving the accuracy of a blood pressure estimation value.

[0032] In addition, the present invention performs a blood pressure estimation algorithm calibration using a cuff label, thereby simplifying the calibration method of the conventional blood pressure estimation algorithm and improving user convenience. In other words, the inconvenience of performing a blood pressure estimation algorithm using the bio-signal obtained from the signal measuring device and the actual blood pressure value from the cuff sphygmomanometer by the user wearing a cuff signal measuring device and a cuff sphygmomanometer simultaneously, as in the past, can be eliminated.

[0033] Figure 1 is a drawing illustrating a signal measuring device according to an embodiment of the present invention.

[0034] Figure 2 is an exploded perspective view of Figure 1.

[0035] Figure 3 is a block diagram of a blood pressure estimation device according to an embodiment of the present invention.

[0036] Figure 4 is a drawing illustrating a water calibration process according to an embodiment of the present invention.

[0037] Figure 5 is a drawing showing a process in which a state determination unit determines different states using a first sensor unit and a second sensor unit.

[0038] Figure 6 is a diagram showing the process of determining the sleep state by the state determination unit.

[0039] FIG. 7 is a diagram illustrating a process of a blood pressure estimation method using pseudo calibration according to an embodiment of the present invention.

[0040] This specification clarifies the scope of the present invention and explains the principles of the invention and discloses embodiments thereof to enable those skilled in the art to practice the invention. The disclosed embodiments may be implemented in various forms.

[0041] Expressions such as “includes” or “may include” that may be used in various embodiments of the present invention indicate the existence of the disclosed function, operation, or component, etc., and do not limit one or more additional functions, operations, or components, etc. In addition, in various embodiments of the present invention, it should be understood that terms such as “includes” or “has” are intended to specify the existence of a feature, number, step, operation, component, part, or combination thereof described in the specification, and do not exclude in advance the possibility of the existence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0042] When a component is referred to as being "connected, coupled" to another component, it should be understood that while the component may be directly connected or coupled to the other component, there may also be a new component between the component and the other component. Conversely, when a component is referred to as being "directly connected" or "directly coupled" to another component, it should be understood that no new component exists between the component and the other component.

[0043] The terms "first," "second," etc., used herein may be used to describe various components, but the components should not be limited by the terms. The terms are used solely to distinguish one component from another.

[0044]

[0045] The present invention relates to a blood pressure estimation device using pseudo calibration and a blood pressure estimation method using pseudo calibration, and more particularly, to a blood pressure estimation device and method using pseudo calibration that calibrate a blood pressure estimation value by utilizing the user's unique blood pressure change characteristics from bio-signals obtained from the user in different states. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings.

[0046]

[0047] Referring to FIG. 3, a blood pressure estimation device (100) applying pseudo calibration according to one embodiment of the present invention includes a signal measuring device (10), a blood pressure estimation algorithm (20), a pseudo labeling unit (31), a high-quality signal selection unit (31), and a control unit (50).

[0048]

[0049] The signal measuring device (10) is provided to measure the user's bio-signal. According to the present embodiment, the signal measuring device (10) may include a first sensor unit (11) that collects the user's photoplethysmography signal (PPG signal). According to the present embodiment, the signal measuring device (10) uses the PPG signal to estimate a blood pressure value. However, the bio-signal collected by the signal measuring device (10) may be any bio-signal other than the PPG signal as long as it is a bio-signal capable of estimating the blood pressure value. In addition, according to the present embodiment, the signal measuring device (10) may include a second sensor unit (12) for detecting the user's movement. The second sensor unit (12) may be an acceleration sensor that measures the magnitude and amount of change in acceleration that occurs according to the movement pattern of the user's finger.

[0050] According to the present embodiment, the signal measuring device (10) may be provided in a wearable device that can be worn by a user. For example, the signal collection unit (10) may be implemented as a device in the form of a ring, band, or watch that can be worn on a finger, wrist, or ankle, or may be provided by being installed on clothing. That is, it may be implemented in various forms such as a ring type, a wristwatch type, a bracelet type, an anklet type, a wrist band type, an ankle band type, a hair band type, or a glasses type. Hereinafter, the signal measuring device (10) will be described in the case where it is a ring type that can be worn on a user's finger.

[0051] Referring to FIGS. 1 and 2, the signal collection device (10) may include an external electrode (101), an internal electrode (102), an insulation unit (103), a top cover (104), and a plurality of sensors (105).

[0052] The external electrode (101) may have an arc shape with an open section. The external electrode may be composed of a conductor and may function as an electrode for measuring an electrocardiogram (ECG) by coming into contact with the user's body. The external electrode (101) may form the exterior of a ring-shaped signal collection device (10).

[0053] The inner electrode (102) may have a ring shape and may be formed with a plurality of openings (106) to install a plurality of sensors (105). The inner electrode (102) may be formed of a conductor and may function as an electrode for measuring an electrocardiogram. In addition, the inner electrode (102) may form the inner ring of a ring and the inner electrode (102) may come into contact with a user's finger.

