WO2024143980A1

WO2024143980A1 - Sensor-based human posture estimation device

Info

Publication number: WO2024143980A1
Application number: PCT/KR2023/020122
Authority: WO
Inventors: 이은영
Original assignee: 주식회사 포스처에이아이
Priority date: 2022-12-28
Filing date: 2023-12-07
Publication date: 2024-07-04
Also published as: KR20240105696A

Abstract

A posture estimation device using a sensor, according to an embodiment of the present invention, comprises: at least one sensor unit which is wearable on a user's body and senses a posture of the user; an image capture unit which captures an image of the user wearing the sensor unit; and a posture estimation unit which estimates a posture of the user by using image processing data obtained as a result of analyzing sensing data sensed by the sensor unit and the image captured by the image capture unit.

Description

Sensor-based human posture estimation device

The present invention relates to a human posture estimation device, and more specifically, to a posture estimation device that estimates human posture using a sensor mounted on the user's body.

Among the common diseases of modern people, diseases related to the musculoskeletal system are caused by poor posture or a lifestyle habit of maintaining the same posture for a long period of time due to the increase in sedentary lifestyles, computer-based work environments, and increased use of portable devices such as smartphones and tablets. There are many, and the trend of outbreaks is increasing. In particular, modern people who lead a sedentary lifestyle often repeat the posture of lowering their head or bending their back forward. If this posture continues, the head will gradually turn forward from the center of the body, the shoulders will slump, and the back will bend. In addition, due to the weight of the head, the muscles connected to the neck muscles become stiff due to tension in the neck muscles, and the shoulder muscles are also constantly tense to support the weight of the head, which easily leads to fatigue. In addition, when the back is bent, the joints of the spine point backwards from the center of the body, so the center of the body is prone to shift or tilt to the left or right.

Therefore, maintaining correct 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.

Human posture estimation technology is being used to estimate the user's posture by analyzing images taken of the user, but it is difficult to accurately measure the user's posture through image analysis alone.

[Prior art literature]

[Patent Document]

(Patent Document 1) Registered Patent Publication No. 10-1480026

The technical problem to be solved by the present invention is to provide a posture estimation device that estimates human posture using a sensor mounted on the user's body.

In order to solve the above technical problem, a posture estimation device according to an embodiment of the present invention includes at least one sensor unit that can be worn on the user's body and senses the user's posture; An imaging unit that photographs a user wearing the sensor unit; and a posture estimation unit that estimates the user's posture using sensing data sensed by the sensor unit and image processing data obtained by analyzing the image captured by the imaging unit.

Additionally, the posture estimation device may include a user terminal including the imaging unit, and display the estimated user's posture on a display unit of the user terminal.

In addition, the posture estimation unit generates big data from the sensing data and image processing data, and analyzes the user's posture in a 360-degree direction from the generated big data using a learned artificial intelligence (AI) neural network, The results of analyzing the big data can be displayed on the display unit.

According to embodiments of the present invention, the user's posture can be accurately measured, and the accuracy of posture estimation can be increased by applying it along with the user's image analysis. Through this, spinal diseases or musculoskeletal diseases can be prevented.

Figure 1 is a block diagram of a posture estimation device according to an embodiment of the present invention.

Figures 2 and 3 are exemplary diagrams of the sensor unit of the posture estimation device according to an embodiment of the present invention.

Figure 4 is a block diagram of the posture estimation unit of the posture estimation device according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings.

However, the technical idea of the present invention is not limited to some of the described embodiments, but may be implemented in various different forms, and as long as it is within the scope of the technical idea of the present invention, one or more of the components may be optionally used between the embodiments. It can be used by combining or replacing.

In addition, terms (including technical and scientific terms) used in the embodiments of the present invention, unless explicitly specifically defined and described, are generally understood by those skilled in the art to which the present invention pertains. It can be interpreted as meaning, and the meaning of commonly used terms, such as terms defined in a dictionary, can be interpreted by considering the contextual meaning of the related technology.

