System for correcting posture by using artificial intelligence, and operation method of system

Abstract

The present disclosure relates to a posture correction system which comprises at least one from among a server, a user terminal and a posture sensor and which corrects posture by using artificial intelligence. An operation method of the posture correction system comprises steps in which: the user terminal sets reference range information of correct posture on the basis of an input of a user; daily posture information of the user is acquired on the basis of a daily posture sensor, which is included in the posture sensor and measures the inclination of the spine of the user; and, when the daily posture information is not included in the reference range information, the user terminal outputs a message indicating that the posture is to be corrected.

Description

System for correcting posture using artificial intelligence and method of operating the system

[0001] The present disclosure relates to a system and its operating method for correcting posture using artificial intelligence. More specifically, the present disclosure utilizes artificial intelligence to three-dimensionally reconstruct a user's current posture. Furthermore, the present disclosure utilizes artificial intelligence to predict diseases that may occur if the current posture is maintained.

[0002]

[0003] Among the common ailments of modern people, musculoskeletal disorders are often caused by poor posture or prolonged use of the same posture due to increased sedentary lifestyles, computer-based work environments, and increased use of portable devices such as smartphones and tablets. Their incidence is on the rise. In particular, modern people with sedentary lifestyles often repeatedly bow their heads or bend forward. If these postures persist, the head gradually shifts forward from the center of the body, leading to hunched shoulders and a rounded back. Furthermore, the weight of the head causes tension in the neck muscles, tightening the muscles connected to them. Shoulder muscles also tense up to support the weight of the head, easily leading to fatigue. Furthermore, a hunched back causes the spinal joints to shift backward from the center of the body, making the body's center of gravity shift to the left or right or tilt.

[0004] Therefore, maintaining proper posture is very important for modern people, so chairs and desks are being made ergonomically for sedentary lifestyles, and recently, technologies for posture correction have been introduced.

[0005] However, until now, users have not received appropriate feedback for posture correction, and posture correction tailored to each user has not been properly implemented.

[0006] [Prior Art Literature]

[0007] [Patent Document]

[0008] (Patent Document 0001) Korean Patent Publication No. 10-1480026 (December 31, 2014)

[0009]

[0010] According to one embodiment of the present disclosure, a method for operating a posture correction system for correcting posture using artificial intelligence, which includes at least one of a server, a user terminal, and a posture sensor, includes a step of allowing the user terminal to set reference range information of a correct posture based on a user's input, a step of obtaining the user's daily posture information based on a daily posture sensor included in the posture sensor and measuring the inclination of the user's spine, a step of outputting a message indicating that the user terminal will correct the posture if the daily posture information is not included in the reference range information, a step of obtaining the user's posture image information based on a motion posture sensor included in the posture sensor and photographing the user's entire body, a step of the server applying at least one of the posture image information and the daily posture information to a posture estimation machine learning model to obtain three-dimensional (3D) posture information, a step of the server determining a posture similarity between the 3D posture information and a predetermined reference posture information, and a step of outputting a message indicating that the user terminal will correct the posture if the posture similarity is less than a predetermined threshold similarity.

[0011] The step of setting reference range information of the operating method of the posture correction system according to one embodiment of the present disclosure includes the step of receiving user input related to reference range information from a user, a step of determining, by the user terminal, predetermined initial range information as reference range information when the user input indicates automatic range setting, a step of displaying, by the user terminal, a user interface for range setting, a step of obtaining, by the user terminal, an inclination of the user's spine based on a daily posture sensor, a step of changing, by the user terminal, an inclination of a body object representing at least a part of a human body included in a user interface based on the inclination of the spine and outputting the changed inclination, and a step of determining, by the user terminal, an inclination of the spine from 0 degrees as reference range information when the user confirms the inclination of the body object and inputs an input for an object indicating confirmation.

[0012] The step of obtaining daily posture information of the operating method of the posture correction system according to one embodiment of the present disclosure includes the step of displaying a daily posture object including a first area, a second area, a third area, and a fourth area, and the step of displaying at least one of the first area, the second area, the third area, and the fourth area in a predetermined color based on the daily posture information.

[0013] The step of displaying in a predetermined color of the operating method of the posture correction system according to one embodiment of the present disclosure includes the step of displaying a first area in a predetermined color when the daily posture information indicates that the user is leaning forward, the step of displaying a second area in a predetermined color when the daily posture information indicates that the user is leaning backward, the step of displaying a third area in a predetermined color when the daily posture information indicates that the user is leaning to the left, and the step of displaying a fourth area in a predetermined color when the daily posture information indicates that the user is leaning to the right.

[0014] A method of operating a posture correction system according to one embodiment of the present disclosure comprises: a first region including regions 1-1 and 1-2, a second region including regions 2-1 and 2-2, a third region including regions 3-1 and 3-2, and a fourth region including regions 4-1 and 4-2; a step of displaying region 1-1 in a predetermined color when daily posture information indicates that the user has leaned forward and the size information of the tilt included in the daily posture information is included in the reference range information; a step of displaying region 1-2 in a predetermined color when daily posture information indicates that the user has leaned forward and the size information of the tilt included in the daily posture information is not included in the reference range information; a step of displaying region 2-1 in a predetermined color when daily posture information indicates that the user has leaned backward and the size information of the tilt included in the daily posture information is included in the reference range information; The method further comprises: a step of indicating that the user is tilted backward, and if the size information of the tilt included in the daily posture detailed information is not included in the reference range information, displaying the 2-2 area in a predetermined color; a step of indicating that the user is tilted to the left, and if the size information of the tilt included in the daily posture detailed information is included in the reference range information, displaying the 3-1 area in a predetermined color; a step of indicating that the user is tilted to the left, and if the size information of the tilt included in the daily posture detailed information is not included in the reference range information, displaying the 3-2 area in a predetermined color; a step of indicating that the user is tilted to the right, and if the size information of the tilt included in the daily posture detailed information is not included in the reference range information, displaying the 4-1 area in a predetermined color; and a step of indicating that the user is tilted to the right, and if the size information of the tilt included in the daily posture detailed information is not included in the reference range information, displaying the 4-2 area in a predetermined color.

[0015] An operating method of a posture correction system according to one embodiment of the present disclosure comprises the steps of: a user terminal accumulating daily posture information at a predetermined cycle for a predetermined collection time to obtain accumulated daily posture information; a step of obtaining a time when the daily posture information, which is not included in the reference range information among the accumulated daily posture information and indicates a forward leaning, as a forward bad posture time; a step of obtaining a time when the daily posture information, which is not included in the reference range information among the accumulated daily posture information and indicates a backward leaning, as a backward bad posture time; a step of obtaining a time when the daily posture information, which is not included in the reference range information among the accumulated daily posture information and indicates a left leaning, as a left bad posture time; a step of obtaining a time when the daily posture information, which is not included in the reference range information among the accumulated daily posture information and indicates a right leaning, as a right bad posture time; a step of displaying a first area darker, darker, or more opaque as the forward bad posture time increases; a step of displaying a second area darker, darker, or more opaque as the backward bad posture time increases. The step includes a step of displaying the third area darker, darker, or more opaque as the left bad posture time increases, and a step of displaying the fourth area darker, darker, or more opaque as the right bad posture time increases.

[0016] An operating method of a posture correction system according to an embodiment of the present disclosure comprises the steps of: a user terminal accumulating daily posture information at a predetermined cycle for a predetermined collection time to obtain accumulated daily posture information; a step of obtaining a time when the daily posture information indicating a forward leaning is present in the accumulated daily posture information as a forward posture time; a step of obtaining a time when the daily posture information indicating a backward leaning is present in the accumulated daily posture information as a backward posture time; a step of obtaining a time when the daily posture information indicating a left leaning is present in the accumulated daily posture information as a left posture time; a step of obtaining a time when the daily posture information indicating a right leaning is present in the accumulated daily posture information as a right posture time; a step of displaying a first area darker, darker, or more opaque as the forward posture time increases; a step of displaying a second area darker, darker, or more opaque as the backward posture time increases; a step of displaying a third area darker, darker, or more opaque as the left posture time increases; and a step of displaying a fourth area darker, darker, or more opaque as the right posture time increases. Includes steps to display dark, dark, or opaque.

[0017] An operating method of a posture correction system according to one embodiment of the present disclosure includes a step of allowing a user terminal to accumulate daily posture information at a predetermined cycle for a predetermined collection time to obtain accumulated daily posture information, a step of obtaining a time when the user terminal has daily posture information included in the reference range information among the accumulated daily posture information as a correct posture time, a step of obtaining a time when the user terminal has daily posture information not included in the reference range information among the accumulated daily posture information as a bad posture time, a step of obtaining a posture score by dividing the correct posture time by the collection time, and a step of displaying the posture score.

[0018] An operating method of a posture correction system according to one embodiment of the present disclosure includes a step of obtaining a posture pattern of a user by at least one of a user terminal and a posture sensor, and a step of selecting at least one of reference range information and a posture estimation machine learning model corresponding to the posture pattern, wherein the posture pattern includes at least one of a standing posture, a sitting posture, a walking posture, and a running posture.

[0019] The three-dimensional posture information of the operating method of the posture correction system according to one embodiment of the present disclosure includes a plurality of body feature nodes at least partially corresponding to joints constituting the body and a skeletal edge connecting the plurality of body feature nodes, and the step of acquiring the three-dimensional posture information includes a step of corresponding the angle of the skeletal edge corresponding to the spine included in the three-dimensional posture information to the ground to the daily posture information.

[0020] An operating method of a posture correction system according to one embodiment of the present disclosure includes a step of selecting exercise information beneficial to the user from among a plurality of candidate exercise information based on daily posture information by the user terminal, a step of outputting the exercise information by the user terminal, and a step of obtaining posture image information of the user based on a movement posture sensor when the user assumes a posture based on the exercise information.

[0021] The step of selecting exercise information in the operating method of the posture correction system according to one embodiment of the present disclosure includes a step of the user terminal obtaining disease information from the user, a step of selecting some candidate exercise information by excluding exercise information harmful to the user from a plurality of candidate exercise information based on the disease information, and a step of selecting exercise information based on daily posture information from among the candidate exercise information.

[0022] The step of obtaining daily posture information in the operating method of the posture correction system according to one embodiment of the present disclosure includes the step of displaying a daily posture object including at least one concentric circle around an object viewed from above by the user, and the step of displaying a color corresponding to size information included in the daily posture information at a position corresponding to direction information included in the daily posture information in the daily posture object.

[0023] An operating method of a posture correction system according to one embodiment of the present disclosure includes a step in which a user terminal accumulates daily posture information at a predetermined cycle for a predetermined collection time to obtain accumulated daily posture information, a step in which a server applies the accumulated daily posture information to a spinal disease estimation machine learning model to obtain a disease probability for each of a plurality of spinal diseases, a step in which a highest probability value is greater than or equal to a predetermined threshold probability among the disease probabilities for each of a plurality of spinal diseases is selected, and a step in which a spinal disease corresponding to the highest probability value is selected from the plurality of spinal diseases and determined as spinal disease information.

[0024] The step of determining the posture similarity between the 3D posture information and the predetermined reference posture information by the user terminal of the operating method of the posture correction system according to one embodiment of the present disclosure includes the step of determining whether the 3D posture information matches the correct posture information for each predetermined body part of the user, the step of increasing the health score of the corresponding body part as the 3D posture information matches the correct posture information, and the step of displaying a graph indicating the health score for each body part in one area of ​​the screen.

[0025] Additionally, a program for implementing the operating method of the posture correction system as described above can be recorded on a computer-readable recording medium.

[0026]

[0027] The posture correction system of the present disclosure can acquire a user's 3D posture information using a single camera. This means that it can acquire the user's 3D posture information accurately and inexpensively. Since the posture correction system of the present disclosure acquires 3D posture information, it can provide the user with a 3D posture correction method. Consequently, the user's posture can be improved more efficiently and accurately.

[0028] Additionally, the posture correction system of the present disclosure can measure a user's posture by linking with various posture sensors, not just a camera. This eliminates the need to film the user with a camera, and measures the user's posture in real time, thereby improving the user's posture in daily life.

[0029]

[0030] FIG. 1 is a drawing showing a user terminal according to one embodiment of the present disclosure.

[0031] FIG. 2 is a drawing for explaining the configuration of a posture correction system according to one embodiment of the present disclosure.

[0032] FIG. 3 is a flowchart showing the operation of a posture correction system according to one embodiment of the present disclosure.

[0033] Figure 4 shows numerical values ​​related to automatic range setting and manual range setting according to one embodiment of the present disclosure.

[0034] FIG. 5 is a flowchart showing the operation of a posture correction system according to one embodiment of the present disclosure.

[0035] FIG. 6 illustrates a user interface related to range setting displayed on a user terminal according to one embodiment of the present disclosure.

[0036] FIG. 7 is a flowchart showing the operation of a posture correction system according to one embodiment of the present disclosure.

[0037] FIG. 8 is a diagram illustrating a self-estimation machine learning model according to one embodiment of the present disclosure.