[0054] An insulating unit (103) may be placed between the external electrode (101) and the internal electrode (102). The insulating unit (103) may enable electrical insulation between the external electrode (101) and the internal electrode (102).

[0055] The top cover (104) may have an arc shape so as to close a portion of the external electrode (101), and may be formed with the same curvature as the external electrode (101) to form a ring together with the external electrode. The top cover (104) may be coupled to a portion of the portion between the external electrodes (101) to form the exterior of the ring-shaped signal collection device (10).

[0056] A plurality of sensors (105) may be arranged to protrude inwardly from the inner surface of the inner electrode (102) and come into contact with the user's finger. The plurality of sensors (105) may be arranged in each of the openings (106) of the inner electrode (102). The plurality of sensors (105) may be first sensor units (11) that collect PPG signals (photoplethysmograms) from the user at different locations. In addition, the signal collection device (10) may have a second sensor unit (12) arranged between the outer electrode (101) and the inner electrode (102). The second sensor unit (12) may be an acceleration sensor arranged between the outer electrode (101) and the inner electrode (102) to detect the user's movement.

[0057]

[0058] The above blood pressure estimation algorithm (20) is provided to estimate blood pressure using the bio-signals. The above blood pressure estimation algorithm (20) may be generated in a predetermined manner and provided in the form of a predetermined memory or application.

[0059] The above blood pressure estimation algorithm (20) can be algorithmized to extract blood pressure by extracting features from the user's bio-signal. For example, the above blood pressure estimation algorithm (20) can be algorithmized to extract a predetermined characteristic point from the user's PPG signal and use it to estimate blood pressure by comparing it with the measured blood pressure value.

[0060] The above characteristic points can be extracted from specific points in time or periodic signals. The specific points in time or periodic signals can be selected from biosignals measured continuously or intermittently. If the characteristic points are periodic signals, certain representative features can be selectively extracted and utilized from the periodic signals. There can be one or more pieces of information about the characteristic points obtained through such characteristic points in time or periodic signals, and these can be combined and applied to estimate a specific blood pressure value.

[0061] In addition, the blood pressure algorithm (20) can be obtained by retrospectively analyzing the waveform of the PPG signal. The PPG signal can be analyzed by applying various methods such as linear regression analysis, multiple regression analysis, and nonlinear regression analysis. There is a predetermined relationship between the waveform of the PPG signal and the actual blood pressure value, and by analyzing this relationship, an algorithm for estimating the blood pressure value from the PPG signal can be secured.

[0062] In addition, the blood pressure estimation algorithm (20) can be derived by training using a machine learning or deep learning method. For example, an artificial neural network (ANN), a recurrent neural network (RNN), etc. can be used. Various hidden layers can be used to derive a blood pressure estimation value, and various trained learning models can be used to estimate a blood pressure value using the PPG signal.

[0063] Meanwhile, the blood pressure estimation algorithm (20) can be implemented to provide an estimated blood pressure value according to physical characteristics by using the user's basic physical information such as age, gender, weight, and height together with the PPG signal in order to improve the accuracy of estimation.

[0064]

[0065] The above-mentioned water labeling unit (31) labels the biosignal obtained from the signal measuring device (10) by assigning a predetermined blood pressure value corresponding to the biosignal. The predetermined blood pressure value assigned corresponding to the biosignal can be assigned through the blood pressure estimation algorithm (20). The above-mentioned water labeled biosignal can be called water labeling in the sense that it does not assign an accurately measured blood pressure value, for example, a blood pressure value acquired using a cuff blood pressure monitor.

[0066] According to the present embodiment, the signal measuring device (10) obtains bio-signals measured in at least two different states of the user, and the blood pressure labeling unit (31) labels the bio-signals measured in at least two different states of the user by assigning blood pressure values.

[0067] To derive a user's unique blood pressure variation characteristics, biosignals measured under different user conditions can be acquired. The more biosignals collected across multiple different environments and conditions, the more precisely the user's unique blood pressure characteristics can be extracted. Furthermore, the more precisely the unique blood pressure variation characteristics are extracted, the more accurate the blood pressure estimation algorithm (20) that reflects these characteristics will be.

[0068] The above different states refer to cases where the factors affecting the user's blood pressure are different. Generally, blood pressure is calculated as 'cardiac output × peripheral resistance', and cardiac output is defined as 'stroke volume × heart rate'. Therefore, different states of stroke volume, heart rate, or peripheral resistance as factors affecting blood pressure can be treated as different states for extracting blood pressure change characteristics. For example, immediately after waking up, the sympathetic nervous system is activated, cardiac output increases, and blood vessels constrict, increasing resistance. In addition, during the day when people are actively working, cardiac output increases and the degree of blood vessel constriction becomes greater. On the other hand, in a situation where one has sufficiently rested (generally just before going to bed), the parasympathetic nervous system is activated, cardiac output decreases, blood vessels dilate, and resistance decreases.