Additionally, the terms used in the embodiments of the present invention are for describing the embodiments and are not intended to limit the present invention.

In this specification, the singular may also include the plural unless specifically stated in the phrase, and when described as "at least one (or more than one) of A and B and C", it is combined with A, B, and C. It can contain one or more of all possible combinations.

Additionally, in describing the components of the embodiments of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are only used to distinguish the component from other components, and are not limited to the essence, sequence, or order of the component.

And, when a component is described as being 'connected', 'coupled', or 'connected' to another component, that component is directly 'connected', 'coupled', or 'connected' to that other component. In addition to cases, it may also include cases where the component is 'connected', 'coupled', or 'connected' by another component between that component and that other component.

Additionally, when described as being formed or disposed “on top” or “bottom” of each component, “top” or “bottom” means that the two components are directly adjacent to each other. This includes not only the case of contact, but also the case where one or more other components are formed or disposed between the two components. In addition, when expressed as “top” or “bottom,” the meaning of not only the upward direction but also the downward direction can be included based on one component.

Figure 1 is a block diagram of a posture estimation device according to an embodiment of the present invention. Figures 2 and 3 are exemplary diagrams of the sensor unit of the posture estimation device according to an embodiment of the present invention, and Figure 4 is a block diagram of the posture estimation unit of the posture estimation device according to an embodiment of the present invention.

The posture estimation device 100 according to an embodiment of the present invention consists of a sensor unit 110, a posture estimation unit 120, and an imaging unit 130.

The sensor unit 110 can be worn on the user's body and includes at least one sensor that senses the user's posture. The sensor unit 110 is implemented as a wearable device, can be worn on the user's body, and can sense the user's posture in a 360-degree direction. The sensor unit 110 may be mounted at the location of at least one feature point for sensing the user's posture. It can be installed in locations necessary to measure the user's posture, such as the user's body (center of the chest, back, shoulders), neck, head, arms, hands, hips, legs, and feet. At this time, the number of sensor units 110 mounted may vary depending on the complexity of the posture to be estimated. The sensor unit 110 may be composed of a separate module.

As shown in FIG. 2, the sensor unit 100 may include various types of devices that can measure the posture of users according to an embodiment of the present invention. The sensor unit 110 measures the user's neck, shoulders, waist, pelvis, and center of the chest as shown in (a) to (e) of FIG. 2 so that users can collect daily data on changes in posture while stopping or moving. ,It may include a sensor device worn on the leg. As shown in Figure 2(a), a sensor device is worn in the form of a belt on the user's waist or in the center of the chest (e.g., solar plexus), which is the center of the body, and as shown in Figure 2(b), a sports bra or shirt worn on the upper part of the body. A clothing type device such as a smart seat placed or installed on a chair where users sit, as shown in Figure 2(c), a turtle neck habit corrector as shown in Figure 2(d), or a body wearer as shown in Figure 2(e). It may include a sensor in the form of a patch that is directly attached to the. Here, a smart seat is related to sitting, but a sports bra or turtleneck habit corrector can be seen as a device related to moving. Furthermore, the sensor unit 110 according to an embodiment of the present invention may include a hybrid type product that combines not only a lifestyle wearable but also a musculoskeletal medical device, so the embodiment of the present invention is not specifically limited to any one type. No.

The sensor unit 110 may be configured to include a wearable sensor for measuring users' posture. For example, a smart seat may include a pressure sensor that can measure how pressure is distributed when the user sits on the chair. If the pressure value measured by the front pressure sensor is greater than the back, the user's posture can be considered to be tilted forward. Above all, the sensor mounted on the sensor unit 110 according to an embodiment of the present invention may include a sensor for sensing body movement or posture of users in a 360-degree direction. For example, it may include a 3-axis gyro sensor to measure the direction of 360 degrees. It may include a 3-axis acceleration sensor, a 3-axis gyro sensor, or a 6-axis or 9-axis sensor, and may include an inertial measurement unit (IMU). Here, the 9-axis sensor may be a high-precision 9-axis IMU sensor, and a 3-axis compass, 3-axis gyro, and 3-axis acceleration sensor can be combined to sense the object's movement, direction, vibration, compass data, and gyro data. It includes acceleration measurement of +/-2.5g and a gyro that can measure at speeds of 125°/s or less, and changes from high precision measurement to low precision measurement can be performed without interruption of sensor data. do. Alternatively, it may include a 5-axis physical sensor that is attachable and detachable from the body. In addition, it may include various sensors that can measure the user's position coordinates, tilt, movement vector, etc. The user's posture angle, posture angular velocity, acceleration, etc. can be measured using at least one of the sensors.

