Device and method for predicting viscosity by using artificial intelligence in viscosity prediction system

Abstract

An embodiment of the present invention provides a device and a method for predicting viscosity by using artificial intelligence in a viscosity prediction system in order to increase accuracy and convenience for viscosity measurement. A system according to an embodiment of the present invention includes: a communication unit for receiving PPG data of a user from a wearable device; a memory; and a control unit for generating a PPG segment by preprocessing the PPG data, predicting at least one parameter of a viscosity model by reflecting the PPG segment to a pre-trained artificial intelligence model, determining the viscosity of the user by using the predicted parameter, and storing the determined viscosity in the memory.

Description

Device and method for predicting astrology using artificial intelligence in an astrology prediction system

[0001] The present invention relates to a viscosity prediction system, and more particularly, to a system and method for predicting viscosity using artificial intelligence in a viscosity prediction system.

[0002] Recently, as various adult diseases, including vascular diseases, have increased, research is actively being conducted to detect the signs of adult diseases in advance. To this end, research is being conducted on devices and diagnostic kits that measure blood viscosity.

[0003] Common methods for testing the viscosity of fluids such as blood include the cone-plate rotation method and the scanning capillary method. The cone-plate rotation method involves placing blood between a rotating cone and a fixed plate, and measuring blood viscosity through rotational resistance. The scanning capillary method calculates viscosity by measuring the pressure required to maintain a constant flow rate of blood through a capillary.

[0004] These methods had the inconvenience of requiring the patient to directly draw blood, and the associated pain. Furthermore, these methods raised concerns about errors caused by environmental changes when measuring viscosity across various shear rates. For example, blood viscosity values ​​can fluctuate frequently due to water loss.

[0005] Therefore, a solution to solve these problems is needed.

[0006] The purpose of the present invention to solve the above-mentioned purpose is to predict viscosity using artificial intelligence in a viscosity prediction system to increase the accuracy of viscosity measurement.

[0007] And the purpose of the present invention is to predict viscosity using artificial intelligence in a viscosity prediction system to increase convenience in viscosity measurement.

[0008] The technical problems to be solved by the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned can be clearly understood by a person having ordinary skill in the technical field to which the present invention belongs from the description below.

[0009] The system of the present invention for achieving the above purpose includes a communication unit for receiving a user's PPG data from a wearable device; a memory; and a control unit for preprocessing the PPG data to generate a PPG segment, reflecting the PPG segment in a pre-learned artificial intelligence model to predict at least one parameter of a viscosity model, determining the user's viscosity using the predicted parameter, and storing the determined viscosity in the memory.

[0010] The method of the present invention for achieving the above object includes a process in which a control unit receives a user's PPG data from a wearable device through a communication unit, a process in which the control unit preprocesses the PPG data to generate a PPG segment, a process in which the control unit predicts at least one parameter of a viscosity model by reflecting the PPG segment in a pre-learned artificial intelligence model, a process in which the control unit determines the user's viscosity using the predicted parameter, and a process in which the control unit stores the determined viscosity in a memory.

[0011] The effect of the present invention according to the above configuration is to increase the accuracy of viscosity measurement by predicting viscosity through artificial intelligence in a viscosity prediction system.

[0012] And the effect of the present invention is to increase the convenience of viscosity measurement by predicting viscosity through artificial intelligence in an astrology prediction system.

[0013] The effects of the present invention are not limited to the effects described above, and should be understood to include all effects that can be inferred from the detailed description of the present invention or the composition of the invention described in the claims.

[0014] FIG. 1 is a simplified configuration diagram of an astrology prediction system according to one embodiment of the present invention.

[0015] Figure 2 is a block diagram of an astrocyte device according to one embodiment of the present invention.

[0016] FIG. 3 is a diagram of generating PPG data in a wearable device according to an embodiment of the present invention.

[0017] FIG. 4 is a diagram for preprocessing PPG data in an astrological device according to one embodiment of the present invention.

[0018] FIG. 5 is a diagram for predicting parameters of a viscosity model in a viscosity prediction system according to an embodiment of the present invention.