[0069] Specifically, by using the bio-signals measured from the signal measuring device (10) in the first state and the blood pressure value measured at that point in time, and the bio-signals measured from the signal measuring device (10) in the second state different from the first state and the blood pressure value measured at that point in time, the user's unique state change characteristics in the first state and the second state can be applied to the blood pressure estimation algorithm. For example, if the user's systolic blood pressure measured in the first state is 120 mmHg in the first measurement and 124 mmHg in the second measurement, and the user's systolic blood pressure measured in the second state is 140 mmHg in the first measurement and 138 mmHg in the second measurement, blood pressure change information according to the change in the bio-signals, for example, the PPG signal, from the first state to the second state can be extracted. At this time, the average value of the measurement values ​​in the first state and the average value of the measurement values ​​in the second state can be used to extract the state of the change. This blood pressure change information is unique information of the user and can provide the effect of enabling more accurate blood pressure correction tailored to the individual.

[0070] The above-mentioned blood pressure labeling unit (31) labels the bio-signals acquired in the different states by assigning a predetermined blood pressure value, thereby securing labeled data reflecting the different states. For example, the different states may include distinction between active / sleep states or distinction between systolic / diastolic states.

[0071]

[0072] The above-described superior quality signal selection unit (31) is provided to select a target biosignal that satisfies a predetermined signal quality standard among the biosignals labeled by the water labeling unit (31). The target biosignal selected by the above-described superior quality signal selection unit (31) is a biosignal having a certain level of superior signal quality or higher and is used when performing water calibration. The above-described predetermined signal quality standard may be appropriately selected in consideration of noise level, signal variability, etc.

[0073] At this time, according to the present embodiment, since the signal measurement device (10) obtains bio-signals measured in at least two different states of the user, and the water labeling unit (31) performs labeling on bio-signals reflecting the different states of the user, the high-quality signal selection unit (31) can select a target bio-signal that satisfies a predetermined signal quality standard for each of the labeled bio-signals in different states.

[0074] Specifically, the superior quality signal selection unit (31) may select a first superior quality signal that satisfies a predetermined criterion among the first measurement values ​​measured in the first state, and may select a second superior quality signal that satisfies a predetermined second criterion among the second measurement values ​​measured in the second state as the target biosignal. For example, the first state may be a resting state, and the second state may be an active state.

[0075] In the first state, the first measurement value may be a measurement value that satisfies a predetermined criterion among measurement values ​​measured multiple times in the first state, or an average value of measurement values ​​that satisfies a predetermined criterion may be used. Measurements may be performed at predetermined time intervals, for example, 10 to 30 minutes, in the first state, and a first high-quality signal that satisfies a predetermined first criterion among the measurement values ​​in the first state may be selected as a target biosignal and applied to calibration. The first criterion may be selected as the first high-quality signal when the difference between the SBP (systolic blood pressure) and DBP (diastolic blood pressure) values ​​acquired through the PPG signal is within a predetermined error range that is set in advance.

[0076] The second measurement value in the second state, similar to the second measurement value in the first state, may be a measurement value that satisfies a predetermined criterion among measurement values ​​measured multiple times in the second state, or an average value of measurement values ​​that satisfies a predetermined criterion may be used. Measurements may be performed at predetermined time intervals, for example, 10 to 30 minutes, in the second state, and a second high-quality signal that satisfies a predetermined second criterion among the measurement values ​​in the second state may be applied to calibration. The second criterion may be selected as the second high-quality signal when the difference between the SBP (systolic blood pressure) and DBP (diastolic blood pressure) values ​​acquired through the PPG signal is within a predetermined error range that is set in advance.

[0077] In this way, among the measurement values ​​measured in different states, a signal of excellent quality that satisfies a predetermined standard can be selected as a target biosignal, and a predetermined blood pressure value assigned to the target biosignal can be used to calibrate the blood pressure estimation algorithm.

[0078]

[0079] Furthermore, the high-quality signal selection unit (31) can extract and combine feature points for the bio-signals obtained from the signal measurement device (10) to extract the user's unique features for cardiac output, heart rate, and vascular resistance. The noise of the bio-signals can be removed and preprocessed, and the feature points can be extracted from the bio-signals of the periodic waveform. The high-quality signal selection unit (31) selects the first high-quality signal and the second high-quality signal as target bio-signals from the noise-removed signals, selects target bio-signals in which unique blood pressure change features are reflected in different states, and applies a predetermined blood pressure value assigned to the target bio-signals to the calibration of the blood pressure estimation algorithm (20).

[0080]

[0081] According to an embodiment of the present invention, a blood pressure estimation device using water calibration includes a status determination unit (40). The status determination unit (40) is provided to receive a bio-signal measured from the second sensor unit (12) and determine the user's status.