[0038] FIG. 9 is a drawing for explaining the operation of a posture correction system according to one embodiment of the present disclosure.

[0039] FIG. 10 is a flowchart for explaining the operation of a posture correction system according to one embodiment of the present disclosure.

[0040] FIG. 11 is a drawing showing a screen of a user terminal according to one embodiment of the present disclosure.

[0041] FIG. 12 is a drawing for explaining a daily posture object according to one embodiment of the present disclosure.

[0042] FIG. 13 is a drawing for explaining the operation of a user terminal according to one embodiment of the present disclosure.

[0043] FIG. 14 is a drawing for explaining the operation of a user terminal according to one embodiment of the present disclosure.

[0044] FIG. 15 is a drawing for explaining the operation of a user terminal according to one embodiment of the present disclosure.

[0045] Fig. 16 is a flowchart showing the operation of a posture correction system according to one embodiment of the present disclosure.

[0046] FIG. 17 is a drawing showing a screen of a user terminal according to one embodiment of the present disclosure.

[0047] FIG. 18 is a drawing for explaining a daily posture object according to one embodiment of the present disclosure.

[0048] FIG. 19 is a drawing for explaining the operation of a posture correction system according to one embodiment of the present disclosure.

[0049] FIG. 20 is a drawing for explaining the operation of a user terminal according to one embodiment of the present disclosure.

[0050]

[0051] The advantages and features of the disclosed embodiments, and the methods for achieving them, will become clearer with reference to the embodiments described below, along with the accompanying drawings. However, the present disclosure is not limited to the embodiments disclosed below and may be implemented in various different forms. These embodiments are provided solely to ensure the completeness of the disclosure and to fully inform those skilled in the art of the present disclosure of the scope of the invention.

[0052] The terms used in this specification will be briefly explained, and the disclosed embodiments will be described in detail.

[0053] The terms used in this specification have been selected from widely used, current terms, taking into account the functions of the present disclosure. However, these terms may vary depending on the intentions of engineers working in the relevant fields, precedents, the emergence of new technologies, etc. Furthermore, in certain cases, terms may be arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the relevant description of the invention. Therefore, the terms used in this disclosure should not be defined simply as names, but rather based on the meanings of the terms and the overall content of the present disclosure.

[0054] In this specification, singular expressions include plural expressions unless the context clearly indicates that they are singular. In addition, plural expressions include singular expressions unless the context clearly indicates that they are plural.

[0055] When a part of a specification 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 stated.

[0056] Also, the term "part" used in the specification means a software or hardware component, and the "part" performs certain functions. However, the "part" is not limited to software or hardware. The "part" may be configured to reside on an addressable storage medium and may be configured to execute one or more processors. Thus, by way of example, the "part" includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functionality provided within the components and "parts" may be combined into a smaller number of components and "parts" or further separated into additional components and "parts."

[0057] According to one embodiment of the present disclosure, a "unit" may be implemented as a processor and memory. The term "processor" should be broadly interpreted to include a general-purpose processor, a central processing unit (CPU), a microprocessor, a digital signal processor (DSP), a controller, a microcontroller, a state machine, and the like. In some circumstances, a "processor" may also refer to an application-specific integrated circuit (ASIC), a programmable logic device (PLD), a field-programmable gate array (FPGA), and the like. The term "processor" may also refer to a combination of processing devices, such as, for example, a combination of a DSP and a microprocessor, a combination of a plurality of microprocessors, a combination of one or more microprocessors in conjunction with a DSP core, or any other such combination of configurations.

[0058] The term "memory" should be interpreted broadly to include any electronic component capable of storing electronic information. The term memory may also refer to various types of processor-readable media, such as random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable PROM (EEPROM), flash memory, magnetic or optical data storage, registers, etc. A memory is said to be in electronic communication with a processor if the processor can read information from and / or write information to the memory. Memory integrated in a processor is in electronic communication with the processor.

[0059] Below, with reference to the attached drawings, a detailed description of the embodiments is provided so that those skilled in the art can easily implement the present disclosure. Furthermore, in order to clearly illustrate the present disclosure, portions irrelevant to the description are omitted from the drawings.

[0060] FIG. 1 is a drawing showing a user terminal according to one embodiment of the present disclosure.

[0061] The user terminal (100) may include a processor (110) and a memory (120). The processor (110) may execute commands stored in the memory (120). Similarly to the user terminal (100), at least one of the server (210) and the attitude sensor (220) may include a processor and a memory. At least some of the operations performed in the user terminal (100) described below may be performed in at least one of the server (210) and the attitude sensor (220). Conversely, at least some of the operations performed in the server (210) and the attitude sensor (220) may be performed in the user terminal (100).

[0062] FIG. 2 is a drawing for explaining the configuration of a posture correction system according to one embodiment of the present disclosure.

[0063] The posture correction system (200) may include at least one of a server (210), a user terminal (100), and a posture sensor (220). The posture correction system (200) may correct posture using artificial intelligence. More specifically, the posture correction system (200) may accurately measure a user's posture using artificial intelligence. Furthermore, the posture correction system (200) may suggest a correct posture based on the accurately measured user's posture. Furthermore, the posture correction system (200) may also predict diseases that may occur if the user continues to maintain a current posture.

[0064] The user terminal (100) is a device used by a user. The user terminal (100) may include at least one of a PC, a smartphone, a tablet, or a smartwatch. The user terminal (100) may be a device that a user can carry around, but is not limited thereto. The user terminal (100) may include at least one of an input unit, an output unit, a communication unit, and a control unit. The input unit may be a component for a user to input information. The input unit may include at least one of a keyboard, a mouse, and a touchscreen. The output unit may be a component for outputting information through sound or light. The output unit may include at least one of a touchscreen, a monitor, and a speaker. The control unit may be a component that controls the components of the user terminal (100). The communication unit may be a component for the user terminal (100) to communicate with at least one of a server (210) and a posture sensor (220). In the present disclosure, the posture sensor (220) communicates with the user terminal (100), and the user terminal (100) can communicate with the server (210). The server (210) and the posture sensor (220) can exchange information through the user terminal (100). However, this is not limited to this, and the server (210) can also communicate directly with the posture sensor (220).

[0065] The server (210) may be a device located remotely from the user. The server (210) may be a remotely fixed device. That is, the server (210) may not be a device that the user can carry around. The server (210) may store information received from multiple user terminals. For example, the server (210) may store at least one of daily posture information and 3D posture information. The server (210) may process information received from the user terminal (100) and transmit result information to the user terminal (100). At least one of the operations performed by the server (210) in the present disclosure may be performed by the user terminal (100). In addition, at least one of the operations performed by the user terminal (100) in the present disclosure may be performed by the server (210).

[0066] The posture sensor (220) may be a device for measuring a user's posture. The posture sensor (220) may obtain posture information. The posture sensor (220) may include at least one of a daily posture sensor and an exercise posture sensor. The posture information may include at least one of daily posture information and three-dimensional posture information. The daily posture sensor may obtain daily posture information, and the exercise posture sensor may obtain three-dimensional posture information.

[0067] According to one embodiment of the present disclosure, a daily posture sensor can obtain daily posture information by measuring the user's inclination. The daily posture information may include direction information and size information of the user's inclination. The direction information may indicate the direction in which the user leans. For example, the direction information may indicate the direction in which the user's spine leans with respect to the ground. Additionally, the size information may indicate the magnitude of the user's inclination. For example, the size information may indicate the magnitude of the inclination of the user's spine with respect to the ground. The user's inclination, which is the daily posture information, may be at least one of the inclination of the user's upper body or the inclination of the user's spine. In the present disclosure, the inclination of the user's upper body may indicate the inclination of the user's spine. The user's inclination may be at least one of the inclination of the spine with respect to the user's lower body (the lengthwise direction of the thigh) or the inclination of the spine with respect to the ground. The direction information of the inclination included in the daily posture information may include forward, backward, left, and right. Direction information included in the daily posture information may include at least one of forward, backward, left, right, left-front, left-back, right-front, and right-back. However, the present invention is not limited thereto, and the direction information may represent the direction in which the user's spine is tilted as one of values ​​greater than or equal to 0 degrees and less than 360 degrees. Size information included in the daily posture information may represent the size of the angle formed by the spine and an axis perpendicular to the ground. That is, the size information included in the daily posture information may have a smaller value as the user's spine is more perpendicular to the ground and a larger value as the spine is more parallel to the ground. In addition, the size information included in the daily posture information may have a smaller value as the angle of the spine with respect to the user's thigh is 180 degrees and a larger value as the angle is less than 180 degrees.

[0068] According to one embodiment of the present disclosure, a motion posture sensor can acquire posture image information. At least one of a user terminal (100) and a server (210) can acquire three-dimensional posture information indicating the posture of the user's skeleton based on the posture image information. The motion posture sensor can include at least one of a camera, a depth camera, and an infrared sensor.

[0069] 3D posture information based on a motion posture sensor may be information including the angle of at least one joint included in the user's body. For example, the 3D posture information may include at least one of an elbow angle, a knee angle, a waist angle, a neck angle, an arm angle relative to the torso, a thighbone angle relative to the torso, a wrist angle, and an ankle angle. However, the 3D posture information is not limited thereto, and the 3D posture information may also represent a set of coordinates of the user's body feature nodes. One body feature node may correspond to one of the head, neck, left shoulder, right shoulder, elbow, hand, pelvis, knee, and foot. A set of multiple body feature nodes included in the user may be the 3D posture information. A body feature node may include identification information and coordinate values. For example, the identification information of a body feature node may represent one of the head, neck, left shoulder, right shoulder, elbow, hand, pelvis, knee, and foot. A connection relationship between multiple body feature nodes may be determined based on the identification information of the body feature nodes. For example, the head can be connected to the neck, the shoulder can be connected to the elbow, and the elbow can be connected to the hand. The coordinate values ​​of the body feature node can represent the three-dimensional coordinate values ​​of the body feature node in virtual space. While the posture image information from the motion posture sensor expresses the user's two-dimensional posture, three-dimensional posture information adds depth information to the two-dimensional information to represent the user's posture in three-dimensional space.

[0070] The posture sensor (220) may be a device independent of the user terminal (100), but is not limited thereto, and may be included in the user terminal (100). The posture sensor (220) may be included in a wearable device. For example, the posture sensor (220) may be included in a smart seat. In the case of a smart seat, a pressure sensor may be included that can measure which part of the body is subjected to more pressure when the user sits on the chair. If a pressure value is measured from the pressure sensor at the front, the user's posture may be considered to be leaning forward.

[0071] The posture sensor (220) according to the present disclosure may include a sensor for sensing body movements or postures of users in a three-dimensional space. The posture sensor (220) may include a three-axis gyro sensor for measuring a 360-degree direction. The posture sensor (220) may include a three-axis acceleration sensor, a three-axis gyro sensor, or a three-axis geomagnetic sensor. The posture sensor (220) may include an inertial measurement unit (IMU). Through this, the posture sensor (220) may be able to measure a user's posture angle, posture angular velocity, posture angular acceleration, acceleration, etc. The user's posture angle, posture angular velocity, posture angular acceleration, or acceleration, etc. may be included in one of the user's posture information or daily posture information.

[0072] In the embodiment of the present invention, the posture sensor (220) can measure the front-back and left-right directions through a 3-axis gyro sensor, a 3-axis acceleration sensor, or a 3-axis geomagnetic sensor, and can measure the movement in the 360-degree direction around the user, that is, the body's daily posture information. Since various types of sensors can be used for measuring the user's posture area in the front-back, left-right, and 360-degree directions, the embodiment of the present invention will not be particularly limited to any one type. In the embodiment of the present invention, the sensing data acquired through the posture sensor (220) is not limited to the posture information of the front-back inclination, and posture information for the left-right and 360-degree directions must be acquired in order to recommend an appropriate posture correction exercise, so posture data related to the 360-degree direction may be absolutely required for the exercise recommendation service according to the embodiment of the present invention.

[0073] Below, the operation of the posture correction system (200) is described.

[0074] FIG. 3 is a flowchart showing the operation of a posture correction system according to one embodiment of the present disclosure.

[0075] The user terminal (100) may perform a step (310) of setting reference range information of a correct posture based on a user's input. The reference range information may be information for determining whether the user's daily posture information is a correct posture or a bad posture. The reference range information may be determined to be 0 degrees or more and x degrees or less. x may be a predetermined value or a value set by the user. When the spine is perpendicular to the ground, the inclination of the user's spine may indicate 0 degrees. In addition, a posture in which the spine is perpendicular to the ground may indicate a correct posture.

[0076] If the user's daily posture information is included in the reference range information, the posture correction system (200) can determine the user's posture as correct. If the user's daily posture information is not included in the reference range information, the posture correction system (200) can determine the user's posture as bad.

[0077] In order to explain the step (310) of setting the reference range information, a description will be given with reference to FIGS. 4 to 6.

[0078] Figure 4 illustrates numerical values ​​related to automatic range setting and manual range setting according to one embodiment of the present disclosure. Figure 5 is a flowchart illustrating the operation of a posture correction system according to one embodiment of the present disclosure. Figure 6 illustrates a user interface related to range setting displayed on a user terminal according to one embodiment of the present disclosure.