[0019] FIG. 6 is a graph illustrating the performance of an astrology prediction system according to an embodiment of the present invention.

[0020] Figure 7 is a flowchart for learning an artificial intelligence model in an astrology prediction system according to an embodiment of the present invention.

[0021] Figure 8 is a flowchart for predicting viscosity using an artificial intelligence model learned in an astrology prediction system according to one embodiment of the present invention.

[0022] Hereinafter, the present invention will be described with reference to the attached drawings. However, the present invention can be implemented in various different forms and is therefore not limited to the embodiments described herein. In the drawings, irrelevant parts have been omitted for clarity of description, and similar parts have been designated with similar reference numerals throughout the specification.

[0023] Throughout the specification, when a part is said to be "connected (connected, contacted, or coupled)" to another part, this includes not only cases where it is "directly connected," but also cases where it is "indirectly connected" with another member in between. Furthermore, when a part is said to "include" a component, this does not mean that it excludes other components, but rather that it may include other components, unless otherwise specifically stated.

[0024] The terminology used herein is merely used to describe specific embodiments and is not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this specification, it should be understood that the terms "comprises" or "has" indicate the presence of a feature, number, step, operation, component, part, or combination thereof described in the specification, but do not exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0025] The present invention will be described in detail with reference to the attached drawings below.

[0026]

[0027] FIG. 1 is a simplified configuration diagram of an astrology prediction system according to an embodiment of the present invention. Referring to FIG. 1, the astrology prediction system includes an astrology device (101) and a wearable device (103).

[0028] Looking at each component, the wearable device (103) generates PPG data using a conventional pulse measurement method (Photoplethysmography, hereinafter referred to as 'PPG') and transmits the generated PPG data to the astrology device (101) via wired or wireless communication. In Fig. 1, the wearable device (103) is illustrated as a watch, but the present invention is not limited thereto. For example, the wearable device (103) may be a ring, bracelet, necklace, or band in addition to a watch.

[0029] For example, the PPG may be a blood pressure wave measuring module equipped in a wearable device (103). For example, the PPG may irradiate light to a user's finger (e.g., a thumb) via a light source (e.g., a red light or an infrared LED) to estimate the viscosity of blood, as shown in the drawing (301) of FIG. 3. In addition, the PPG may detect light reflected or transmitted through the thumb via a light sensor (e.g., a photodetector) and convert it into an electrical signal to generate PPG data. For example, the PPG data may include an AC Pulsatile Part representing the systolic phase and the diastole phase of the pulsatile component of arterial blood, and a DC Steady Part representing the non-pulsatile component of arterial blood.

[0030] The astrology device (101) trains an artificial intelligence model using pre-provided PPG data. For example, in FIG. 1, the astrology device (101) is depicted as a desktop, but is not limited thereto. For example, the astrology device (101) may be any device equipped with an artificial intelligence model for predicting astrology. For example, the astrology device (101) may be a smartphone, a laptop, or a wearable device.

[0031] For example, the AI ​​model may be a model for deriving at least one parameter for a blood viscosity model. For example, the AI ​​model may be a 1-D convolutional neural network (CNN). For example, the blood viscosity model may be the Carreau-Yasuda model, which can simulate the behavior of blood.

[0032] The viscosity device (101) receives PPG data from a wearable device (103) via wireless or wired communication, analyzes the received PPG data using a pre-learned artificial intelligence model, and derives at least one parameter of a viscosity model. The viscosity device (101) applies the derived parameter to the viscosity model to predict the user's blood viscosity, and stores the predicted blood viscosity corresponding to the user.

[0033] From now on, the detailed configuration of the astrology device (101) will be explained with reference to Fig. 2.

[0034]

[0035] Figure 2 is a block diagram of an astrocyte device (101) according to one embodiment of the present invention.

[0036] Referring to FIG. 2, the astrological device (101) includes a control unit (201), a memory (203), a display unit (205), a communication unit (207), and an input / output unit (209).