[0082] As illustrated in FIG. 5, the state determination unit (40) can determine that the user is in the first state when a predetermined first pattern is measured for a specified time by the second sensor unit (12), and can determine that the user is in the second state when a predetermined second pattern is measured for a specified time by the second sensor unit (12). At this time, the first state may be a resting state, and the second state may be an active state. Alternatively, when a predetermined third pattern is measured by the second sensor unit, the user may be determined to be in a third state, for example, a sleeping state.

[0083] In addition, the state determination unit (40) can determine whether the user is wearing the signal measurement device (10) based on the size or shape of the signal measured by the second sensor unit (12). Furthermore, the state determination unit (40) can determine whether the user is sleeping or active based on the bio-signal. The sleep or activity status can be considered as different environments corresponding to the different conditions of the states.

[0084] According to the present embodiment, the second sensor unit (12) may be an acceleration sensor that detects the movement of a finger, and the second sensor unit (12) may continuously or periodically measure the acceleration of the movement of the user's finger. The status determination unit (40) may determine the user's status using the ACC signal information measured by the acceleration sensor.

[0085] The above-described status determination unit (40) may be provided in the signal measurement device (10) or may be provided in various electronic devices such as a mobile device, a wearable device, a PC, etc. that can receive a bio-signal from the signal measurement device (10). In addition, the blood pressure estimation algorithm (20) may also be provided in the signal measurement device (10) or may be provided in various electronic devices such as a mobile device, a wearable device, a PC, etc. that can receive a bio-signal from the signal measurement device (10).

[0086] Specifically, when the state determination unit (40) determines that the user is wearing the ring (10) through the first sensor unit (11), and a first pattern is measured from the second sensor unit (12) for a specified period of time, the state can be determined as a first state in which the user is resting for a specified period of time. Here, the first pattern may be a regular pattern, and the first pattern may be a stable pattern. Specifically, the first pattern may be a regular pattern in which a specific pattern is repeated. That is, when the state determination unit (40) determines that the user is wearing the ring (10) through the first sensor unit (11), and a stable first pattern is measured from the second sensor unit (12), the state determination unit (40) determines that the user is resting. When the state determination unit (40) determines that the user is resting, the user can be set to be in the first state.

[0087] In addition, the state determination unit (40) determines that the user has worn the ring (10) for a specified period of time through the first sensor unit (11), and when a second pattern is measured from the second sensor unit (12), it can determine that the user is in a second state in which the user is active for a specified period of time. At this time, the second pattern may be an irregular pattern, and the second pattern may be an unstable pattern. The state determination unit (40) determines that the user is active when the second sensor unit (12) measures the unstable second pattern when it determines that the user has worn the ring-type signal measurement device (10) through the first sensor unit (11). When the state determination unit (40) determines that the user is active, the user can be set to be in the second state.

[0088] In the embodiments of the present invention, the first pattern is described as being a regular pattern and the second pattern is described as being an irregular pattern, but this is not limited thereto. The first pattern may be composed of various patterns as long as it can measure whether the user is resting, and the second pattern may be composed of various patterns as long as it can measure whether the user is active.

[0089] Additionally, the first pattern capable of measuring whether the user is resting and the second pattern capable of measuring whether the user is active may be derived from patterns occurring in the user's previous activity and previous rest data.

[0090] According to an embodiment of the present invention, the state determination unit (40) can determine whether the user is sleeping based on the number of times the first state is determined to be resting. Specifically, as illustrated in FIG. 6, the state determination unit (40) can determine that the user is sleeping when the number of times the first state is determined to be resting exceeds a predetermined reference number for a certain period of time. The sleeping state can be defined as the third state.

[0091] According to an embodiment of the present invention, a case where a user transitions from a resting state to an active state cannot be determined as a sleep state, but a case where the user continues to remain in a resting state can be determined as a sleep state. For example, when a user transitions from a resting state to an active state, the first sensor unit (11) may measure a stable first pattern and then measure an unstable second pattern. In this case, the state determination unit (40) may determine that the user has transitioned to an active state that is not sleeping or resting.

[0092] When the user continues to remain in a resting state and enters a sleeping state, the first sensor unit (11) can continuously measure the first stable pattern. In this case, the state determination unit (40) can determine that the user is in a sleeping state.

[0093] At this time, the state determination unit (40) determines that the user is sleeping when the number of times the third state is determined to be resting for a certain period of time is greater than or equal to a specified standard number. More specifically, the state determination unit (40) determines that the user is sleeping when the number of times the third state is determined to be resting for a certain period of time (a specified period of time) is greater than or equal to a specified standard value, thereby improving the accuracy of determining whether or not the user is sleeping. That is, the state determination unit (40) can determine whether or not the third state is sleeping through the number of repetitions of a short unit of time determined to be resting for a certain period of time.