[0079] The step (310) of setting the reference range information may include the following steps.

[0080] Referring to FIGS. 4 and 5, the user terminal (100) may perform a step (510) of receiving user input related to reference range information from a user. The user terminal (100) may display a user interface for setting reference range information when a program related to a posture correction system is first installed in the user terminal (100). The user may input user input indicating whether to automatically or manually set the reference range information into the user terminal (100).

[0081] If the user's input indicates automatic range setting, the user terminal (100) may perform a step (520) of determining predetermined initial range information as reference range information. The initial range information may be information previously stored in the user terminal (100). The initial range information may be set for the front, back, left, and right of the user. However, the present invention is not limited thereto, and the initial range information may be set for at least one of the front, back, left, right, left-front, left-back, right-front, and right-back of the user.

[0082] For example, referring to FIG. 4, the initial range information may be 20 degrees forward, 15 degrees backward, 10 degrees to the left, and 10 degrees to the right. The posture correction system (200) may determine the initial range information as reference range information. If the reference range information is 20 degrees forward, it may mean that leaning forward by 0 degrees or more and 20 degrees or less is a correct posture. Since 0 degrees means that the user's spine is perpendicular to the ground, which is naturally included in the range of a correct posture, the reference range information may be displayed with only the maximum value of the inclination. Similarly, if the reference range information is 10 degrees to the left, it may mean that leaning left by 0 degrees or more and 10 degrees or less is a correct posture.

[0083] The posture correction system (200) can determine whether the user's posture is correct based on reference range information. For example, if the user bends forward less than 20 degrees, the posture correction system (200) can determine that the user's posture is correct. However, if the user bends forward more than 20 degrees, the posture correction system (200) can determine that the user's posture is bad. Automatically determining reference range information in this manner can minimize user input, thereby increasing user convenience.

[0084] If the user's input indicates manual range setting, the user terminal (100) may perform a step (530) of displaying a user interface for range setting. The user interface for range setting may be a screen for setting reference range information. Referring to FIG. 4, the user terminal (100) may set the reference range information to a portion of a range from 0 degrees to 30 degrees based on the user interface for range setting. More specifically, the reference range information may be set to a portion of a range from 0 degrees to 30 degrees for the front, the reference range information may be set to a portion of a range from 0 degrees to 30 degrees for the back, the reference range information may be set to a portion of a range from 0 degrees to 30 degrees for the left, and the reference range information may be set to a portion of a range from 0 degrees to 30 degrees for the right. As previously explained, since 0 degrees naturally falls within the range of a proper posture, the reference range information may be expressed only by the maximum inclination value determined to be a proper posture. That is, the reference range information is set to one value between 0 degrees and 30 degrees for the front, one value between 0 degrees and 30 degrees for the back, one value between 0 degrees and 30 degrees for the left, and one value between 0 degrees and 30 degrees for the right.

[0085] A user interface for setting a range according to one embodiment of the present disclosure may be as shown in FIG. 6. For example, referring to (a) of FIG. 6, the user terminal (100) may display a screen for obtaining reference range information for the front. In addition, referring to (b) of FIG. 6, the user terminal (100) may display a screen for obtaining reference range information for the rear. Referring to (c) of FIG. 6, the user terminal (100) may display a screen for obtaining reference range information for the left. Referring to (d) of FIG. 6, the user terminal (100) may display a screen for obtaining reference range information for the right.

[0086] When a user interface for setting a range is displayed as in FIG. 6, the user can set reference range information based on the guidance of the user terminal (100). More specifically, the user terminal (100) can perform a step (540) of acquiring the inclination of the user's spine based on the daily posture sensor. The user terminal (100) can display a message indicating to the user to incline the user's spine at a desired angle. The user can incline the spine to a range considered to be a normal incline while wearing the daily posture sensor. The daily posture sensor can measure the inclination of the user's spine. The inclination of the spine can indicate at least one of the inclination of the spine with respect to the ground or the inclination of the spine with respect to the legs. The user terminal (100) can acquire the inclination of the spine with respect to the front, top, left, and right, respectively.

[0087] Although it is described that the inclination of the spine is acquired using a daily posture sensor in step (540), it is not limited thereto. In step (540), the user terminal (100) may also acquire the inclination of the spine using a motion posture sensor.

[0088] Referring to FIG. 5, the user terminal (100) may perform a step (550) of changing and outputting the inclination of a body object (610) representing at least a part of a human body included in a user interface based on the inclination of the spine. The body object (610) may be an object representing the user's spine. However, it is not limited thereto, and the body object (610) may represent the entire body of the user. The body object (610) may be an illustration corresponding to the user's body. However, it is not limited thereto, and the body object (610) may be an image of the user captured by a camera. The user terminal (100) may incline the inclination of the body object (610) based on the inclination of the spine acquired in step (540). Therefore, the user can visually check the inclination of his / her body. The user can determine whether the inclination of his / her body is normal. The user terminal (100) may provide an interface for accurately determining reference range information, thereby guiding the user to maintain a correct posture.

[0089] Referring to FIG. 6, the user terminal (100) may output a numerical value (620) of the user's spinal inclination. By outputting the numerical value (620) of the spinal inclination, the user terminal (100) may enable the user to know what value the reference range information will be set to.

[0090] The user terminal (100) may include an object (625) representing a predetermined threshold slope. The object (625) representing the threshold slope may have an arc shape. For example, the threshold slope may be 30 degrees. The threshold slope may be the maximum value of the reference range information. The reference range information may be determined to be less than or equal to the threshold slope value.

[0091] Referring to FIG. 5, a user may maintain a desired inclination of the spine. At this time, if the user confirms at least one of the inclination of the body object (610) or the numerical value (620) and inputs an object (630) indicating confirmation, the user terminal (100) may perform a step (560) of determining a range from 0 degrees to the inclination of the spine as reference range information. The object (630) indicating confirmation may be a button displayed on a display. However, it is not limited thereto, and the object (630) indicating confirmation may be implemented as a physical button. If the user terminal (100) receives an input for the object (630), the user terminal (100) may determine a range from 0 degrees to the inclination of the spine as reference range information. The inclination of the spine may be the inclination of the spine that the user is maintaining.

[0092] In this way, since the user determines the reference range information, which is the range of correct posture, by directly taking a posture, the posture correction system of the present disclosure does not require the user to correct the posture while maintaining an unnecessarily uncomfortable posture. In addition, since the user sets the posture by directly taking a posture, there is little trial and error, so the user terminal (100) can quickly determine the reference range information. Correct posture may vary from person to person. The posture correction system (200) of the present disclosure can recommend a correct posture to the user by reflecting the correct posture that varies from person to person.

[0093] Referring again to FIG. 3, the user terminal (100) can perform a step (320) of acquiring the user's daily posture information based on a daily posture sensor included in the posture sensor and measuring the inclination of the user's spine.

[0094] A daily posture sensor can obtain daily posture information by measuring the user's inclination. The daily posture sensor may include at least one of a pressure sensor and an inertial sensor for measuring the user's inclination. The daily posture sensor may be included in a wearable device that the user can wear. The daily posture sensor may be positioned in close contact with the user's chest or back. The daily posture sensor may also be included in a chair or a smart seat. The daily posture sensor can measure the user's inclination, which is daily posture information. The user's inclination may be the user's inclination with respect to the ground, or the spine's inclination with respect to the lower body. Furthermore, the user's inclination may be the cervical spine's inclination with respect to the ground, or the cervical spine's inclination with respect to the spine. For convenience of explanation in the present disclosure, the daily posture information measured by the daily posture sensor is expressed as the spine's inclination, but is not limited thereto.

[0095] If the daily posture information is not included in the reference range information, the user terminal (100) can perform step (330) of outputting a message indicating that the posture should be corrected. For example, if the reference range information is 0 degrees or more and 15 degrees or less forward, and the daily posture information is 25 degrees forward, the daily posture information is not included in the reference range information. In this case, the user terminal (100) can determine the user's posture as bad. In addition, the user terminal (100) can output a message indicating that the posture should be corrected. The user can correct the posture based on the message. The posture correction system (200) of the present disclosure can output a message whenever the user does not correct his / her posture, thereby guiding the user to maintain the correct posture. Accordingly, the user's posture can be gradually corrected.

[0096] Furthermore, the posture correction system (200) of the present disclosure acquires the user's relatively simple inclination as daily posture information, eliminating the need for complex and expensive sensors. Furthermore, the posture correction system (200) of the present disclosure simply compares values ​​to recommend a proper posture, eliminating the need for high computer processing power. Therefore, it can also output messages for quick posture correction. In other words, since there is no delay, the user can immediately recognize that they are in a bad posture and correct their posture.

[0097] The above describes the process by which the posture correction system (200) acquires daily posture information using the daily posture sensor and guides the user to maintain a correct posture. In this way, when the posture correction system (200) utilizes the daily posture information from the daily posture sensor, it can provide quick feedback to the user, but it has the disadvantage of making it difficult to know the user's exact posture. In particular, when the posture correction system (200) recommends an exercise to correct the user's posture, it is difficult to confirm whether the user is accurately following the recommended exercise. The exercise recommended by the posture correction system (200) is to correct the user's posture, but if the user does not adopt the exact posture instructed in the exercise, the user's posture may actually worsen. To this end, the posture correction system may further utilize an exercise posture sensor.

[0098] First, the process of recommending exercise to the user through the user terminal (100) by the posture correction system (200) will be described. This process can be performed between steps (330) and (340).

[0099] FIG. 7 is a flowchart showing the operation of a posture correction system according to one embodiment of the present disclosure.

[0100] The user terminal (100) may perform a step (710) of selecting exercise information beneficial to the user from among a plurality of candidate exercise information based on daily posture information. The user terminal (100) may accumulate and store daily posture information to obtain accumulated daily posture information. The accumulated daily posture information may be information accumulated over a predetermined collection period. The predetermined collection period may be one day, one week, or one month.

[0101] Multiple candidate exercise information can identify exercises that can be recommended to the user. The multiple candidate exercise information may be exercises that can improve the user's posture. Depending on the user's preferred posture, at least one of the multiple candidate exercise information can be recommended to the user.

[0102] The user terminal (100) can obtain the posture that the user mainly takes based on the accumulated daily posture details. For example, the user terminal (100) can obtain in which direction the user has leaned the longest among forward, backward, left, and right. The user terminal (100) may store a movement information selection table. The movement information selection table may be used to select movement information. The movement information selection table may be a table that corresponds to movement information for the direction in which the user has leaned the longest among the postures taken. For example, if the user has leaned forward the longest, the user terminal (100) can select the movement information corresponding to the time when the user has leaned the most forward from the movement information selection table.

[0103] According to various embodiments of the present disclosure, the user terminal (100) can obtain an ordered tilt direction by chronologically sorting the time for which it was tilted forward, backward, left, or right. For example, the user may have been tilted forward for the most time, then left, then right, and then backward. In this case, the user terminal (100) can obtain the ordered tilt direction as forward-left-right-backward. The user terminal (100) may store a motion information selection table that corresponds to the ordered tilt direction and motion information. The user terminal (100) can select motion information corresponding to the ordered tilt direction from the motion information selection table.

[0104] According to various embodiments of the present disclosure, the user terminal (100) can obtain bad posture time by direction based on accumulated daily posture details. The user terminal (100) can also select exercise information based on the direction in which the bad posture time is the longest. More specifically, the user terminal (100) can select exercise information corresponding to the direction in which the bad posture time is the longest from the exercise information selection table. For example, if the direction in which the bad posture time is the longest is forward, the user terminal (100) can select exercise information corresponding to the forward direction. Bad posture time will be described in detail later.

[0105] According to various embodiments of the present disclosure, the user terminal (100) can obtain sorted tilt directions by sorting directions in order of increasing bad posture time. For example, if the directions are sorted in order of increasing bad posture time, it may be forward-left-right-backward. The user terminal (100) can select exercise information corresponding to the sorted tilt directions from the exercise information selection table.

[0106] The step (710) of selecting exercise information may further include the following steps. The user terminal (100) may perform a step of obtaining disease information from the user. The disease information may include a disease code. The user terminal (100) may select one from a list of multiple predetermined disease information items based on the user's input.

[0107] The user terminal (100) may perform a step of selecting some candidate exercise information by excluding exercise information harmful to the user from a plurality of candidate exercise information based on disease information. The user terminal (100) may pre-store a harmful exercise table. The harmful exercise table may be a table that corresponds disease information to harmful exercise information. The user terminal (100) may select harmful exercise information corresponding to the disease information from the harmful exercise table. The harmful exercise information may be information related to an exercise that suggests a posture that the user cannot assume or an exercise that worsens the user's disease. The user terminal (100) may obtain some candidate exercise information by excluding harmful exercise information from the plurality of candidate exercise information. The some candidate exercise information may be a subset of the plurality of candidate exercise information.