[0037] Looking at each component, the memory (203) stores various programs and data necessary for the operation of the astrocyte device (101). For example, the memory (203) may be implemented as a non-volatile memory, a volatile memory, a flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

[0038] The display unit (205) displays image data under the control of the control unit (201). The implementation method of the display unit (205) is not limited, and may be implemented as various types of displays such as, for example, an LCD (Liquid Crystal Display), an OLED (Organic Light Emitting Diodes) display, an AMOLED (Active-Matrix Organic Light-Emitting Diode), a PDP (Plasma Display Panel), etc. The display unit (205) may additionally include additional components depending on its implementation method. For example, when the display unit (205) is a liquid crystal display, the display unit (205) may include an LCD display panel (not shown), a backlight unit (not shown) that supplies light thereto, and a panel driving substrate (not shown) that drives the panel (not shown). The display unit (205) may be combined with a touch panel (not shown) of the input / output unit (209) to be provided as a touch screen (not shown).

[0039] The input / output unit (209) receives various commands from the user or outputs various data to an external device. For example, the input / output unit (209) may include at least one of a key, a touch panel, and a pen recognition panel.

[0040] The communication unit (207) communicates with various types of external devices according to various types of wired or wireless communication methods. For example, the communication unit (207) may include at least one of a Wi-Fi module and an LTE (Long Term Evolution) module. For example, the Wi-Fi module may communicate using the Wi-Fi method, and the LTE module may communicate according to various communication standards such as IEEE and LTE.

[0041] The control unit (201) controls the overall operation of the viscosity device (101) using various programs stored in the memory (204).

[0042] For example, the control unit (201) can train an artificial intelligence model (or artificial intelligence algorithm) that can predict the non-Newtonian fluid viscosity behavior of blood.

[0043] To explain in more detail, the control unit (201) receives PPG data from the wearable device (103) through the communication unit (207). For example, the PPG data may be data representing the pulse waveform of the user's skin blood vessels.

[0044] And the control unit (201) preprocesses the received PPG data to generate a PPG segment. For example, the control unit (201) can generate a PPG segment by removing the first or second deviation signals (Wandering1, Wandering2) from the PPG data through a noise filter (not shown), as shown in the graph (401) illustrated in FIG. 4. For example, the deviation signal can indicate a signal that rises or falls by 20% relative to the signal baseline over a period of 1 minute.

[0045] The control unit (201) then applies the PPG segment, which is preprocessed PPG data, to the artificial intelligence model to predict at least one parameter of the viscosity model. For example, if the viscosity model is the Carreau-Yasuda model, the Carreau-Yasuda model can be expressed by the following <Mathematical Formula 1>.

[0046]

[0047]

[0048]

[0049] For example, μ ∞ represents the infinite-shear-dependent viscosity, μ0 can represent the zero-shear-dependent viscosity, and n represents the slope of the power law, and α can represent a parameter that affects the shape of the transition region. And λ represents the time constant that determines the point where it changes from a constant to a power law, can represent the shear rate. The shear rate quantifies the velocity gradient between adjacent blood layers under an applied shear stress, which is an important parameter in describing the non-Newtonian viscosity behavior of blood in the Carreau-Yasuda model.

[0050]

[0051] For example, if the artificial intelligence model is a 1-D CNN, the 1-D CNN can predict multiple parameters of the Carreau-Yasuda model by analyzing the PPG segment, which is an input signal, twice through the 1-D CNN, as shown in the drawing (501) in Fig. 5. For example, the multiple parameters are Y1, Y2, Y3, Y4, and Y5, and Y1 is μ, a parameter of the Carreau-Yasuda model. ∞, Y2 is μ0, a parameter of the Carreau-Yasuda model, Y3 is γ, a parameter of the Carreau-Yasuda model, Y4 is α, a parameter of the Carreau-Yasuda model, and Y5 can be n, a parameter of the Carreau-Yasuda model.

[0052] The control unit (201) then determines the first viscosity for the user by applying the predicted parameters to the viscosity model. For example, the control unit (201) may determine the first viscosity by applying the predicted parameters to Equation 1 of the Carreau-Yasuda model.

[0053] The control unit (201) receives viscosity data from the viscometer via the communication unit (207). For example, the viscometer can determine blood viscosity based on blood directly collected from the user and generate viscosity data. For example, the viscosity data may be data representing the viscosity of the user's blood.