[0094] According to an embodiment of the present invention, the state determination unit (40) can calculate the reference number based on statistical values ​​of the first state measured over a specified period of time. Specifically, the state determination unit (40) can calculate the reference number of the first state based on the user's previous sleep data to determine whether or not the user is asleep.

[0095] According to an embodiment of the present invention, the second sensor unit (12) can measure the movement patterns of the user's fingers that occur when the user is in a sleeping state and when the user is in a state where the user is repeating activity and rest. The state determination unit (40) can apply statistical values ​​(e.g., average values) of the measurement values ​​measured by the second sensor unit (12) when in a sleeping state and the measurement values ​​when the user is in a state where the user is repeating activity and rest when determining the state.

[0096] In addition, the state determination unit (40) according to an embodiment of the present invention can calculate statistical values ​​by accumulating signals measured by the second sensor unit (12) over a specified period of time. Specifically, the state determination unit (40) can calculate statistical values ​​in each state, such as when the user is in a sleeping state and when the user is in a state of alternating activity and rest.

[0097] For example, the state determination unit (40) may set a designated period of 24 hours prior to the sleep determination time, and accumulate signals measured by the second sensor unit (12) to calculate statistical values. In addition, the state determination unit (40) may set a designated period of 24 hours prior to the sleep determination time, and calculate statistical values ​​according to each state, whether the user is in a sleeping state or in a state of repeating activity and rest. The number of criteria required to determine whether or not sleep is present may vary depending on the user's lifestyle. Therefore, in order to increase the accuracy of sleep determination, it is necessary to optimize the number of criteria required to determine sleep for each user. By adjusting the number of criteria for each user, it is possible to determine whether or not sleep is present in an optimized manner for the user.

[0098] According to an embodiment of the present invention, the determination of rest, activity, and sleep can be optimized for each user, and the blood pressure estimation algorithm (20) can be calibrated using a target biosignal reflecting the characteristics of each state for each user and a predetermined blood pressure value assigned thereto, thereby further improving the accuracy of the estimated blood pressure value estimated by the blood pressure estimation algorithm (20).

[0099]

[0100] The control unit (50) performs pseudo-calibration of the blood pressure estimation algorithm (20) using the target biosignal and the blood pressure value labeled in the target biosignal. That is, the control unit (50) applies the target biosignal that satisfies a predetermined signal quality standard by the high-quality signal selection unit (31) and the predetermined blood pressure value assigned to the target biosignal to the calibration of the blood pressure estimation algorithm (20).

[0101] To improve the accuracy of a blood pressure estimation algorithm (20), a user can perform multiple calibrations by wearing a cuff blood pressure monitor and a signal measuring device (10) as in the past, but this incurs the inconvenience of having to wear a cuff blood pressure monitor. The present invention can improve user convenience and accuracy by having the control unit (50) utilize labeled bio-signals for calibration.

[0102] According to the present embodiment, the control unit (50) can calibrate the blood pressure estimation algorithm (20) using the target bio-signals reflecting the different states and the blood pressure values ​​labeled in the target bio-signals. That is, since the high-quality signal selection unit (31) selects a signal satisfying a predetermined signal quality standard among the bio-signals labeled by reflecting the different states of the user as the target bio-signal, the blood pressure estimation algorithm (20) can be calibrated to reflect the different states of the user, thereby deriving a more accurate blood pressure estimation value.

[0103] Additionally, according to an embodiment of the present invention, the control unit (50) may provide the user with guidance instructions for calibration when performing calibration. Specifically, the control unit (50) may provide the guidance instructions through a display unit (60) that displays information to the user. The control unit (50) and the display unit (60) may be provided in various electronic devices, such as the signal measuring device (10), mobile devices, wearable devices, and PCs.

[0104] The above guide instructions can provide guidance to the user regarding the timing and method of measurement, as well as connection to external devices. For example, the measurement timing can be guided to measure and compensate during the active period (e.g., immediately after waking up) and the resting period (e.g., immediately before going to bed). In addition, the user can be guided on compliance matters to be observed during measurement. For example, the guide instructions can include instructions such as measuring before showering or bathing, refraining from smoking or consuming caffeine within 30 minutes before measurement, and measuring after resting for 5 minutes after using the bathroom. In addition, the guide instructions can guide to comply with the blood pressure measurement guide recommended by medical institutions, such as twice in the morning and twice in the evening. The activity or resting period can be distinguished by collecting the ACC (acceleration) signal through the above-described acceleration sensor to distinguish the user's status. If the above guide instructions are not followed, the user can be notified of an alarm or additionally requested to perform compensation at a necessary time.

[0105] For example, according to the present embodiment, the ACC data is collected for about a week through the signal measuring device (10), and the data is analyzed to analyze the user's weekly lifestyle pattern. The pattern according to the user's state of waking up, sleeping, activity, and rest can be analyzed. The control unit (50) can induce manual calibration of the blood pressure estimation algorithm (20) at different time periods so as to reflect the user's different states.