Through a 3-axis gyro sensor or a 3-axis acceleration sensor, the direction of front and back and left and right can be measured, and movement in 360 degrees, that is, the user's body posture, can be measured. In addition, it is natural that various types of sensors can be included to measure the user's posture area in front and back, left and right, and 360 degrees.

Sensing data acquired through the sensor of the sensor unit 110 is not limited to posture information of forward and backward inclinations, but includes posture information for left and right and 360 degrees, so that the user's accurate posture can be estimated.

The imaging unit 130 can capture images of a user wearing the sensor unit 110. The imaging unit 130 can capture images necessary to estimate the user's posture. At this time, when estimating the user's body posture, the image may be captured to include only part of the user's body, and when estimating the posture of the entire body, the image may be captured to include the user's entire body. Alternatively, only specific parts of the user can be photographed.

The imaging unit 130 may be included in the user terminal. Here, the user terminal may be a mobile terminal such as the user's mobile phone or tablet. Or, of course, it could be a PC terminal. The imaging unit 130 may include an image sensor and an image processing processor (ISP) mounted on the user terminal. The imaging unit 130 may be configured as a separate module.

The posture estimation unit 120 estimates the user's posture using the sensing data sensed by the sensor unit 110 and image processing data obtained by analyzing the image captured by the imaging unit 130.

The posture estimation unit 120 analyzes the image captured by the imaging unit 130 of the user, and estimates the user's posture using image processing data that is the result of analyzing the image and the result of analyzing the sensing data. can do.

The imaging unit 130 photographs the user, and the posture estimation unit 120 separates feature points, such as the user's joints, from the user's image captured by the imaging unit 130, and estimates the user's posture using the relative positions of the feature points. It can be used to do so. The process by which the posture estimation unit 120 estimates the posture from the user's image includes pictorial structure technique, depth-based posture estimation technique, deep learning-based 2D human posture estimation technique, and human posture estimation technique using heatmap. , Human pose estimation algorithms such as SOTA 2D human pose estimation technique, RGB-based human pose estimation technique, modeling-based human pose estimation technique, and Posenet can be used.

The posture estimation unit 120 estimates the user's posture using 360-degree sensing data sensed by the sensor unit 110. The posture estimation unit 120 estimates the user's current posture using the sensing data sensed by the sensor unit 110. Sensing data may include coordinate information or tilt information of a body whose movement is fixed, and may include coordinate information of a joint capable of movement or bent angle information.

The location of the feature point analyzed from the image captured by the imaging unit 130 may correspond to the user's wearing position where the sensor unit 110 is attached. Accurate posture information can be estimated by comparing the results of analyzing the feature points analyzed from the image and the sensing data sensed by the sensor unit 110 through a sensor mounted at a position corresponding to the feature point. Decisions are made based on image processing data, but can be supplemented with sensing data. For example, when feature points overlap in image processing data, the feature points may be distinguished or the relative position may be recognized differently from the actual posture. However, using sensing data, the relative position between each feature point can be accurately known, thereby estimating the posture. accuracy can be increased. Alternatively, the decision can be made based on the sensing data, supplemented with image processing data, or the user's posture can be estimated using both data.

The user's posture estimated by the posture estimation unit 120 may be displayed on the display of the user terminal. The user terminal may include a display unit, such as a display, and provide posture information of the user by displaying a human posture estimation result of estimating the user's posture on the display portion.