[0054] The control unit (201) then determines the second viscosity and actual parameters for the user using the received viscosity data. For example, the actual parameters may represent parameters of a viscosity model derived based on viscosity data received from a viscometer.

[0055] And the control unit (201) calculates the loss of the artificial intelligence model using the first viscosity, the second viscosity, and the predicted parameters and the actual parameters.

[0056] For example, Loss can be expressed by the following <Mathematical Formula 2>.

[0057]

[0058]

[0059]

[0060]

[0061]

[0062] For example, the predicted value of a viscosity model may be a first viscosity, while the actual viscosity may be a second viscosity. The viscosity model parameters may represent parameters predicted by the AI ​​model, while the actual parameters may represent parameters of the viscosity model derived from viscosity data received from a viscometer. w1 and w2 may represent weights.

[0063]

[0064] At this time, the control unit (201) sets the weights (e.g., w) so that the loss is minimized. 1, w2) can be determined.

[0065] The control unit (201) trains the artificial intelligence model by reflecting the losses and weights determined in the artificial intelligence model. That is, the control unit (201) can train the artificial intelligence model so that the artificial intelligence model can accurately predict blood viscosity by repeatedly performing this process.

[0066] For example, the control unit (201) can predict the user's astrology using a pre-learned artificial intelligence model.

[0067] To explain in more detail, the control unit (201) receives PPG data from the wearable device (103) through the communication unit (207).

[0068] And the control unit (201) preprocesses the received PPG data to generate a PPG segment. For example, the control unit (201) can generate a PPG segment by removing the first or second deviation signals (Wandering1, Wandering2) from the PPG data through a noise filter (not shown), as shown in the graph (401) illustrated in FIG. 4.

[0069] And the control unit (201) predicts at least one parameter of the viscosity model using the PPG segment, which is the PPG data preprocessed through a pre-learned artificial intelligence model. For example, if the artificial intelligence model is a 1-D CNN, the 1-D CNN can predict multiple parameters of the Carreau-Yasuda model by analyzing the PPG segment, which is the input signal, through the 1-D CNN, as shown in the drawing (501) in FIG. 5.

[0070] The control unit (201) determines blood viscosity for the user using at least one predicted parameter. For example, the control unit (201) can determine blood viscosity for the user by applying at least one predicted parameter to <Mathematical Formula 1> of the Carreau-Yasuda model.

[0071] And the control unit (201) stores the viscosity determined in response to the user in the memory (203) and displays the stored viscosity according to the user's request through the display unit (205).

[0072] Thereafter, the control unit (201) derives a blood viscosity model that reflects individual user (or patient) characteristics based on the stored viscosity. That is, the control unit (201) can predict not only the viscosity at the systolic or diastolic point, but also the viscosity behavior of the entire blood.

[0073] The control unit (201) then acquires flow data that varies depending on the viscosity based on the stored viscosity. For example, the flow data can be utilized for learning an artificial intelligence algorithm that predicts a flow field in real time when shape and user (or patient) information are provided through the acquired flow data.

[0074] Through this configuration, one embodiment of the present invention can predict viscosity using artificial intelligence in a viscosity prediction system, thereby increasing the accuracy of viscosity measurement. Furthermore, one embodiment of the present invention can predict viscosity using artificial intelligence in a viscosity prediction system, thereby increasing the convenience of viscosity measurement.

[0075]

[0076] FIG. 6 is a graph illustrating the performance of an astrology prediction system according to an embodiment of the present invention.

[0077] Referring to the graph (601) illustrated in FIG. 6, it can be confirmed that the accuracy (81.8%) in all ranges for the artificial intelligence algorithm according to one embodiment of the present invention that considers weights is higher than the accuracies of other artificial intelligence algorithms.

[0078] And it can be confirmed that the accuracy (84%) in the physiological sheer range for the artificial intelligence algorithm according to one embodiment of the present invention is higher than the accuracies of other artificial intelligence algorithms.

[0079]

[0080] Figure 7 is a flowchart for learning an artificial intelligence model in an astrology prediction system according to an embodiment of the present invention.