[0106] In this way, the blood pressure estimation device applying pseudo calibration according to an embodiment of the present invention performs pseudo labeling, which assigns a predetermined blood pressure value to a signal acquired from a signal measuring device (10), selects a signal of excellent quality among the labeled bio-signals as a target bio-signal, and performs calibration of the blood pressure estimation algorithm (20), and at this time, reflects the different states of the users to provide an effect in which the blood pressure estimation algorithm (20) is optimized and calibrated for each user.

[0107]

[0108] Meanwhile, according to another aspect of the present invention, a blood pressure estimation method using pseudo calibration is provided. As illustrated in FIG. 7, the blood pressure estimation method according to an embodiment of the present invention includes a signal measurement step (S1), a pseudo labeling step (S2), a high-quality signal selection step (S3), and a pseudo calibration step (S4).

[0109] The signal measuring step (S1) is a step of measuring the user's bio-signal. The signal measuring step (S1) can be performed through the signal measuring device (10). The signal measuring step (S1) can collect the user's photoplethysmography signal (PPG signal) through the first sensor unit (11). In addition, the user's movement can be detected by the second sensor unit (12). The second sensor unit (12) can be an acceleration sensor that measures the magnitude or change amount of acceleration that occurs according to the movement pattern of the user's finger. The first sensor unit (11) can continuously or periodically collect bio-signals from the user in the different states, and the second sensor unit (12) can continuously or periodically measure the acceleration of the user's finger. The bio-signals collected from the first sensor unit (11) and the second sensor unit (12) are used to extract unique blood pressure characteristics and determine the user's state.

[0110] The above signal measurement step (S1) is substantially performed by the signal measurement device (10), and a repetitive description of the biosignals collected by the signal measurement device (10) as described above will be omitted. According to the present embodiment, the signal measurement step (S1) is performed through a ring worn on the user's finger. However, the measurement and collection of biosignals are not limited to the ring-type collection device.

[0111]

[0112] The above-described blood pressure labeling step (S2) is a step of labeling by assigning a predetermined blood pressure value corresponding to the biosignal acquired from the signal measuring device (10). The predetermined blood pressure value assigned corresponding to the biosignal can be assigned through the blood pressure estimation algorithm (20). The above-described blood pressure labeling step (S2) is performed by the blood pressure labeling unit (31). The labeled biosignal can be referred to as a blood pressure labeling in the sense that it does not assign an accurately measured blood pressure value, for example, a blood pressure value acquired using a cuff blood pressure monitor. The blood pressure estimation algorithm (20) is a pre-algorithmized one, and as described above, it can be algorithmized by utilizing the correlation between the blood pressure value actually measured from the user (e.g., acquired using a cuff blood pressure monitor) and the PPG signal. The above-described algorithm can be implemented by regression analysis or deep learning, etc., and a detailed description thereof will be omitted as described above.

[0113] According to the present embodiment, the signal measuring device (10) obtains bio-signals measured in at least two different states of the user, and the blood pressure labeling step (S2) labels the bio-signals measured in at least two different states of the user by assigning blood pressure values.

[0114] The above different states refer to cases where factors affecting the user's blood pressure are different. Generally, blood pressure is calculated as 'cardiac output × peripheral resistance', and cardiac output is defined as 'stroke volume × heart rate'. Therefore, states in which stroke volume, heart rate, or peripheral resistance are different as factors affecting blood pressure can be treated as different states for extracting blood pressure change characteristics. As described above, a resting state, an active state, and a sleeping state can be treated as different states that satisfy different environments or conditions. For example, the different states may include a distinction between an active / sleeping state, or a distinction between a systolic / diastolic state.

[0115]

[0116] The above-described excellent quality signal selection step (S3) is a step for selecting a target biosignal that satisfies a predetermined signal quality criterion among the biosignals labeled by the above-described water labeling step (S2). The above-described excellent quality signal selection step (S3) is performed by the excellent quality signal selection unit (31). The target biosignal selected by the above-described excellent quality signal selection unit (31) is a biosignal having a signal quality of a certain level or higher and is used when performing water calibration. The above-described signal quality criterion may be appropriately selected in consideration of noise level, signal variability, etc.

[0117] At this time, according to the present embodiment, since the water labeling step (S2) performs labeling on bio-signals reflecting different states of the user, the high-quality signal selection unit (31) can select a target bio-signal that satisfies a predetermined signal quality standard for each labeled bio-signal of different states.

[0118] Specifically, the above-described excellent quality signal selection step (S3) may select a first excellent quality signal that satisfies a predetermined criterion among the first measurement values ​​measured in the first state, and select a second excellent quality signal that satisfies a predetermined second criterion among the second measurement values ​​measured in the second state as the target biosignal. For example, the first state may be a resting state, and the second state may be an active state. The selection process or operation of the above-described target biosignal has been described above, and a repeated description thereof will be omitted.