As shown in FIG. 3, the sensor unit 110 may be mounted on the solar plexus and back portion of the user's body, and sensing data measured by the first sensor (S) attached to the solar plexus and the second sensor (S) attached to the back. You can estimate the user's posture using . In the case of Figure 3 (a), where the first sensor (S) and the second sensor (S') are maintained perpendicular to the ground using the tilt information of the first sensor (S) and the second sensor (S'), The user's posture can be estimated as standing upright. However, in the case where the first sensor S and the second sensor S' are not perpendicular to the ground but are tilted as shown in FIG. 3(b), it can be assumed that the user's posture is in a slouching state. In Figure 3(b), the first sensor (S) and the second sensor (S') are tilted in opposite directions, while the first sensor (S) and the second sensor (S') are tilted in the same direction. If you lose, it can be assumed that you are leaning back or lying down, or, conversely, that you are leaning forward or lying down.

Figure 3 illustrates the body as an example, but the posture can be estimated by analyzing the movements of joints such as the neck, arms, or legs. Sensors are installed on both sides of the joint, and it is possible to estimate in which direction and how much the joint is bent through the relative coordinates and tilt between the sensors.

The posture estimation unit 120 generates big data from the sensing data, and analyzes the direction in which the user's body is tilted in a 360-degree direction from the generated big data using a learned artificial intelligence (AI) neural network. , the results of analyzing the big data can be provided to the user terminal.

The posture estimation unit 120 can estimate the posture using a machine-learned artificial intelligence neural network to estimate the posture from sensing data. Here, the artificial intelligence neural network may include various artificial intelligence neural networks such as convolutional neural network (CNN) and deep learning neural network (DNN). The training set of the artificial intelligence neural network can use sensing data measured in a specific posture as input data, and the specific posture can be used as comparison data, and the parameters of the artificial intelligence neural network learned through training can be stored in the storage unit.

As shown in FIG. 4, the posture estimation unit 120 may include a processing unit 121, a communication interface unit 122, a storage unit 123, and a posture correction unit 124. The processing unit 121 may include at least one processor that estimates the posture from sensing data. The processing unit 121 may include the artificial intelligence neural network described above.

The communication interface unit 122 may transmit and receive signals through communication with the sensor unit 110, a user terminal, or a server. In the process of performing communication, the communication interface unit 122 can perform operations such as modulation/demodulation, encoding/decoding, muxing/demuxing, and resolution conversion. In the case of user terminals such as smartphones, the screen size Since they are all different, the resolution can be converted to provide the optimal screen.

Communication of the communication interface unit 122 includes both wired and wireless communication networks. For example, a wired or wireless Internet network may be used or linked as a communication network. Here, the wired network includes Internet networks such as cable networks and public switched telephone networks (PSTN), and the wireless communication network includes CDMA, WCDMA, GSM, EPC (Evolved Packet Core), LTE (Long Term Evolution), and Wibro networks. It means including. Of course, the communication network 120 according to an embodiment of the present invention is not limited to this, and can be used in a cloud computing network, 5G network, etc. in a cloud computing environment as an access network for a next-generation mobile communication system to be implemented in the future. For example, if the communication network 120 is a wired communication network, the access point within the communication network can connect to the telephone company's exchange office, etc., but in the case of a wireless communication network, data is processed by connecting to the SGSN or GGSN (Gateway GPRS Support Node) operated by the communication company, or Data can be processed by connecting to various repeaters such as BTS (BaseTransceiver Station), NodeB, and e-NodeB.