[0108] The user terminal (100) may perform a step of selecting exercise information based on daily posture information from among some candidate exercise information. Unlike step (710), the user terminal (100) may select exercise information from some candidate exercise information rather than selecting exercise information from a plurality of candidate exercise information. In this way, the user terminal (100) does not recommend exercise information that is harmful to the user, so the posture correction system (200) may not worsen the user's illness. In addition, the posture correction system (200) does not provide the user with exercise information that includes postures that the user cannot assume, so that only exercises that are effective in improving the user's posture can be provided to the user.

[0109] The user terminal (100) can perform a step (720) of outputting exercise information. The user terminal (100) can output identification information of the exercise information. In addition, the user terminal (100) can output a document describing the posture to be taken by the user included in the exercise information, a still image or a video that lists the postures to be taken by the user in chronological order. The user can assume a posture based on the still image or video output by the user terminal (100). By the user assuming a posture corresponding to the exercise information, the user's posture can be improved.

[0110] When a user assumes a posture based on movement information, the user terminal (100) may perform a step (730) of acquiring the user's posture image information based on a movement posture sensor. The step (730) of acquiring the posture image information may correspond to step (340).

[0111] The process of FIG. 7 may be performed between steps (330) and (340). However, the steps performed between steps (330) and (340) are not limited to those of FIG. 7, and other steps may be performed. Furthermore, the steps of FIG. 7 may not be performed between steps (330) and (340). Refer back to FIG. 3 to explain step (340).

[0112] The server (210) or the user terminal (100) may perform a step (340) of acquiring posture image information of the user based on a motion posture sensor included in the posture sensor and capturing the entire body of the user. The motion posture sensor may include at least one of a camera, a depth camera, and an infrared sensor. The posture image information may be a two-dimensional image having color information as a pixel value. The color information may be a value according to the wavelength of visible light. However, the present invention is not limited thereto, and the posture image information may also be an image having a depth value as a pixel value. The depth value may represent the distance between the motion posture sensor and the subject of the photograph.

[0113] The server (210) may perform a step (350) of obtaining 3D posture information by applying at least one of posture image information and daily posture information to a posture estimation machine learning model. The posture estimation machine learning model may be a model for outputting 3D posture information based on at least one of posture image information and daily posture information.

[0114] Refer to Fig. 8 to explain the posture estimation machine learning model (820).

[0115] FIG. 8 is a diagram illustrating a self-estimation machine learning model according to one embodiment of the present disclosure.

[0116] The pose estimation machine learning model may be a model generated according to "DRPose3D: Depth Ranking in 3D Human Pose Estimation". The pose estimation machine learning model (820) may be a machine learning model for predicting 3D pose information from pose image information. The pose estimation machine learning model (820) may be implemented based on a machine learning model such as CNN. The pose estimation machine learning model (820) may use past pose image information (811) and past 3D pose information (812) as learning data. The past pose image information (811) may be an analysis target, and the past 3D pose information (812) may be label information acquired by a person or a device. The past 3D pose information (812) may be a ground truth value. The past pose image information (811) and the past 3D pose information (812) may correspond one-to-one. The posture information of a person in a three-dimensional space that appears in the past posture image information (811) may be past three-dimensional posture information (812).

[0117] The server (210) can update the posture estimation machine learning model (820) through forward propagation and back propagation. The server (210) can machine learn the correlation between past 3D posture information (812) and past posture image information (811). When the accuracy of the posture estimation machine learning model (820) is sufficiently high, the server (210) can stop updating the posture estimation machine learning model (820). The server (210) can transmit the posture estimation machine learning model (820) to the user terminal (100) or another server. The server (210) can store the posture estimation machine learning model (820).

[0118] According to FIG. 8, the server (210) uses past posture image information (811) and past 3D posture information (812) to generate a posture estimation machine learning model (820). However, the present invention is not limited thereto. The server (210) can generate a posture estimation machine learning model (820) based on past posture image information (811), past 3D posture information (812), and past daily posture information. The server (210) can machine learn the correlation between past 3D posture information (812) and past daily posture information.

[0119] The server (210) can receive posture image information (831) from the user terminal (100) or the posture sensor (220). The posture image information (831) may be information acquired by a motion posture sensor included in the posture sensor (220). The process of acquiring the posture image information has been described in step (340). The server (210) can perform step (350) of acquiring 3D posture information (840) by applying at least one of the posture image information (831) and the daily posture information to a posture estimation machine learning model (820). The 3D posture information (840) acquired by applying it to the posture estimation machine learning model (820) may be predicted information. The 3D posture information (840) predicted by the posture estimation machine learning model (820) may be very similar to actual 3D posture information.

[0120] According to FIG. 8, the posture estimation machine learning model (820) can determine predicted 3D posture information (840) by only receiving posture image information (831). However, it is not limited thereto. The posture estimation machine learning model (820) can further receive daily posture information to determine predicted 3D posture information (840). In this case, the server (210) can utilize the posture estimation machine learning model (820) that machine-learns the correlation between past posture image information (811) and past daily posture information regarding past 3D posture information (812).

[0121] Predicted 3D posture information (840) and past 3D posture information (812) may be included in the 3D posture information. While posture image information represents the user's posture on a 2D plane, 3D posture information may include information for expressing the user's posture in 3D space. In other words, 3D posture information may include depth information.

[0122] According to one embodiment of the present disclosure, 3D posture information may include an angle of at least one predetermined joint included in the user's body. The 3D posture information may include identification information corresponding to a body feature node. That is, the 3D posture information may include identification information and angles of joints corresponding to the identification information. Here, at least some of the body feature nodes may correspond to joints. Alternatively, the body feature node may include at least one joint. The at least one predetermined joint may include at least one of the neck, elbow, knee, waist, shoulder, pelvis, wrist, and ankle. Furthermore, the 3D posture information may include at least one of the elbow angle, knee angle, waist angle, neck angle, arm angle relative to the torso (shoulder angle), thigh bone angle relative to the torso (pelvic angle), wrist angle, and ankle angle. Relationships between multiple joints may be predetermined. For example, the neck may be connected to the shoulder by a skeletal edge corresponding to the collarbone, and the shoulder may be connected to the elbow by a skeletal edge corresponding to the arm bone. The distance between multiple joints can be estimated by a posture estimation machine learning model (820) or can be predetermined. In this way, when the angles of multiple joints are determined, the relationships between the multiple joints are predetermined, and the distances between the multiple joints are determined, the posture correction system can generate a skeletal model representing the overall shape of the user's body. The skeletal model can be expressed in three-dimensional space.

[0123] According to various embodiments of the present disclosure, 3D posture information may include a set of coordinates of body feature nodes of the user. One body feature node may correspond to a point on the body. The point on the body may include at least one of the head, neck, left shoulder, right shoulder, elbow, hand, pelvis, knee, and foot. That is, one body feature node may correspond to one of the head, neck, left shoulder, right shoulder, elbow, hand, pelvis, knee, and foot. A set of multiple body feature nodes included in the user may be 3D posture information. A body feature node may include identification information and coordinate values. For example, the identification information of a body feature node may indicate one of the head, neck, left shoulder, right shoulder, elbow, hand, pelvis, knee, and foot. A connection relationship between the multiple body feature nodes may be determined based on the identification information of the body feature node. For example, the head may be connected to the neck, the shoulder may be connected to the elbow, and the elbow may be connected to the hand. The coordinate values ​​included in the body feature node may be three-dimensional coordinate values ​​indicating the location of the body feature node in real space or virtual space. When the coordinates of the body feature node are determined in this way and the connection relationship between the body feature nodes is predetermined based on the identification information, the posture correction system can generate a skeletal model representing the overall shape of the user's body. At least one of the server (210) and the user terminal (100) can display the skeletal model. The skeletal model can correspond to three-dimensional posture information.

[0124] FIG. 9 is a drawing for explaining the operation of a posture correction system according to one embodiment of the present disclosure.

[0125] 3D posture information may include multiple body feature nodes, at least some of which correspond to joints constituting the body, and skeletal edges connecting the multiple body feature nodes. Some of the body feature nodes may correspond to joints, while others may correspond to non-joints, such as hands or feet. The corresponding body parts may be implemented in various ways depending on the design of the posture correction system (200).

[0126] FIG. 9 illustrates a skeletal model (910). The posture correction system (200) can generate a skeletal model (910) corresponding to the three-dimensional posture information acquired by the posture estimation machine learning model (820). The skeletal model (910) can be displayed in three-dimensional space. For example, the three-dimensional posture information can include a first body feature node (911) corresponding to the neck and a second body feature node (912) corresponding to the center of the pelvis. The first body feature node (911) and the second body feature node (912) can include identification information and coordinate information, respectively. However, the present invention is not limited thereto. In addition, the three-dimensional posture information can include a first skeletal edge (913) connecting the first body feature node (911) and the second body feature node (912). The first skeletal edge (913) may correspond to the spine.

[0127] Additionally, the 3D posture information may include a third body feature node (914) corresponding to the shoulder. Additionally, the 3D posture information may include a second skeletal edge (915) connecting the first body feature node (911) and the third body feature node (914). The second skeletal edge (915) may correspond to the clavicle.

[0128] As described above, the user terminal (100) can perform the step (320) of acquiring daily posture information. The daily posture information (920) acquired by the step (320) can be displayed as a vector in a three-dimensional space. The vector by the daily posture information (920) can indicate the extension direction of the spine. The vector by the daily posture information (920) can be the extension direction of the spine from the pelvis toward the head. However, it is not limited thereto, and the vector direction can be the extension direction of the spine from the head toward the pelvis. The user terminal (100) acquires daily posture information based on the posture sensor (220) including the inertial sensor, and thus can acquire the inclination of the user's spine with respect to the ground. Therefore, the user terminal (100) can display the daily posture information (920) in a three-dimensional space.

[0129] The posture correction system (200) can further perform the following process when performing the step (350) of acquiring 3D posture information.

[0130] At least one of the server (210) and the user terminal (100) may perform a step of matching the angle of the skeletal edge corresponding to the spine included in the 3D posture information with respect to the ground to the daily posture information. More specifically, the skeletal edge corresponding to the spine may be the first skeletal edge (913). The server (210) or the user terminal (100) may match the extension direction of the first skeletal edge (913) with the direction of the daily posture information (920). Therefore, the posture correction system (200) of the present disclosure can accurately determine the inclination of the spine with respect to the ground and accurately determine the angle at which the skeletal model is positioned with respect to gravity. Therefore, the posture correction system (200) can determine whether the user is accurately assuming the posture included in the exercise information recommended to the user.

[0131] The reason why the posture correction system (200) corrects the 3D posture information based on the daily posture information as described above is as follows. The skeletal model (910) corresponding to the 3D posture information is acquired based on a motion posture sensor. The motion posture sensor may not include angular information with respect to the ground. For example, the motion posture sensor may be a camera, and the camera captures the object with the horizontal direction of the camera generally parallel to the ground, but depending on the environment, the horizontal direction of the camera may not be parallel to the ground. In such a case, the 3D posture information may have difficulty including information about the inclination of the spine with respect to the ground. This is because the horizontal direction of the posture image information captured by the camera is not parallel to the direction of the ground. Furthermore, if the ground is not perpendicular to gravity, even if the posture image information is captured with the camera parallel to the ground, the 3D posture information based on the posture image information cannot include any information about gravity.

[0132] However, whether the user is maintaining a correct posture may be one of the important factors in terms of the force exerted on the spine by gravity. This is because, if the extension direction of the spine is tilted at a large angle with the direction of gravity, a large force may be exerted on the spine, which may be judged as not being in a correct posture. Therefore, the fact that the 3D posture information based on the motion posture sensor does not reflect any information about the direction of gravity may be a major problem. Therefore, the posture correction system (200) of the present disclosure can correct the 3D posture information based on the daily posture information.

[0133] Referring to FIG. 9, the posture correction system can obtain a skeletal model (930) by rotating the skeletal model (910). More specifically, the posture correction system can obtain the skeletal model (930) by rotating the skeletal model (910) so that the direction of the first skeletal edge (913) included in the skeletal model (910) matches the direction of the daily posture information (920). For example, the inclination of the spine of the skeletal model (910) may be included in the reference range information and thus may be a correct posture, but the inclination of the spine of the skeletal model (930) may not be included in the reference range information and thus may be a bad posture. Accordingly, the posture correction system (200) may display a message indicating a change in posture. The user can correct their posture by viewing the message. Therefore, the posture correction system (200) can recommend exercises to correct the user's posture more precisely and accurately determine whether the user is taking a posture according to the recommended exercises.

[0134] Referring back to FIG. 3, at least one of the server (210) and the user terminal (100) may perform a step (360) of determining a posture similarity between the 3D posture information and the predetermined reference posture information. The predetermined reference posture information may be information for determining whether the user is accurately assuming a posture based on exercise information. The reference posture information may be acquired based on a posture taken by an exercise expert. The following description will be based on the server (210) performing step (360), but is not limited thereto. The user terminal (100) may also perform step (360).