[0081] Referring to FIG. 7, a wearable device (103) shines light on a part of the user's body, generates PPG data using the reflected light, and transmits the generated PPG data to a viscosity device (101).

[0082] The control unit (201) of the astrology device (101) receives PPG data from the wearable device (103) through the communication unit (207) in step 701.

[0083] In step 703, the control unit (201) preprocesses the received PPG data to generate a PPG segment. For example, the control unit (201) may generate a PPG segment by removing the first or second deviation signals (Wandering1, Wandering2) from the PPG data through a noise filter (not shown), as shown in the graph (401) illustrated in FIG. 4.

[0084] At step 705, the control unit (201) applies the PPG segment, which is the preprocessed PPG data, to the artificial intelligence model to predict at least one parameter of the viscosity model.

[0085] For example, if the viscosity model is the Carreau-Yasuda model, the Carreau-Yasuda model can be expressed as in <Mathematical Formula 1>. For example, if the artificial intelligence model is a 1-D CNN, the 1-D CNN can predict a number of parameters of the Carreau-Yasuda model by analyzing the PPG segment, which is an input signal, through the 1-D CNN, as shown in the drawing (501) in FIG. 5.

[0086] At step 707, the control unit (201) determines the first viscosity for the user by applying the predicted parameters to the viscosity model. For example, the control unit (201) can determine the first viscosity by applying the predicted parameters to <Mathematical Formula 1> representing the Carreau-Yasuda model.

[0087] At step 709, the control unit (201) receives viscosity data from the viscometer via the communication unit (207). For example, the viscometer can generate viscosity data by determining blood viscosity based on blood directly collected from a user.

[0088] At step 711, the control unit (201) determines the second viscosity and actual parameters for the user using the received viscosity data.

[0089] In step 713, the control unit (201) calculates the loss of the artificial intelligence model using the first viscosity, the second viscosity, the predicted parameters, and the actual parameters. For example, the loss can be expressed as in <Mathematical Formula 2>. At this time, the control unit (201) sets the weights (e.g., w) so that the loss is minimized. 1, w2) can be determined.

[0090] In step 715, the control unit (201) trains the artificial intelligence model by reflecting the determined losses and weights in the artificial intelligence model. Thereafter, the control unit (201) can repeatedly perform this process to train the artificial intelligence model so that it can accurately predict blood viscosity.

[0091]

[0092] Figure 8 is a flowchart for predicting viscosity using an artificial intelligence model learned in an astrology prediction system according to one embodiment of the present invention.

[0093] Referring to FIG. 8, a wearable device (103) shines light on a part of the user's body, generates PPG data using the reflected light, and transmits the generated PPG data to a viscosity device (101).

[0094] The control unit (201) of the astrology device (101) receives PPG data from the wearable device (103) through the communication unit (207) in step 801.

[0095] In step 803, the control unit (201) preprocesses the received PPG data to generate a PPG segment. For example, the control unit (201) may generate a PPG segment by removing the first or second deviation signals (Wandering1, Wandering2) from the PPG data through a noise filter (not shown), as shown in the graph (401) illustrated in FIG. 4.

[0096] In step 805, the control unit (201) predicts at least one parameter of the viscosity model using the PPG segment, which is the preprocessed PPG data, through a pre-learned artificial intelligence model. For example, if the artificial intelligence model is a 1-D CNN, the 1-D CNN can predict multiple parameters of the Carreau-Yasuda model by analyzing the PPG segment, which is the input signal, through the 1-D CNN, as shown in the drawing (501) in FIG. 5.

[0097] At step 807, the control unit (201) determines blood viscosity for the user using at least one predicted parameter. For example, the control unit (201) may determine blood viscosity for the user by applying at least one predicted parameter to <Mathematical Formula 1> of the Carreau-Yasuda model.

[0098] At step 809, the control unit (201) stores the viscosity determined in response to the user in the memory (203) and displays the stored viscosity according to the user's request through the display unit (205).