A communication network may also include access points. Access points include small base stations such as femto or pico base stations that are often installed in buildings. Here, femto or pico base stations are classified according to the maximum number of sensor units 110 or user terminals that can be connected in the classification of small base stations. Of course, the access point may include a short-distance communication module for performing short-distance communication such as Zigbee and Wi-Fi with the sensor unit 110 or a user terminal. Access points can use TCP/IP or RTSP (Real-TimeStreaming Protocol) for wireless communication. Here, in addition to Wi-Fi, short-range communication can be performed using various standards such as Bluetooth, Zigbee, infrared (IrDA), RF (Radio Frequency) such as UHF (Ultra High Frequency) and VHF (Very High Frequency), and ultra-wideband communication (UWB). there is. Accordingly, the access point can extract the location of the data packet, designate the best communication path for the extracted location, and forward the data packet to the next device, such as the posture correction exercise recommendation device 130, along the designated communication path. Access points can share multiple lines in a typical network environment and may include, for example, routers, repeaters, and repeaters.

When the sensor unit 110 operates, the communication interface unit 122 can receive sensing data in real time and transmit it to the processing unit 121 if the corresponding sensor unit 110 is already registered. For example, when the user looks at FIG. 2, use a posture measurement device as shown in FIG. 2(a) and use a smart sheet as shown in FIG. 2(c) so that the user can know the 360-degree direction of body movement. If so, the data collected from the two sensor units 110 can be classified based on the sensor unit 110 possessed by the user and provided to the processing unit 121. According to an embodiment of the present invention, when a user registers his or her sensor unit 110, the device identification information of the sensor unit 110 is acquired and stored in the storage unit 123, etc., and the user uses the information to determine which user You can find out what sensor it has. This is because when the sensor unit 110 operates and transmits sensing data, it also transmits its own device identification information. Of course, the sensor unit 110 can also transmit information on the time when the sensing data is generated.

Additionally, the communication interface unit 122 may communicate with the user terminal when an app is executed on the user terminal, and may provide analysis results of data related to the user's posture according to the user's service request. Typically, a UI screen for posture in front and back, left and right, and 360 degree directions can be provided. Of course, in addition to this, the communication interface unit 122 may, at the request of the processing unit 121, send a notification for posture correction when the user's posture deviates from the standard posture. When using an application, notifications in the form of a push may be notified through the application.

The processing unit 121 may be responsible for the overall control operation of the communication interface unit 122, the posture correction unit 124, and the storage unit 123. Typically, when sensing data based on posture measurement for each user is received from the communication interface unit 122, the processing unit 121 temporarily stores it in the storage unit 123, retrieves it, and provides it to the posture correction unit 124. . In this process, when the processing unit 121 processes and provides data in a specified format to classify and store the data in the storage unit 123, the processing unit 121 systematically classifies and stores the data in the specified format in the storage unit 123. The communication interface unit 400 can be controlled.

The processing unit 121 may retrieve posture data of specific users from the storage unit 123 and perform an analysis operation. In addition, when the user terminal accesses a service and requests various types of data related to the user terminal, the processing unit 121 may provide a UI screen that graphically processes the user data to the user terminal. For example, it provides a screen related to posture measurement of precise inclinations in the front and back, left and right, and 360-degree directions, provides a screen for 1:1 coaching by tracking posture habits, or provides exercise based on the analysis results of the user's posture measurement data. Videos can be provided for recommendations. For this operation, the processing unit 121 may operate in conjunction with the posture correction unit 124.

The processing unit 121 may analyze sensing data related to posture measurement for each user acquired from the sensor unit 110. The processing unit 121 may learn posture-related data in advance using an artificial intelligence neural network and then perform posture estimation based on the learning results. Typically, machine learning includes supervised learning and unsupervised learning. Supervised learning is when the program designer sets the data as a reference posture when constructing learning data. On the other hand, unsupervised learning is a process in which an artificial intelligence program determines the standard posture on its own by analyzing big data. Therefore, unsupervised learning requires a lot of data for learning. Of course, even in the case of unsupervised learning, data related to the reference posture may be needed initially, so in embodiments of the present invention, it is possible to combine supervised learning and unsupervised learning, so it will not be limited to any one form. Artificial intelligence's deep learning program can increase accuracy when measuring posture. Therefore, based on the accurate results, matching posture correction exercises can be recommended to users' user terminals.