[0135] As previously described, 3D posture information may include identification information corresponding to a body feature node and angles of joints corresponding to the identification information. Reference posture information may also include identification information corresponding to a body feature node and angles of joints corresponding to the identification information. The server (210) may determine the posture similarity between the angles of joints of 3D posture information having the same identification information and the angles of joints of the reference posture information. For example, the posture similarity may be determined by a difference ratio determined by the following equation.

[0136]

[0137] Difference ratio = |Joint angle of 3D posture information - Joint angle of reference posture information| / Joint angle of reference posture information

[0138]

[0139] The smaller the difference ratio, the higher the posture similarity. The above equation is for a single joint; for all joints, the average of the difference ratios for all joints can be used. However, this is not limited to this, and various similarity determination methods can be used to determine posture similarity.

[0140] If the posture similarity is below a predetermined threshold similarity, the user terminal (100) may perform a step of outputting a message indicating that the user should correct the posture. The user can confirm the message and correct the posture so that it is similar to the posture provided in the exercise information. Therefore, the user can assume an accurate posture, thereby maximizing the effectiveness of the posture correction exercise.

[0141] FIG. 10 is a flowchart illustrating the operation of a posture correction system according to one embodiment of the present disclosure. FIG. 11 is a diagram illustrating a screen of a user terminal according to one embodiment of the present disclosure. FIG. 12 is a diagram illustrating a daily posture object according to one embodiment of the present disclosure.

[0142] The step (320) of obtaining daily posture information may include the steps of FIG. 10.

[0143] Referring to FIG. 10, the user terminal (100) can perform a step (1010) of displaying a daily posture object (1120) including a first area, a second area, a third area, and a fourth area. The daily posture object (1120) can express 360 degrees around the user. 360 degrees around the user can mean a circumference including front, back, left, and right. It should be noted that FIGS. 11 to 15 illustrate some embodiments of the daily posture object, and that daily posture objects of different forms than those of FIGS. 11 to 15 are also possible.

[0144] In this regard, referring to FIG. 11, the user terminal (100) may include an output unit. The output unit may include a touch screen. The user terminal (100) may display a daily posture object display area (1110). A daily posture object (1120) may be displayed in the daily posture object display area (1110). The daily posture object (1120) may be an object that displays statistics based on daily posture information. The daily posture object (1120) may include at least one of a first area (1130), a second area (1140), a third area (1150), and a fourth area (1160).

[0145] The first region (1130) may be located in the first direction in the daily posture object (1120). The first region (1130) may be an region indicating that the user has leaned in the first direction. For example, the first region (1130) may be an region indicating that the user has leaned forward.

[0146] The second area (1140) may be located in the second direction from the daily posture object (1120). The second area (1140) may be an area to indicate that the user has tilted in the second direction. For example, the second area (1140) may be an area to indicate that the user has tilted backward.

[0147] The third area (1150) may be located in the third direction from the daily posture object (1120). The third area (1150) may be an area indicating that the user has tilted in the third direction. For example, the third area (1150) may be an area indicating that the user has tilted to the left.

[0148] The fourth area (1160) may be located in the fourth direction of the daily posture object (1120). The fourth area (1160) may be an area indicating that the user has tilted in the fourth direction. For example, the fourth area (1160) may be an area indicating that the user has tilted to the right.

[0149] In Fig. 11, the daily posture object (1120) is shown to be composed of four fan shapes, but this is not limited thereto. Referring to Fig. 12, as in the daily posture object (1210), the first region (1130), the second region (1140), the third region (1150), and the fourth region (1160) can each be divided into two.

[0150] However, it is not limited to this, and the first area (1130), the second area (1140), the third area (1150), and the fourth area (1160) can each be divided into three, such as the daily posture object (1230). In addition, the first area (1130), the second area (1140), the third area (1150), and the fourth area (1160) can each be divided into two or more.

[0151] Referring to FIGS. 10 and 11, the user terminal (100) may perform a step (1020) of displaying at least one of the first area (1130), the second area (1140), the third area (1150), and the fourth area (1160) in a predetermined color based on the daily posture information. The daily posture information may include direction information and size information of the user's inclination. For example, when the direction information of the user's inclination included in the daily posture information indicates the first direction, the user terminal (100) may display the first area (1130) in a predetermined color. In addition, the larger the size information of the user's inclination included in the daily posture information, the darker or more intense the predetermined color may become.

[0152] As described above, the first region (1130), the second region (1140), the third region (1150), and the fourth region (1160) can each be divided into two or more regions. For example, the first region (1130) can be divided into the 1-1 region (1221) and the 1-2 region (1222). The posture correction system (200) can display a region further from the center of the circle as the angle at which the user's spine is tilted increases. For example, when the angle at which the user's spine is tilted in the first direction is large, the 1-2 region (1222) can be displayed in a dark or opaque color. A large angle at which the user's spine is tilted in the first direction can mean that the angle at which the spine is tilted in the first direction is not included in the reference range information. Additionally, if the user's spine has a small tilt angle in the first direction, the 1-1 area (1221) may be displayed in a dark or opaque color. A small tilt angle of the user's spine in the first direction may mean that the tilt angle in the first direction is included in the reference range information.

[0153] According to various embodiments of the present disclosure, the step (320) of obtaining daily posture information may include the following steps. The user terminal (100) may display an object (1240) viewed from above the user. The user terminal (100) may perform a step of displaying a daily posture object (1210) including at least one concentric circle around the object (1240) viewed from above the user. The concentric circle may surround 360 degrees of the object (1240). The concentric circle may represent an omnidirectional direction including the front, back, left, and right of the user. The concentric circle may represent an angle of the user's inclination. The outermost concentric circle may be related to a larger inclination of the user. The user's inclination may refer to the inclination of the user's spine. The user's inclination may be the largest when the user's spine is perpendicular to the ground, and the acute angle formed by the spine and the ground may correspond to the user's inclination.

[0154] The user terminal (100) can perform a step of displaying a color corresponding to the size information included in the daily posture information at a location corresponding to the direction information included in the daily posture information in the daily posture object. Unlike FIG. 12, the daily posture object may not be divided into a first direction, a second direction, a third direction, and a fourth direction. The user terminal (100) can measure the inclination for 360 degrees around the user. The direction information of the daily posture information can represent one angle among the 360 ​​degrees around the user. The size information of the daily posture information can represent the inclination angle of the user's spine. To explain this, reference will be made to FIG. 18 for a moment.

[0155] FIG. 18 is a drawing for explaining a daily posture object according to one embodiment of the present disclosure.

[0156] Referring to the daily posture object (1810), the user terminal (100) can determine the location of the item (1811) based on the direction information (1812) included in the daily posture detailed information. In addition, the user terminal (100) can determine in which concentric circle the item (1811) will be located based on the size information included in the daily posture detailed information. If the size information is small, the item (1811) may be located in a concentric circle closer to the center of the circle, and if the size information is large, the item (1811) may be located in a concentric circle farther from the center of the circle. In addition, the user terminal (100) can determine the color of the item (1811) based on either the good posture time or the bad posture time based on the accumulated daily posture detailed information. If the item (1811) is related to the good posture, the user terminal (100) will refer to the good posture time. Additionally, if the item (1811) is related to bad posture, the user terminal (100) will refer to the bad posture time. The daily posture object (1810) displays only one item (1811), but this is not limiting. The user terminal (100) can display at least one item (1811) in the daily posture object (1810).

[0157] Also, in the daily posture object (1810) of FIG. 18, the item (1811) is positioned only on concentric circles, but this is not limited thereto. According to the daily posture object (1820), the item (1821) may be positioned between concentric circles. The distance from the center of the circle to the item (1821) may be directly proportional to the size information of the daily posture information or accumulated posture information. Since the size information is a continuous value, the item (1821) may be positioned between concentric circles.

[0158] Refer to FIGS. 13 to 15 below to explain step (1020) in more detail.

[0159] FIG. 13 is a drawing for explaining the operation of a user terminal according to one embodiment of the present disclosure.

[0160] The step (1020) of displaying in a predetermined color of Fig. 10 may further include the following steps.

[0161] Direction information included in daily posture details can indicate one of the following: forward, backward, left, or right. However, this is not limited to these, and the direction information can have a value greater than or equal to 0 and less than 360. In this case, when the user's front is considered 0 degrees,

[0162] If the direction information is 315 degrees or more but less than 360 degrees, or 0 degrees or more but less than 45 degrees, the user terminal (100) can determine that the user is leaning forward. If the direction information is 45 degrees or more but less than 135 degrees, the user terminal (100) can determine that the user is leaning to the right. If the direction information is 135 degrees or more but less than 225 degrees, the user terminal (100) can determine that the user is leaning backward. If the direction information is 225 degrees or more but less than 315 degrees, the user terminal (100) can determine that the user is leaning to the left.

[0163] Referring to FIG. 13, if the daily posture information indicates that the user is leaning forward, the user terminal (100) may perform a step of displaying the first area (1130) in a predetermined color. Additionally, the greater the user's inclination included in the daily posture information, the darker or more opaque the predetermined color may be displayed.

[0164] Additionally, if the daily posture information indicates that the user is leaning backwards, the user terminal (100) may perform a step of displaying the second area (1140) in a predetermined color. Additionally, the greater the user's inclination included in the daily posture information, the darker or more opaque the predetermined color may be displayed.

[0165] Additionally, if the daily posture information indicates that the user is leaning to the left, the user terminal (100) may perform a step of displaying the third area (1150) in a predetermined color. Furthermore, the greater the user's inclination included in the daily posture information, the darker or more opaque the predetermined color may be displayed.

[0166] Additionally, if the daily posture information indicates that the user is leaning to the right, the user terminal (100) may perform a step of displaying the fourth area (1160) in a predetermined color. Furthermore, the greater the user's inclination included in the daily posture information, the darker or more opaque the predetermined color may be displayed.

[0167] According to various embodiments of the present disclosure, the user terminal may perform a step of displaying at least one of the first area (1130), the second area (1140), the third area (1150), and the fourth area (1160) in a predetermined color based on accumulated daily posture details.

[0168] More specifically, the user terminal (100) can perform a step of acquiring accumulated daily posture information by accumulating daily posture information at a predetermined cycle during a predetermined collection time. The user terminal (100) can perform a step of acquiring accumulated daily posture information by accumulating daily posture information at a predetermined cycle during a predetermined collection time. The predetermined collection time can be one of one day, one week, or one month. The predetermined cycle can be one of one second to one hour. For example, the user terminal (100) can acquire accumulated daily posture information for one day by accumulating and storing daily posture information at one-minute intervals.

[0169] The user terminal (100) can obtain bad posture time from the accumulated daily posture detailed information. The bad posture time can include at least one of forward bad posture time, backward bad posture time, left bad posture time, and right bad posture time. The bad posture time can indicate the time during which the user's inclination is maintained at an angle not included in the reference range information. The bad posture time can indicate the time during which the user maintains an unhealthy posture. As described above, the accumulated daily posture detailed information is obtained at a predetermined cycle during a predetermined collection time, and if the user's inclination measured at one predetermined cycle is not included in the reference range information, the user terminal (100) can determine that the user maintained a bad posture during the corresponding cycle. The user terminal (100) can accumulate and measure the time during which the user maintained a bad posture during the predetermined collection time. The user terminal (100) can accumulate and measure the time during which the user maintained a bad posture for each direction information. The user terminal (100) can obtain the time the user maintains a bad posture while leaning forward as the forward bad posture time. The user terminal (100) can obtain the time the user maintains a bad posture while leaning backward as the back bad posture time. The user terminal (100) can obtain the time the user maintains a bad posture while leaning to the left as the left bad posture time. The user terminal (100) can obtain the time the user maintains a bad posture while leaning to the right as the right bad posture time.

[0170] The user terminal (100) can perform a step of acquiring a time when the user has daily posture information indicating a forward lean, which is not included in the reference range information in the accumulated daily posture detailed information, as a forward bad posture time. The user terminal (100) can perform a step of acquiring a time when the user has daily posture information indicating a backward lean, which is not included in the reference range information in the accumulated daily posture detailed information, as a backward bad posture time. The user terminal (100) can perform a step of acquiring a time when the user has daily posture information indicating a left lean, which is not included in the reference range information in the accumulated daily posture detailed information, as a left bad posture time. The user terminal (100) can perform a step of acquiring a time when the user has daily posture information indicating a right lean, which is not included in the reference range information in the accumulated daily posture detailed information, as a right bad posture time.

[0171] Referring to FIG. 13, the user terminal (100) can perform a step of displaying the first area (1130) darker, darker, or more opaque as the forward bad posture time increases. In addition, the user terminal (100) can perform a step of displaying the second area (1140) darker, darker, or more opaque as the rear bad posture time increases. In addition, the user terminal (100) can perform a step of displaying the third area (1150) darker, darker, or more opaque as the left bad posture time increases. In addition, the user terminal (100) can perform a step of displaying the fourth area (1160) darker, darker, or more opaque as the right bad posture time increases. In the above, it is described that the user terminal (100) displays the areas (1130, 1140, 1150, 1160) darker, thicker, or opaque as the bad posture time increases, but it is not limited thereto. The user terminal (100) may display the areas (1130, 1140, 1150, 1160) brighter, blurrier, or more transparent as the bad posture time increases. In addition, the user terminal (100) may display the bad posture time as numbers in the areas (1130, 1140, 1150, 1160).