[0099] Through this process, one embodiment of the present invention can predict viscosity using artificial intelligence in a viscosity prediction system, thereby increasing the accuracy of viscosity measurement. Furthermore, one embodiment of the present invention can predict viscosity using artificial intelligence in a viscosity prediction system, thereby increasing the convenience of viscosity measurement.

[0100]

[0101] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will readily appreciate that the present invention can be readily modified into other specific forms without altering the technical spirit or essential characteristics of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single entity may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined manner.

[0102] The scope of the present invention is indicated by the claims described below, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention.

Claims

1. A communication unit that receives the user's PPG data from a wearable device; memory; and A device for predicting viscosity using artificial intelligence in a viscosity prediction system, comprising a control unit that preprocesses the PPG data to generate a PPG segment, reflects the PPG segment in a pre-learned artificial intelligence model to predict at least one parameter of a viscosity model, determines the viscosity of the user using the predicted parameter, and stores the determined viscosity in the memory.

2. In paragraph 1, A device for predicting viscosity using artificial intelligence in a viscosity prediction system characterized in that the above artificial intelligence model is a 1-deconvolutional neural network.

3. In paragraph 1, A device for predicting viscosity using artificial intelligence in a viscosity prediction system characterized in that the above viscosity model is a Karo-Yasuda model.

4. In paragraph 1, A device for predicting viscosity using artificial intelligence in a viscosity prediction system, characterized in that the control unit receives PPG data from a wearable device through the communication unit, preprocesses the PPG data to generate a PPG segment, reflects the PPG segment in an artificial intelligence model to predict at least one parameter of a viscosity model, determines a first viscosity using the predicted parameter, receives viscosity data from a viscometer through the communication unit, determines a second viscosity using the viscosity data, determines a loss of the artificial intelligence model using the first viscosity and the second viscosity, determines at least one weight for the loss so that the loss is minimized, and trains the artificial intelligence model by reflecting the determined loss and the determined weight in the artificial intelligence model.

5. In paragraph 4, A device for predicting viscosity using artificial intelligence in a viscosity prediction system, characterized in that the loss includes a first loss comparing the first viscosity and the second viscosity, a second loss comparing the predicted parameter and an actual parameter derived using the viscosity data, and a third loss determined using the determined weight and the first loss and the second loss.

6. The process of the control unit receiving the user's PPG data from the wearable device through the communication unit. The above control unit preprocesses the PPG data to generate a PPG segment, The above control unit predicts at least one parameter of the viscosity model by reflecting the PPG segment in a pre-learned artificial intelligence model. The process of the above control unit determining the user's viscosity using the predicted parameters, and A method for predicting viscosity using artificial intelligence in a viscosity prediction system including a process in which the control unit stores the determined viscosity in a memory.

7. In paragraph 6, A method for predicting viscosity using artificial intelligence in a viscosity prediction system, wherein the artificial intelligence model is a 1-deconvolutional neural network.

8. In paragraph 6, A method for predicting viscosity using artificial intelligence in a viscosity prediction system characterized in that the above viscosity model is a Karo-Yasuda model.

9. In paragraph 6, The above control unit receives PPG data from a wearable device through the above communication unit, The above control unit preprocesses the PPG data to generate a PPG segment, The above control unit predicts at least one parameter of the viscosity model by reflecting the PPG segment in the artificial intelligence model, The process of the above control unit determining the first viscosity using the predicted parameters, The process of the above control unit receiving viscosity data from the viscometer through the communication unit and determining the second viscosity using the viscosity data, The process of the above control unit determining the loss of the artificial intelligence model using the first viscosity and the second viscosity, The process of the control unit determining at least one weight for the loss so that the loss is minimized, and A method for predicting viscosity using artificial intelligence in an astrology prediction system, wherein the control unit further includes a process of training the artificial intelligence model by reflecting the determined loss and the determined weight to the artificial intelligence model.

10. In paragraph 9, A method for predicting viscosity using artificial intelligence in a viscosity prediction system, characterized in that the loss includes a first loss comparing the first viscosity and the second viscosity, a second loss comparing the predicted parameter and an actual parameter derived using the viscosity data, and a third loss determined using the determined weight and the first loss and the second loss.

Images