Above all, the processing unit 121 collects daily data about sitting, standing, and moving from the sensor unit 110 to generate big data, and when analyzing the big data, it is possible to precisely measure the 360-degree tilt in relation to the user's posture. It is possible. Of course, such 360-degree orientation measurement data can be acquired and provided through the 3-axis, 6-axis, and 9-axis gyro sensors of the sensor unit 110, and the posture correction unit 124 analyzes the data and provides We can recommend customized exercises based on your posture habits.

Of course, since the types of sensor units 110 carried by each user may be different, it is desirable for the posture correction unit 124 to recommend exercise based on data acquired from the sensor units 110 carried by each user. However, since the posture correction unit 124 collects and analyzes big data, for example, it can further increase accuracy when recommending posture correction exercise by reflecting data from the sensor unit 110 that users A and B do not possess. You can. In other words, it may be more accurate to recommend posture correction exercises based on the analysis results of data acquired by 100 people rather than recommending posture correction exercises based on the analysis results of data acquired by 5 people. Therefore, the posture correction unit 124 is not limited to the sensing data of the sensor unit 110 carried by the user, but can recommend exercise by comprehensively judging the posture measurement data of other users.

For example, if posture correction exercise is recommended based on the analysis results of the four sensor units 110 as shown in FIG. 2, when the user only possesses the posture measurement device as shown in FIG. 2(a), the corresponding posture measurement Posture correction exercise can be recommended by reflecting the sensing data of the device as well as data from the remaining three sensor units 110 that the user does not possess, and by reflecting data commonly extracted from many users. In this process, the posture correction unit 124 can reflect the characteristics of users with the same or similar gender, age, physical condition, and same or similar occupation, and retrieve their data and reflect it when recommending exercise. Accordingly, for example, if exercise A is recommended based on one sensor unit 110, exercise B may be recommended by also referring to data from other users. In other words, if exercise A is a result that reflects only the sensing data related to the neck, exercise B can be seen as a recommended result that comprehensively reflects the correlation between the neck, spine, and waist.

The storage unit 123 can temporarily store various types of data that are processed under the control of the processing unit 121. For example, when the posture measurement data of the sensor unit 110 is temporarily stored, the storage unit 123 may provide the posture measurement data to the posture correction unit 124 to enable data analysis.

The posture correction unit 124 can provide services such as posture correction or exercise recommendations to users. To this end, the posture correction unit 124 can build an all-in-one posture integration solution so that various services (e.g., exercise recommendations, posture correction subscriptions, etc.) can be provided in one place. For example, the posture correction unit 124 may measure users' real-time posture and provide a vibration notification directly to the sensor unit 110 for poor posture, or to a user terminal carried by the users. In addition, upon request from users holding the user terminal, analysis results related to individual postural habits can be provided through data collection and analysis. In addition, the posture correction unit 124 provides a posture correction subscription service and can perform AI-based 1:1 customized posture coaching through posture AI chatbot and messenger-based chat, and further provides posture correction exercises through the user terminal. It is also possible to recommend .

The posture correction unit 124 may provide exercise prescription services. In this way, the posture correction unit 124 can track posture habits, periodically determine changes in posture habits, and perform 1:1 coaching when posture correction is necessary. Additionally, when rehabilitation for posture habits is necessary, customized coaching is provided. Rehabilitation exercise prescriptions can be performed. The posture correction unit 124 can provide an integrated posture correction solution, or service, related to these operations. For example, it may be possible to predict the user's posture habits through big data analysis, and preventive actions can be taken by informing the user before a serious situation occurs and performing coaching actions to correct posture. . In this process, the posture correction unit 124 may recommend a specific exercise to the user for coaching and receive data on whether the user is performing the recommended exercise. It can be confirmed by analyzing the sensing data acquired from the sensor unit 110, etc., whether the user is actually performing posture correction exercises according to the coaching, and in this way, various forms of coaching can be performed depending on the user's situation.