[0172] In this way, the posture correction system (200) of the present disclosure shows the bad posture time by direction, so that the user can know in which direction the user mainly maintains the posture leaning. For example, the user may feel that he or she is sitting comfortably on a chair, but in reality, he or she may maintain a posture leaning in a certain direction, which continuously applies force to the user's spine, causing a disc problem in the user's spine. The posture correction system (200) of the present disclosure can analyze the user's posture for an extended period of time based on accumulated daily posture information. In addition, the user's posture can be visually displayed. Through this, the user can identify a bad posture that he or she was not aware of and make efforts to correct it.

[0173] According to various embodiments of the present disclosure, the user terminal (100) can determine the color of the first area (1130) to the fourth area (1160) according to the time (frequency) corresponding to the direction of the user's tilt based on the accumulated daily posture detailed information. For example, the longer the time (frequency) indicating that the user has leaned forward in the accumulated daily posture detailed information, the darker or more opaque the first area (1130) can be displayed. In addition, the longer the time (frequency) indicating that the user has leaned backward in the accumulated daily posture detailed information, the darker or more opaque the second area (1140) can be displayed. The longer the time (frequency) indicating that the user has leaned to the left in the accumulated daily posture detailed information, the darker or more opaque the third area (1150) can be displayed. The longer the time (frequency) that the user indicates that he or she is leaning forward in the cumulative daily posture details, the darker or more opaque the fourth area (1160) may be displayed.

[0174] More specifically, the user terminal (100) may perform a step of acquiring accumulated daily posture information by accumulating daily posture information at a predetermined cycle for a predetermined collection time. The user terminal (100) may perform a step of acquiring a time when the daily posture information indicating a forward leaning is present in the accumulated daily posture information as a forward posture time. The user terminal (100) may perform a step of acquiring a time when the daily posture information indicating a backward leaning is present in the accumulated daily posture information as a backward posture time. The user terminal (100) may perform a step of acquiring a time when the daily posture information indicating a left leaning is present in the accumulated daily posture information as a left posture time. The user terminal (100) may perform a step of acquiring a time when the daily posture information indicating a right leaning is present in the accumulated daily posture information as a right posture time. User

[0175] The user terminal (100) may perform a step of displaying the first area darker, darker, or more opaque as the forward posture time increases. The user terminal (100) may perform a step of displaying the second area darker, darker, or more opaque as the rear posture time increases. The user terminal (100) may perform a step of displaying the third area darker, darker, or more opaque as the left posture time increases. The user terminal (100) may perform a step of displaying the fourth area darker, darker, or more opaque as the right posture time increases. In the above, it has been described that the areas (1130, 1140, 1150, 1160) are displayed darker, darker, or more opaque as the posture time corresponding to the direction information increases, but the present invention is not limited thereto. The user terminal (100) may display areas (1130, 1140, 1150, 1160) as bright, blurry, or transparent as the posture time corresponding to the direction information increases. In addition, the user terminal (100) may display the posture time corresponding to the direction information as numbers in areas (1130, 1140, 1150, 1160).

[0176] Referring to Fig. 13, for example, in the accumulated posture details, the time the user was leaning to the left may be the longest, the time the user was leaning forward or backward may be intermediate, and the time the user was leaning to the right may be the shortest. In this case, the user terminal (100) may display the third area (1150) as the darkest, the first area (1130) and the second area (1140) as intermediately dark, and the fourth area (1160) as the brightest.

[0177] In this way, the posture correction system (200) of the present disclosure can display the user's predominant posture. Even if the user's posture is not tilted to the extent that it would cause strain to the spine, a persistent posture leaning to one side can cause strain to the spine. The posture correction system (200) of the present disclosure can enable the user to continuously pay attention to their posture by allowing the user to check the posture they predominantly assume.

[0178] FIG. 14 is a drawing for explaining the operation of a user terminal according to one embodiment of the present disclosure.

[0179] As described in FIG. 12, at least one of the first region (1130), the second region (1140), the third region (1150), and the fourth region (1160) can be divided into two or more regions.

[0180] For example, the first region (1130) may include a 1-1 region (1411) and a 1-2 region (1412). The second region (1140) may include a 2-1 region (1421) and a 2-2 region (1422). The third region (1150) may include a 3-1 region (1431) and a 3-2 region (1432). The fourth region (1160) may include a 4-1 region (1441) and a 4-2 region (1442).

[0181] If the daily posture information indicates that the user is leaning forward, and the size information of the tilt included in the daily posture information is included in the reference range information, the user terminal (100) may perform a step of displaying the 1-1 area (1411) in a predetermined color. The 1-1 area (1411) may be displayed in a predetermined color when the user is leaning forward but has a correct posture. The fact that the size information of the tilt included in the daily posture information is included in the reference range information may indicate that the user is leaning forward but is leaning within a normal range.

[0182] If the daily posture information indicates that the user is leaning forward and the size information of the tilt included in the daily posture information is not included in the reference range information, the user terminal (100) may perform a step of displaying the 1-2 area (1412) in a predetermined color. The 1-2 area (1412) may be displayed in a predetermined color when the user is leaning forward and has a bad posture. The fact that the size information of the tilt included in the daily posture information is not included in the reference range information may indicate that the user is leaning excessively. In other words, it may indicate that the user is having a bad posture.

[0183] If the daily posture information indicates that the user is leaning backwards and the size information of the tilt included in the daily posture information is included in the reference range information, the user terminal (100) may perform a step of displaying the 2-1 area (1421) in a predetermined color. The 2-1 area (1421) may be displayed in a predetermined color when the user is leaning backwards but has a correct posture. The fact that the size information of the tilt included in the daily posture information is included in the reference range information may indicate that the user is leaning but is leaning within a normal range.

[0184] If the daily posture information indicates that the user is leaning backwards and the size information of the tilt included in the daily posture information is not included in the reference range information, the user terminal (100) may perform a step of displaying the 2-2 area (1422) in a predetermined color. The 2-2 area (1422) may be displayed in a predetermined color when the user is leaning backwards and has a bad posture. The fact that the size information of the tilt included in the daily posture information is not included in the reference range information may indicate that the user is leaning excessively. In other words, it may indicate that the user is having a bad posture.

[0185] If the daily posture information indicates that the user is leaning to the left, and if the size information of the tilt included in the daily posture information is included in the reference range information, the user terminal (100) may perform a step of displaying the 3-1 area (1431) in a predetermined color. The 3-1 area (1431) may be displayed in a predetermined color when the user is leaning to the left but has a correct posture. The fact that the size information of the tilt included in the daily posture information is included in the reference range information may indicate that the user is leaning but is leaning within a normal range.

[0186] If the daily posture information indicates that the user is leaning to the left and the size information of the tilt included in the daily posture information is not included in the reference range information, the user terminal (100) may perform a step of displaying the 3-2 area (1432) in a predetermined color. The 3-2 area (1432) may be displayed in a predetermined color when the user is leaning to the left and has a bad posture. The fact that the size information of the tilt included in the daily posture information is not included in the reference range information may indicate that the user is leaning excessively. In other words, it may indicate that the user is having a bad posture.

[0187] If the daily posture information indicates that the user is leaning to the right, and if the size information of the tilt included in the daily posture information is included in the reference range information, the user terminal (100) may perform a step of displaying the 4-1 area (1441) in a predetermined color. The 4-1 area (1441) may be displayed in a predetermined color when the user is leaning to the right but has a correct posture. The fact that the size information of the tilt included in the daily posture information is included in the reference range information may indicate that the user is leaning but is leaning within a normal range.

[0188] If the daily posture detailed information indicates that the user is leaning to the right and the size information of the tilt included in the daily posture detailed information is not included in the reference range information, the user terminal (100) may perform a step of displaying the 4-2 area (1442) in a predetermined color. The 4-2 area (1442) may be displayed in a predetermined color when the user is leaning to the right and has a bad posture. The fact that the size information of the tilt included in the daily posture detailed information is not included in the reference range information may indicate that the user is leaning excessively. In other words, it may indicate that the user is having a bad posture.

[0189] Referring to FIG. 14, it can be confirmed that the 3-2 area (1432) displayed on the user terminal (100) is the darkest, so that the user often takes a bad posture to the left. In addition, among the 1-1 area (1411), the 2-1 area (1421), the 3-1 area (1431), and the 4-1 area (1441) displayed on the user terminal (100), the 4-1 area (1441) is the darkest, so that the user mostly leans to the right when taking a good posture.

[0190] In this way, the posture correction system (200) of the present disclosure can identify the direction in which the user is leaning, not only when the user is in a good posture but also when the user is in a bad posture. Accordingly, the user can make a careful effort to correct their posture. Since it is difficult for the user to observe themselves from the outside, it is difficult for the user to know what posture they are in. However, the posture correction system (200) of the present disclosure can provide the user with various information, thereby enabling the user to objectively check their posture.

[0191] In Fig. 14, the posture correction system (200) divides each of the areas (1130, 1140, 1150, 1160) into two areas, but is not limited thereto. The posture correction system (200) can divide each of the areas (1130, 1140, 1150, 1160) into three or more areas. The posture correction system (200) can indicate that the user's inclination becomes greater the farther away from the center of the circle. The posture correction system (200) can divide each of the areas (1130, 1140, 1150, 1160) into three or more areas based on predetermined reference range information. Additionally, the posture correction system (200) can display predetermined colors in areas (1130, 1140, 1150, 1160) based on at least one of daily posture information or accumulated daily posture information. As this has already been explained, a duplicate explanation will be omitted.

[0192] FIG. 15 is a drawing for explaining the operation of a user terminal according to one embodiment of the present disclosure.

[0193] Fig. 15 illustrates a case where, unlike Figs. 12 to 14, the direction information of the slope included in the daily posture information includes forward, backward, left, right, left-front, left-back, right-front, and right-back. The first region (1130), the second region (1140), the third region (1150), and the fourth region (1160) have already been described, so any redundant description will be omitted.

[0194] The user terminal (100) can further display a fifth area (1510), a sixth area (1520), a seventh area (1530), and an eighth area (1540). The user terminal (100) can divide each of the fifth area (1510), the sixth area (1520), the seventh area (1530), and the eighth area (1540) into at least one area. FIG. 15 illustrates a case where the user terminal (100) divides each of the fifth area (1510), the sixth area (1520), the seventh area (1530), and the eighth area (1540) into two areas.

[0195] The fifth region (1510) may be related to a case where the direction information of the daily posture information indicates left-front. The sixth region (1520) may be related to a case where the direction information of the daily posture information indicates left-back. The seventh region (1530) may be related to a case where the direction information of the daily posture information indicates right-back. The eighth region (1540) may be related to a case where the direction information of the daily posture information indicates right-front.

[0196] The fifth region (1510) may include the 5-1 region (1511) and the 5-2 region (1512). The 5-1 region (1511) may be displayed in a predetermined color when the direction information of the daily posture information indicates left-front and the size information is included in the reference range information. The user terminal (100) may obtain the time for maintaining the correct posture while leaning left-front based on the accumulated daily posture information, and may display the 5-1 region (1511) in a predetermined color based on this time.

[0197] The 5-2 area (1512) may be displayed in a predetermined color when the direction information of the daily posture information indicates left-front and the size information is not included in the reference range information. The user terminal (100) may obtain the time for which a bad posture was maintained while leaning left-front based on the accumulated daily posture information, and may display the 5-2 area (1512) in a predetermined color based on this time.

[0198] Area 6-1 (1521) may be displayed in a predetermined color when the direction information of the daily posture information indicates left-backward and the size information is included in the reference range information. The user terminal (100) may obtain the time for which the correct posture was maintained while leaning left-backward based on the accumulated daily posture information, and may display Area 6-1 (1521) in a predetermined color based on this time.

[0199] The 6-2 area (1522) may be displayed in a predetermined color when the direction information of the daily posture information indicates left-backward and the size information is not included in the reference range information. The user terminal (100) may obtain the time for which a bad posture was maintained while leaning left-backward based on the accumulated daily posture information, and may display the 6-2 area (1522) in a predetermined color based on this time.

[0200] Area 7-1 (1531) may be displayed in a predetermined color when the direction information of the daily posture information indicates right-backward and the size information is included in the reference range information. The user terminal (100) may obtain the time for which the correct posture was maintained while leaning to the right-backward based on the accumulated daily posture information, and may display Area 7-1 (1531) in a predetermined color based on this time.

[0201] The 7-2 area (1532) may be displayed in a predetermined color when the direction information of the daily posture information indicates right-backward and the size information is not included in the reference range information. The user terminal (100) may obtain the time for which the user maintained a bad posture while leaning to the right-backward based on the accumulated daily posture information, and may display the 7-2 area (1532) in a predetermined color based on this time.

[0202] Area 8-1 (1541) may be displayed in a predetermined color when the direction information of the daily posture information indicates right-forward and the size information is included in the reference range information. The user terminal (100) may obtain the time for which the correct posture was maintained while leaning to the right-forward based on the accumulated daily posture information, and may display Area 8-1 (1541) in a predetermined color based on this time.