[0119]

[0120] According to the present embodiment, a state determination step (SS) for determining the state of the user may be further included, so that bio-signals in different environments or conditions can be distinguished. For example, in the state determination step (SS), if a predetermined first pattern is measured by the second sensor unit (12) for a specified time, it may be determined that the user is in the first state, and if a predetermined second pattern is measured by the second sensor unit (12) for a specified time, it may be determined that the user is in the second state. The first state may be defined as a state in which the user is resting, and the second state may be defined as a state in which the user is active. Of course, the state of the user is not limited to the first and second states, and various states may be determined by differently distinguishing the state of a factor affecting blood pressure as described above. The state determination step (SS) may determine whether the user is sleeping or active from the bio-signals received from the signal measurement step (S1). For example, when the user is in a sleeping state, it may be defined as a third state.

[0121] In addition, according to the present embodiment, the state determination step (SS) can determine whether the ring is worn or not based on the size or shape of the signal measured by the first sensor unit (11). It can be utilized to define a state change such as a change from an active state to a resting state, a change from a resting state to a sleeping state, or a change from an active state to a resting state while the ring is worn. According to the present embodiment, the state determination step (SS) can be substantially performed through the state determination unit (40). Since the process of the state determination unit (40) defining each state through analysis of a biosignal has been described above, a detailed description thereof will be omitted.

[0122]

[0123] The above-described pseudo-calibration step (S4) is a step of pseudo-calibrating the blood pressure estimation algorithm (20) using the target bio-signal and the blood pressure value labeled in the target bio-signal. The pseudo-calibration step (S4) is performed by the control unit (50). That is, the pseudo-calibration step (S4) applies the target bio-signal that satisfies a predetermined signal quality standard by the high-quality signal selection unit (31) and the predetermined blood pressure value assigned to the target bio-signal to the calibration of the blood pressure estimation algorithm (20). The operation or effect of the pseudo-calibration step (S4) has been specifically described in the operation or effect of the control unit (50), and thus, a repeated description thereof will be omitted.

[0124] According to the present embodiment, the water calibration step (S4) can calibrate the blood pressure estimation algorithm (20) using the target bio-signals reflecting the different states and the blood pressure values ​​labeled in the target bio-signals. That is, in the water labeling step (S2), labeled bio-signals reflecting the different states of the user are secured, and the high-quality signal selection step (S3) selects a signal satisfying a predetermined signal quality standard among the labeled bio-signals as the target bio-signal, so that the blood pressure estimation algorithm (20) is calibrated using data reflecting the different states of the user, so that the estimated blood pressure value estimated by the blood pressure estimation algorithm (20) can be improved more accurately.

[0125] In this way, water calibration (S4) can be performed by extracting high-quality signals from a user's specific state. Of course, water calibration (S4) can also be performed by selecting high-quality signals from among water-labeled biosignals without reflecting a specific state. These biosignals can be acquired periodically or aperiodically.

[0126] The blood pressure estimation method using pseudo-calibration according to an embodiment of the present invention can provide substantially the same function or effect as the blood pressure estimation method using pseudo-calibration described above. That is, it provides the effect of calibrating a blood pressure estimation algorithm to optimize it for the user based on biometric information collected under different environments where major factors affecting blood pressure are different. A blood pressure estimation algorithm calibrated in this way can provide more accurate estimated blood pressure values.

[0127]

[0128] While the present invention has been described with reference to the embodiments illustrated in the drawings, these are merely exemplary, and those skilled in the art will appreciate that various modifications and variations of the embodiments are possible. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

Claims

1. A signal measuring device that measures the user's bio-signals; A blood pressure estimation algorithm that estimates blood pressure using the above biosignals; A blood pressure labeling unit that labels the bio-signal obtained from the signal measuring device by assigning a predetermined blood pressure value; A high-quality signal selection unit that selects a target biosignal that satisfies a predetermined signal quality standard among the biosignals labeled by the above-mentioned water labeling unit; and A blood pressure estimation device using pseudo calibration, characterized in that it includes a control unit that performs pseudo calibration on the blood pressure estimation algorithm using the target biosignal and the blood pressure value labeled in the target biosignal.

2. In paragraph 1, A blood pressure estimation device using pseudo calibration, characterized in that the signal measuring device includes a first sensor unit that collects a user's photoplethysmography signal (PPG signal).

3. In paragraph 1, The above signal measuring device obtains the bio-signals measured in at least two different states of the user, A blood pressure estimation device using pseudo calibration, characterized in that the above different states are states in which the user's cardiac output or peripheral resistance is in different environments.

4. In paragraph 1, A blood pressure monitoring device using pseudo calibration, characterized in that the signal measuring device includes a second sensor unit that detects the user's movement.

5. In paragraph 4, A blood pressure estimation device using pseudo calibration, characterized in that the second sensor unit is an acceleration sensor that measures the magnitude or amount of change in acceleration that occurs according to the movement pattern of the user's finger.