The processing unit 121 can perform the following operations: analysis of time to maintain correct posture, analysis of data by cycle, provision of posture scores, and analysis of 360-degree posture area. Analysis of the time to maintain a correct posture can, for example, distinguish and determine the time to maintain a correct posture among the total time used by the sensor unit 110, generate daily activity data, and provide analysis results to the user terminal. Additionally, by analyzing the posture data, the daily posture can be scored and provided to the user terminal. Of course, with regard to posture, the standard data that serves as the standard for posture can be built in advance as a big data DB, so the user's posture correction range or posture correction index is calculated based on comparison with the standard data and scored based on this. The action can be performed. Of course, such a standard posture can also be determined by analyzing the video of a person who serves as a standard for posture, such as an expert, and comparing the analysis results. Furthermore, as sensing data about 360-degree body tilt is acquired, the direction in which the user's body is tilted can be visualized in 360 degrees through analysis and a visualization screen can be provided to the user's user terminal.

The posture correction unit 124 may be able to recognize the user's current state depending on the type of the sensor unit 110 and recommend appropriate exercise for the current state. For example, if you are using a turtle neck habit corrector as shown in (d) of Figure 2, videos related to appropriate neck exercises can be recommended, and if you use a smart sheet as shown in (c) of Figure 2, If it is determined that it exists, it may be possible to recommend appropriate exercise videos that can be performed in a sitting position. Of course, in order to provide this service, the posture correction unit 124 can construct recommended images such as appropriate videos that can be provided according to the type of each sensor unit 110 and the user's posture and provide services based on this. there is. In other words, the posture correction unit 124 can determine the user's current state and the type of the sensor unit 110 from the currently received sensing data. If it is determined that the user is in a seated state, for example, assuming that there are dozens of recommended images that can be recommended according to various postures, recommended images that match the results are extracted based on the results of analyzing the user's posture while seated. It can be provided to the user's user terminal.

In addition, the embodiment of the present invention provides posture correction in various fields related to posture correction, such as golf swing posture correction, fitness and virtual training, weightlifting posture correction, office worker posture correction, musculoskeletal system and physical therapy, through the posture correction unit 124. Since it is possible to provide any number of services, embodiments of the present invention will not be specifically limited to posture correction operations in daily life.

In addition to the above, the communication interface unit 122, processing unit 121, posture correction unit 124, and storage unit 123 of FIG. 4 can perform various operations. The communication interface unit 122, the processing unit 121, the posture correction unit 124, and the storage unit 123 may be composed of hardware modules that are physically separated from each other, and each module has internal functions for performing the above operations. You can save software and run it. However, the software is a set of software modules, and each module can be formed as hardware, so there is no particular limitation on the configuration of software or hardware. For example, the storage unit 123 may be hardware, such as storage or memory. However, since it is possible to store information through software (repository), there is no particular limitation to the above.

Meanwhile, as another embodiment of the present invention, the processing unit 121 may include a CPU and memory, and may be formed as a single chip. The CPU includes a control circuit, an arithmetic unit (ALU), an instruction interpretation unit, and a registry, and the memory may include RAM. The control circuit performs control operations, the operation unit performs operations on binary bit information, and the command interpretation unit includes an interpreter or compiler, which can convert high-level language into machine language and machine language into high-level language. , the registry can be involved in software data storage. According to the above configuration, the data operation processing speed can be rapidly increased by initially copying the program stored in the posture correction unit 124, loading it into memory, that is, RAM, and then executing it. In the case of deep learning models, they can be loaded into GPU memory rather than RAM and executed by accelerating the execution speed using GPU.