[0203] The 8-2 area (1542) may be displayed in a predetermined color when the direction information of the daily posture information indicates right-forward and the size information is not included in the reference range information. The user terminal (100) may obtain the time for which a bad posture was maintained while leaning to the right-forward based on the accumulated daily posture information, and may display the 8-2 area (1542) in a predetermined color based on this time.

[0204] In this way, the posture correction system (200) of the present disclosure can provide more precise information to the user by displaying detailed information on the user's inclination direction. Accordingly, the posture correction system (200) can propose more precise posture corrections, and the user can perform posture corrections more precisely.

[0205] Fig. 16 is a flowchart showing the operation of a posture correction system according to one embodiment of the present disclosure.

[0206] The user terminal (100) can perform a step (1610) of acquiring accumulated daily posture information by accumulating daily posture information at a predetermined cycle for a predetermined collection time. As this has already been explained, a duplicate explanation will be omitted.

[0207] The user terminal (100) can perform a step (1620) of acquiring the time for which the daily posture information is included in the reference range information among the accumulated daily posture detailed information as the correct posture time. As previously described, the accumulated daily posture detailed information can include multiple pieces of daily posture information acquired periodically. If the size information of one piece of daily posture information acquired in a predetermined cycle among the multiple pieces of daily posture information is included in the reference range information, the user terminal (100) can determine that the correct posture has been maintained during the corresponding cycle. The user terminal (100) can accumulate and measure the time for which the correct posture is maintained during the predetermined collection time.

[0208] The user terminal (100) can perform a step (1630) of acquiring a time when the user has daily posture information that is not included in the reference range information among the accumulated daily posture information as a bad posture time. If the size information of one daily posture information acquired in a predetermined cycle among a plurality of daily posture information is not included in the reference range information, the user terminal (100) can determine that the user has maintained a bad posture during the corresponding cycle. The user terminal (100) can accumulate and measure the time when the user has maintained a bad posture during the predetermined collection time.

[0209] The user terminal (100) can perform the step (1640) of obtaining a posture score by dividing the correct posture time by the collection time. However, it is not limited thereto. The user terminal (100) can also obtain a posture score by multiplying the value obtained by dividing the correct posture time by the collection time by 100. Using other methods, the user terminal (100) can obtain a posture score that increases as the correct posture time increases.

[0210] Additionally, the user terminal (100) can perform a step (1650) of displaying a detailed score. To explain step (1650), FIG. 17 is described.

[0211] FIG. 17 is a drawing showing a screen of a user terminal according to one embodiment of the present disclosure.

[0212] The user terminal (100) can display a screen (1700). This screen may display weekly statistics. That is, the screen (1700) indicates a case where the collection time is one week.

[0213] The user terminal (100) can display the correct posture time (1710). The correct posture time (1710) can be a value obtained by dividing the correct posture time accumulated over a week by 7 days. In addition, the correct posture time (1710) can also indicate an increase in the correct posture time compared to the previous week.

[0214] The correct posture time may be information acquired in step (1620). The user terminal (100) may display a posture score (1720). The posture score (1720) quantifies the time spent maintaining the correct posture, but does not indicate how long the user actually maintained the correct posture. Therefore, the user terminal (100) may simultaneously display the correct posture time (1710) and the posture score (1720) on a single screen.

[0215] The user terminal (100) can obtain a posture score based on the correct posture time for a week. However, this is not limited to this, and the user terminal (100) can obtain a posture score based on the correct posture time for each day, and can also display the average of the posture scores obtained each day on the screen (1700).

[0216] The user terminal (100) can display the correct posture time as a graph (1730). The graph (1730) can display the user's correct posture time for a week by day of the week.

[0217] The posture correction system of the present disclosure indicates whether the user's posture is correct through a posture score, allowing the user to strive to improve their posture score. Furthermore, since the user's posture score will increase in proportion to the effort the user puts into maintaining good posture, the user will be motivated to maintain good posture. Therefore, the user can use the posture correction system (200) for a long period of time.

[0218] At least one of the user terminal (100) and the posture sensor (220) of the present disclosure can perform a step of acquiring a posture pattern of a user. The posture pattern can include at least one of a standing posture, a crouching posture, a sitting posture on a chair, a walking posture, and a running posture. The posture pattern can include a plurality of exercise postures. For example, the posture pattern can include squats, sit-ups, Pilates, swimming, running, etc. At least one of the user terminal (100) and the posture sensor (220) can determine the current posture pattern of the user based on an inertial sensor.

[0219] The user terminal (100) can select at least one of reference range information and a posture estimation machine learning model corresponding to the posture pattern. The reference range information may vary depending on the user's posture pattern. That is, the correct posture may be determined differently depending on whether the user is standing or sitting. The user terminal (100) can determine the reference range information based on the user's current posture pattern, thereby determining that the correct posture varies depending on the posture pattern. By selecting the reference range information based on the posture pattern in this way, the performance of the posture correction system (200) can be improved.

[0220] In addition, the posture estimation machine learning model may be configured to output 3D posture information based on posture image information as previously described. The performance of the posture estimation machine learning model may vary depending on the posture pattern. The user terminal (100) or the server (210) may select the posture estimation machine learning model based on the user's current posture pattern. Accordingly, the user terminal (100) or the server (210) may utilize a posture estimation machine learning model optimized for the posture pattern. Accordingly, the performance of the posture correction system (200) may be improved.

[0221] FIG. 19 is a drawing for explaining the operation of a posture correction system according to one embodiment of the present disclosure.

[0222] The posture correction system (200) may further include a spinal disease estimation machine learning model (1920). The spinal disease estimation machine learning model (1920) may be a machine learning model for predicting spinal disease information from accumulated daily posture information. The spinal disease estimation machine learning model (1920) may be a model for selecting one of multiple spinal diseases based on the accumulated daily posture information. The spinal disease estimation machine learning model (1920) may be implemented based on a CNN or transformer-based machine learning model. The spinal disease estimation machine learning model (1920) may use past accumulated daily posture information (1911) and past spinal disease information (1912) as learning data. The past accumulated daily posture information (1911) may be an analysis target, and the past spinal disease information (1912) may be label information acquired by a person or a device. The past spinal disease information (1912) may be a ground truth value. The past cumulative daily posture detailed information (1911) and the past spinal disease information (1912) may correspond one-to-one. The posture correction system (200) may collect the past cumulative daily posture detailed information (1911) from multiple users. In addition, the posture correction system (200) may obtain the spinal disease information that has developed within a predetermined analysis time for the user corresponding to the past cumulative daily posture detailed information (1911) as the past spinal disease information (1912). The predetermined analysis time may be a value of one day or more and one year or less. The spinal disease information may be identification information of the spinal disease, and may be, for example, a disease code included in the spinal disease.

[0223] At least one of the past correct posture time or past bad posture time may be used for machine learning, along with the past cumulative daily posture details (1911). However, this is not limited to the past cumulative daily posture details (1911), and at least one of the past correct posture time or past bad posture time may be used for machine learning instead. At least one of the past correct posture time or past bad posture time may include the direction of inclination and the time spent in the inclination state.

[0224] The server (210) can update the spinal disease estimation machine learning model (1920) through forward propagation and back propagation. The server (210) can machine learn the correlation between past spinal disease information (1912) and past accumulated daily posture detailed information (1911). When the accuracy of the spinal disease estimation machine learning model (1920) is sufficiently high, the server (210) can stop updating the spinal disease estimation machine learning model (1920). The server (210) can transmit the spinal disease estimation machine learning model (1920) to the user terminal (100) or another server. The server (210) can store the spinal disease estimation machine learning model (1920).

[0225] The server (210) can receive cumulative daily posture detailed information (1931) about the current user from the user terminal (100) or the posture sensor (220). Since the process of obtaining the cumulative daily posture detailed information (1931) has already been described, a duplicate description will be omitted. The server (210) can perform a step of obtaining predicted spinal disease information (1940) by applying the cumulative daily posture detailed information (1931) to a spinal disease estimation machine learning model (1920). The spinal disease information (1940) obtained by applying the spinal disease estimation machine learning model (1920) may be predicted information. The spinal disease information (1940) predicted by the spinal disease estimation machine learning model (1920) may be very similar to actual spinal disease information.

[0226] The output information of the spinal disease estimation machine learning model (1920) may be expressed as a disease probability for each of multiple spinal diseases. That is, the server (210) may perform a step of applying the accumulated daily posture information to the spinal disease estimation machine learning model to obtain a disease probability for each of multiple spinal diseases.

[0227] The posture correction system (200) can select the highest probability value that is greater than or equal to a predetermined threshold probability among the disease probabilities for each of a plurality of spinal diseases. The posture correction system (200) can select the spinal disease corresponding to the highest probability value from the plurality of spinal diseases and determine it as spinal disease information (1940). The posture correction system (200) can output a probability value corresponding to the spinal disease information (1940). If the probability value corresponding to the spinal disease information (1940) is very high, it can be considered that the probability of contracting the disease corresponding to the spinal disease information (1940) within the analysis time is high. In addition, if the probability value corresponding to the spinal disease information (1940) is relatively low, it can be considered that the probability of contracting the disease corresponding to the spinal disease information (1940) within the analysis time is low. The user can check the probability of contracting the disease corresponding to the spinal disease information (1940) and be motivated to correct his / her posture.

[0228] Below, the process by which the posture correction system (200) determines whether a user is maintaining a healthy posture for each body part using 3D posture information is described. If in step (360) the posture correction system (200) determines whether the user is properly following the reference posture information based on exercise information, in the following steps the posture correction system (200) determines for each body part whether the user is maintaining a proper posture that is beneficial to health.

[0229] FIG. 20 is a drawing for explaining the operation of a user terminal according to one embodiment of the present disclosure.

[0230] At least one of the server (210) and the user terminal (100) may perform a step of determining the posture similarity between the 3D posture information and the predetermined correct posture information. The step of determining the posture similarity may be similar to step (360).

[0231] At least one of the server (210) and the user terminal (100) can determine whether the 3D posture information for each predetermined body part of the user matches the correct posture information. The predetermined correct posture information may be information for determining whether the user is taking a healthy posture for the predetermined body part. The posture correction system (200) can increase the health score of the corresponding body part as the 3D posture information matches the correct posture information. However, the present invention is not limited thereto, and the posture correction system (200) can also use previously determined bad posture information instead of the correct posture information. The posture correction system (200) can lower the score of the corresponding body part as the 3D posture information matches the bad posture information. The correct posture information or the bad posture information may be determined by an expert such as a medical professional or a trainer.

[0232] The predetermined body part may include at least one of the torso, left arm, right arm, left leg, and right leg. For example, if the user is sitting with the right shoulder raised higher than the left shoulder, the posture correction system (200) may lower the health score of the left arm or right arm. Additionally, if the user is sitting with the right leg raised above the left leg, the posture correction system (200) may lower the health score of the left leg or right leg. Additionally, if the user is excessively tilting the spine with respect to the ground, the posture correction system (200) may lower the health score of the torso.

[0233] The posture correction system (200) can determine at least one of correct or bad posture information based on a posture pattern. Since the method for determining the posture pattern has already been described, a detailed explanation will be omitted. Depending on the posture pattern, correct and bad postures may differ. For example, if the posture pattern is a squat, crouching with the knees out in front of the feet may be detrimental to the user's legs. However, if the user sits on a chair and the knees are in front of the feet, this may not be a problem at all.

[0234] As previously described, 3D posture information may include identification information corresponding to a body feature node and angles of joints corresponding to the identification information. Correct posture information may also include identification information corresponding to a body feature node and angles of joints corresponding to the identification information. The server (210) may determine the posture similarity between the angles of joints of 3D posture information having the same identification information and the angles of joints of the reference posture information. For example, the posture similarity may be determined by a difference ratio determined by the following equation.

[0235]

[0236] Difference ratio = |Joint angle of 3D posture information - Joint angle of correct posture information| / Joint angle of correct posture information

[0237]

[0238] The smaller the difference ratio, the higher the health score for a predetermined body part. The posture correction system (200) can obtain health scores for each predetermined body part. The user terminal (100) can display a graph indicating the health score for each body part on a portion of the screen (2010). For example, the farther away from the center of the graph, the higher the health score for the corresponding body part.

[0239] According to various embodiments of the present disclosure, the posture correction system (200) can acquire three-dimensional posture information at predetermined intervals during a collection time. The posture correction system (200) can determine that the three-dimensional posture information matches the correct posture information when the difference ratio for one of the plurality of three-dimensional posture information is below a predetermined threshold ratio. The posture correction system (200) can determine that the three-dimensional posture information matches the correct posture information for a predetermined body part. In addition, when the three-dimensional posture information matches the correct posture information, the posture correction system (200) can determine that the user maintained the correct posture for a period corresponding to the three-dimensional posture information. The posture correction system (200) can acquire the correct posture time, which is the time for which the user maintained the correct posture, for each predetermined body part based on the plurality of three-dimensional posture information during the collection time. The posture correction system (200) can determine a score based on a value obtained by dividing the correct posture time by the collection time. Additionally, the user terminal (100) can display a graph indicating health scores for each body part on one area (2010) of the screen.