6. In paragraph 4, A blood pressure estimation device using pseudo calibration, characterized in that it further includes a state determination unit that determines that the user is in a first state when a predetermined first pattern is measured for a specified time by the second sensor unit, and that determines that the user is in a second state when a predetermined second pattern is measured for a specified time by the second sensor unit.

7. In paragraph 6, The above signal measuring device is in the form of a ring that is worn on the user's finger. A blood pressure estimation device using pseudo calibration, characterized in that the status determination unit determines whether the user is wearing the signal measuring device based on the size or shape of the signal measured by the first sensor unit.

8. In paragraph 6, A blood pressure estimation device that applies pseudo calibration, characterized in that the above-mentioned state determination unit determines whether the user is sleeping or active from the bio-signal received from the signal measuring device.

9. In paragraph 3, A blood pressure estimation device using pseudo calibration, characterized in that the labeling unit assigns and labels blood pressure values ​​to the bio-signals measured in at least two different states of the user.

10. In paragraph 8, The above-mentioned high-quality signal selection unit selects a target biosignal that satisfies a predetermined signal quality standard for each of the labeled biosignals of different states, A blood pressure estimation device using pseudo calibration, characterized in that the control unit calibrates the blood pressure estimation algorithm using the target biosignals reflecting the different states and the blood pressure values ​​labeled in the target biosignals.

11. In paragraph 1, The signal measuring device includes a first sensor unit that collects a user's photoplethysmography signal (PPG signal), and a second sensor unit that detects the user's movement. The first sensor unit collects bio-signals from the user continuously or periodically in the different states, A blood pressure estimation device using pseudo calibration, characterized in that the second sensor unit continuously or periodically measures the acceleration of the user's finger.

12. Signal measurement step for measuring the user's bio-signals; A blood pressure estimation step for estimating blood pressure using a blood pressure estimation algorithm that estimates blood pressure using the above biosignals; A water labeling step for labeling by assigning a predetermined blood pressure value corresponding to the bio-signal obtained from the signal measuring device; and A high-quality signal selection step for selecting a target biosignal that satisfies a predetermined signal quality standard among the biosignals labeled by the above-mentioned water labeling step; and A blood pressure estimation method using pseudo calibration, characterized in that it includes a pseudo calibration step of calibrating the blood pressure estimation algorithm using the target biosignal and the blood pressure value labeled in the target biosignal.

13. In paragraph 12, A blood pressure estimation method using pseudo calibration, characterized in that the signal measurement step is performed by a first sensor unit that collects a user's photoplethysmography signal (PPG signal).

14. In paragraph 12, The above signal measuring device obtains the bio-signals measured in at least two different states of the user, A blood pressure estimation method using pseudo calibration, characterized in that the above different states are states in which the user's cardiac output or peripheral resistance is in different environments.

15. In paragraph 12, The above signal measurement step is a blood pressure estimation method using pseudo calibration, characterized in that the user's movement is detected by a second sensor unit.

16. In paragraph 15, A blood pressure estimation method using pseudo calibration, characterized in that the second sensor unit is an acceleration sensor that measures the magnitude or amount of change in acceleration that occurs according to the movement pattern of the user's finger.

17. In paragraph 14, A blood pressure estimation method using pseudo calibration, characterized in that it further includes a state determination step of determining that the user is in the first state when a predetermined first pattern is measured for a specified time by the second sensor unit, and determining that the user is in the second state when a predetermined second pattern is measured for a specified time by the second sensor unit.

18. In paragraph 16, The above signal measurement step is performed through a ring worn on the user's finger. A blood pressure estimation method using pseudo calibration, characterized in that the above-mentioned status determination step determines whether the user is wearing the ring based on the size or shape of the signal measured by the first sensor unit.

19. In paragraph 16, The above state determination step is a blood pressure estimation method using pseudo calibration, characterized in that it determines whether the user is sleeping or active from the bio-signal received from the signal measurement step.

20. In paragraph 12, A blood pressure estimation method using pseudo calibration, characterized in that the labeling unit assigns and labels blood pressure values ​​to the bio-signals measured in at least two different states of the user.

21. In paragraph 20, The above-mentioned high-quality signal selection unit selects a target biosignal that satisfies a predetermined signal quality standard for each of the labeled biosignals of different states, A blood pressure estimation method using pseudo calibration, characterized in that the control unit calibrates the blood pressure estimation algorithm using the target biosignals reflecting the different states and the blood pressure values ​​labeled in the target biosignals.

22. In paragraph 12, The above signal measurement step collects the user's photoplethysmography signal (PPG signal) by the first sensor unit, and detects the user's movement by the second sensor unit. The first sensor unit collects bio-signals from the user continuously or periodically in the different states, A blood pressure estimation method using pseudo calibration, characterized in that the second sensor unit continuously or periodically measures the acceleration of the user's finger.

Images