[0240] In the above, the health score was obtained using correct posture information, but the posture correction system (100) can also use bad posture information.

[0241]

[0242] Difference ratio = |Joint angle of 3D posture information - Joint angle of bad posture information| / Joint angle of bad posture information

[0243]

[0244] At this time, the smaller the difference ratio, the lower the health score for the predetermined body part. The user terminal (200) can display a graph indicating the health score for each body part on a section (2010) of the screen.

[0245] As previously described, the posture correction system (200) can obtain a posture score for the spine. The posture correction system (200) can divide the spine into the cervical, thoracic, and lumbar vertebrae and obtain a posture score for each. The posture correction system (200) can compare the 3D posture information with the correct posture information to determine the correct posture time. The posture correction system (200) can determine the correct posture time for at least one of the cervical, thoracic, and lumbar vertebrae. The correct posture time may mean the time during which the 3D posture information matches the correct posture information. Matching may mean a case where the difference ratio based on the 3D posture information and the correct posture information is below a predetermined threshold difference. However, the present invention is not limited thereto, and the correct posture time may also mean the time during which the 3D posture information does not match the bad posture information.

[0246] The posture correction system (200) can obtain a posture score by dividing the correct posture time by the collection time. The posture score can include at least one of a cervical posture score, a thoracic posture score, and a lumbar posture score. The user terminal (100) can display at least one of a cervical posture score, a thoracic posture score, and a lumbar posture score in an area (2020). In addition, the user terminal (100) can display a posture score for the spine in an area (2030). The posture score for the spine can be the posture score obtained in step (1640).

[0247] The posture correction system (200) can recommend exercise information based on at least one of a posture score for the spine, a cervical posture score, a thoracic posture score, a lumbar posture score, and a health score for each predetermined body part. More specifically, the posture correction system (200) can store an exercise information selection table that corresponds exercise information to at least one of a posture score for the spine, a cervical posture score, a thoracic posture score, a lumbar posture score, and a health score for each predetermined body part. The posture correction system can select exercise information corresponding to at least one of a posture score for the spine, a cervical posture score, a thoracic posture score, a lumbar posture score, and a health score for each predetermined body part from the exercise information selection table. The user can check the exercise information by inputting an input to a button (2040) into the user terminal (100).

[0248] As such, the posture correction system of the present disclosure can display scores for not only the spine but also each body part, allowing users to monitor their physical health by body part. Furthermore, users can identify areas requiring attention for health and strive to improve those areas. Therefore, by knowing which areas to focus on to maintain their health, users can avoid investing resources in unnecessary areas. Therefore, the posture correction system of the present disclosure can dramatically improve the user's health.

[0249] We have discussed various embodiments so far. Those skilled in the art will appreciate that the present invention can be implemented in modified forms without departing from its essential characteristics. Therefore, the disclosed embodiments should be considered illustrative rather than restrictive. The scope of the present invention is set forth in the claims, not the foregoing description, and all differences within the scope equivalent thereto should be construed as being encompassed by the present invention.

[0250] Meanwhile, the embodiments of the present invention described above can be written as a program that can be executed on a computer, and can be implemented in a general-purpose digital computer that runs the program using a computer-readable recording medium. The computer-readable recording medium includes storage media such as magnetic storage media (e.g., ROM, floppy disk, hard disk, etc.) and optical reading media (e.g., CD-ROM, DVD, etc.).

Claims

1. A method for operating a posture correction system for correcting posture using artificial intelligence, comprising at least one of a server, a user terminal, and a posture sensor, A step for the user terminal to set the standard range information of the correct posture based on the user's input; A step of acquiring the user's daily posture information based on a daily posture sensor included in the above-mentioned posture sensor and measuring the inclination of the user's spine; A step of outputting a message indicating that the user terminal should correct the posture when the above daily posture details are not included in the above standard range information; A step of acquiring user's posture image information based on a motion posture sensor included in the above posture sensor and capturing the user's entire body; A step in which the server applies at least one of the detailed posture image information and the daily posture detailed information to a posture estimation machine learning model to obtain three-dimensional posture information; The step of the server determining the posture similarity between the 3D posture information and the predetermined reference posture information; and An operating method of a posture correction system, comprising a step of outputting a message indicating that the user terminal should correct the posture when the posture similarity is less than a predetermined threshold similarity.

2. In paragraph 1, The step of setting the above standard range information is: A step of receiving user input related to reference range information from a user; When the user's input indicates automatic range setting, the user terminal determines the preset initial range information as the reference range information; If the user's input indicates manual range setting, the user terminal displays a user interface for range setting; A step in which the user terminal acquires the inclination of the user's spine based on the daily posture sensor; A step of the user terminal changing and outputting the inclination of a body object representing at least a part of a human body included in the user interface based on the inclination of the spine; and A method of operating a posture correction system, comprising a step of the user terminal determining a range of inclination of the spine from 0 degrees as reference range information when the user confirms the inclination of the body object and inputs an object indicating confirmation.

3. In paragraph 1, The steps for obtaining the above daily posture information are: A step of displaying a daily posture object including a first area, a second area, a third area, and a fourth area; An operating method of a posture correction system, comprising a step of displaying at least one of the first area, the second area, the third area, and the fourth area with a predetermined color based on the daily posture information.

4. In paragraph 3, The steps indicated by the above predetermined colors are: A step of displaying the first area in a predetermined color when the above daily posture details indicate that the user is leaning forward; A step of displaying the second area in a predetermined color when the above daily posture details indicate that the user is leaning backwards; If the above daily posture details indicate that the user is leaning to the left, a step of displaying the third area in a predetermined color; and An operating method of a posture correction system, comprising a step of displaying the fourth area in a predetermined color when the above daily posture detailed information indicates that the user is leaning to the right.

5. In paragraph 3, The above first region includes the first-1 region and the first-2 region, The above second area includes the 2-1 area and the 2-2 area, The third area above includes the 3-1 area and the 3-2 area, The above fourth area includes the 4-1 area and the 4-2 area, A step of displaying the 1-1 area in a predetermined color when the above daily posture detailed information indicates that the user is leaning forward and the size information of the tilt included in the above daily posture detailed information is included in the above reference range information; If the above daily posture detailed information indicates that the user is leaning forward and the size information of the tilt included in the above daily posture detailed information is not included in the above reference range information, a step of displaying the 1-2 area in a predetermined color; A step of displaying the 2-1 area in a predetermined color when the above daily posture detailed information indicates that the user is leaning backwards and the size information of the tilt included in the above daily posture detailed information is included in the above reference range information; If the above daily posture detailed information indicates that the user is leaning backwards and the size information of the tilt included in the above daily posture detailed information is not included in the above reference range information, a step of displaying the 2-2 area in a predetermined color; When the above daily posture detailed information indicates that the user is tilted to the left and the size information of the tilt included in the above daily posture detailed information is included in the above reference range information, a step of displaying the 3-1 area in a predetermined color; If the above daily posture detailed information indicates that the user is tilted to the left and the size information of the tilt included in the above daily posture detailed information is not included in the above reference range information, a step of displaying the 3-2 area in a predetermined color; If the above daily posture detailed information indicates that the user is tilted to the right and the size information of the tilt included in the above daily posture detailed information is included in the above reference range information, a step of displaying the 4-1 area in a predetermined color; and An operating method of a posture correction system, comprising a step of displaying the 4-2 area in a predetermined color when the above daily posture detailed information indicates that the user is leaning to the right and the size information of the tilt included in the above daily posture detailed information is not included in the above reference range information.

6. In paragraph 3, A step in which the user terminal acquires accumulated daily posture information by accumulating daily posture information at a predetermined cycle for a predetermined collection time; A step of acquiring the time having daily posture information that is not included in the above-mentioned standard range information among the above-mentioned accumulated daily posture details and indicates a forward lean as the forward bad posture time; A step of acquiring the time having daily posture information that is not included in the above-mentioned standard range information among the above-mentioned accumulated daily posture details and indicates a backward lean as the backward bad posture time; A step of obtaining the time having daily posture information that is not included in the above-mentioned standard range information among the above-mentioned accumulated daily posture details and indicates a tilt to the left as the left bad posture time; A step of obtaining the time having daily posture information that is not included in the above-mentioned standard range information among the above-mentioned accumulated daily posture details and indicates a tilt to the right as the right bad posture time; The step of displaying the first area darker, thicker, or more opaque as the above-mentioned bad posture time increases; The step of displaying the second area darker, thicker, or more opaque as the above-mentioned bad posture time increases; The step of displaying the third area darker, thicker, or more opaque as the above left bad posture time increases; and A method of operating a posture correction system, comprising a step of displaying the fourth area darker, thicker, or more opaque as the above-mentioned right bad posture time increases.

7. In paragraph 3, A step in which the user terminal acquires accumulated daily posture information by accumulating daily posture information at a predetermined cycle for a predetermined collection time; A step of obtaining the time having the daily posture information indicating that the above accumulated daily posture information is tilted forward as the forward posture time; A step of obtaining the time when the daily posture information indicating that the above accumulated daily posture information is tilted backwards is used as the backward posture time; A step of obtaining the time having the daily posture information indicating that the above accumulated daily posture information is tilted to the left as the left posture time; A step of obtaining the time having daily posture information indicating that the above accumulated daily posture information is tilted to the right as the right posture time; A step of displaying the first area darker, thicker, or more opaque as the forward posture time increases; A step of displaying the second area darker, thicker, or more opaque as the above posterior posture time increases; The step of displaying the third area darker, thicker, or more opaque as the left posture time increases; and A method of operating a posture correction system, comprising a step of displaying the fourth area darker, thicker, or more opaque as the right posture time increases.

8. In paragraph 1, A step in which the user terminal acquires accumulated daily posture information by accumulating daily posture information at a predetermined cycle for a predetermined collection time; A step of obtaining the time having daily posture information included in the above standard range information among the above accumulated daily posture details as the correct posture time; A step of acquiring the time having daily posture information that is not included in the above standard range information among the above accumulated daily posture details as bad posture time; A step of obtaining a posture score by dividing the correct posture time by the collection time; and An operating method of a posture correction system including a step of displaying the above-mentioned posture score.

9. In paragraph 1, A step in which at least one of the user terminal and the posture sensor acquires a user's posture pattern; and Including a step of selecting at least one of the reference range information and the posture estimation machine learning model corresponding to the above posture pattern, A method of operation of a posture correction system, wherein the posture pattern includes at least one of a standing posture, a sitting posture, a walking posture, and a running posture.

10. In paragraph 1, The above three-dimensional posture information includes a plurality of body feature nodes at least partially corresponding to joints forming the body and skeletal edges connecting the plurality of body feature nodes. The step of obtaining the above 3D posture information is: An operating method of a posture correction system, including a step of making the angle of a skeletal edge corresponding to the spine included in the above three-dimensional posture information relative to the ground correspond to the above daily posture information.

11. In paragraph 1, A step in which the user terminal selects exercise information beneficial to the user from among a plurality of candidate exercise information based on the daily posture details; A step in which the user terminal outputs the exercise information; and An operating method of a posture correction system, comprising a step of the user terminal obtaining posture image information of the user based on the exercise posture sensor when the user assumes a posture based on the above exercise information.

12. In paragraph 11, The steps for selecting the above exercise information are: A step in which the user terminal obtains disease information from the user; A step of selecting some candidate exercise information by excluding exercise information harmful to the user from the plurality of candidate exercise information based on the above disease information; and An operating method of a posture correction system, comprising a step of selecting exercise information based on the daily posture detailed information among the above-mentioned partial candidate exercise information.

13. In paragraph 1, The steps for obtaining the above daily posture information are: A step of displaying a daily pose object including at least one concentric circle around an object viewed from above the user; and An operating method of a posture correction system, comprising a step of displaying a color corresponding to size information included in the daily posture information at a location corresponding to direction information included in the daily posture detailed information in the daily posture object.

14. In paragraph 1, A step in which the user terminal acquires accumulated daily posture information by accumulating daily posture information at a predetermined cycle for a predetermined collection time; A step in which the server applies the accumulated daily posture details to a spinal disease estimation machine learning model to obtain a disease probability for each of multiple spinal diseases; A step of selecting the highest probability value that is greater than or equal to a predetermined critical probability among the disease probabilities for each of the plurality of spinal diseases; and An operating method of a posture correction system, comprising a step of selecting a spinal disease corresponding to the highest probability value from the plurality of spinal diseases and determining the spinal disease information.

15. In paragraph 1, The step of the above user terminal determining the posture similarity between the above 3D posture information and the predetermined reference posture information is: A step of determining whether the above 3D posture information matches the correct posture information for each body part of the user determined in advance; The step of increasing the health score of the corresponding area as the above 3D posture information matches the correct posture information; and A method of operating a posture correction system, comprising the step of displaying a graph representing health scores for each body part on an area of ​​a screen.

Images