The user terminal can access the posture estimation device and check the analysis results of data related to each user's posture. Of course, the UX/UI screens provided to the user terminal may vary. The user terminal may be provided with an exercise recommendation service based on analysis of artificial intelligence (AI) according to an embodiment of the present invention. For example, exercise recommendations may be provided in the form of a video, or URL information that can access the video may be provided. In addition, the user terminal can receive a bad posture vibration notification, that is, when a bad posture is detected, a notification signal can be transmitted to vibrate the user terminal or the sensor unit 110, and individual information based on the analysis results of the collected data can be provided. It is also possible to check the results of postural habits analysis in the form of graphs or charts. For example, when the user is watching TV, the user's posture can be detected and a notification service such as "Your shoulders are bent too much and you need to stretch your deltoid muscle!" can be provided through the user terminal. Of course, the notification can be in the form of a text message, a messenger-based chat service, or a voice outputting the content through a speaker, or when a nearby artificial intelligence (AI) speaker is detected, a voice can also be output through it. Anything is possible. For example, the user terminal and the artificial intelligence speaker can perform direct communication such as beacon communication, and through this, the posture correction exercise recommendation device 130 can recognize what peripheral devices are present and provide the results. You can output voice through an artificial intelligence speaker.

In addition, the user terminal can check the results of posture data analysis through an application (hereinafter referred to as app) for using the service according to an embodiment of the present invention. Typically, a posture score may be provided as an analysis result. You can check your score by grading your daily posture. Of course, the calculation method can be performed in various ways, so it will not be specifically limited to any one method. For example, if the total time to maintain correct posture is 10 hours, and the correct posture was maintained for only 5 hours and the remaining time was not correct, the daily score could be 50 points. In this way, in addition to the score, the user terminal can check daily activity data by distinguishing the time for maintaining correct posture from the total usage time using the sensor unit 110. It may be possible to check this in the form of a bar graph.

The user terminal can check the user's real-time posture data periodically on a daily, weekly, and monthly basis. Above all, the user's 360-degree posture area analysis results can be checked using the 360-degree sensing data acquired from the sensor unit 110. there is. You can visualize the direction your body is tilted in 360 degrees and see it at a glance.

The user terminal can also have a conversation with a posture AI chatbot. Through this, users can also receive AI-based 1:1 customized posture coaching. In addition, the posture correction exercise recommendation device 130 learns answers to various questions in advance through an artificial intelligence program of an AI chatbot, so that the user terminal can receive answers to questions. Above all, artificial intelligence programs are different from existing rule-based chatbots in that they recognize text-type sentences like humans, so they can predict and provide answers. In other words, when a user asks a question, the chatbot cannot answer if it is not in the rules. However, in the case of artificial intelligence, predictions are possible based on the analysis results of the question, so by providing the user with the expected answer, it is possible to continue the conversation and provide an accurate answer to the user's desired query through data learning. . Therefore, even in the case of posture AI chatbots, when there is an inquiry from a user, an appropriate response is provided through an artificial intelligence deep learning program. In the case of posture AI according to an embodiment of the present invention, it is possible to learn data related to posture correction or exercise recommendations in advance and provide appropriate answers to users' inquiries based on this. Of course, the embodiment of the present invention will not be limited to such posture correction or exercise recommendation answers. The user terminal may also be able to use a posture correction subscription service. It may also be possible to check materials such as videos for AI-based 1:1 customized posture coaching through the subscription service screen.

The user terminal may be various types of devices other than smartphones. For example, the user terminal may be another wearable device that users wear on their wrists, etc., and mobile-based devices such as tablet PCs, or even PC-based terminal devices such as desktop computers or laptop computers may also be possible. You can. Therefore, in embodiments of the present invention, there will be no particular limitation on the type of user terminal.

In order to access the analyzed posture data, the user can perform authentication with the server that stores and provides it. First, the server may encrypt the identification number of a user group that can access the current body tilt value as posture data, generate a first password, and transmit it to the user terminal. The user can receive the first password, decode the first password using the identification number of the user group to which the user belongs, and read the body tilt value. You can generate a second password by encrypting the read body tilt value and your user identification number and send it to the server. The server reads the body tilt value from the second password using the user identification number, checks whether there is a user identification number whose body tilt value is the same as the body tilt value, and determines whether the user identification number is the same as the body tilt value. The user can be authenticated by determining whether it matches the user's user identification number. Since the first password and second password exchanged between the server and the user are encrypted using different data and are different for each user, even if one of them is hacked, other information can be prevented from being hacked.

Meanwhile, embodiments of the present invention can be implemented as computer-readable code on a computer-readable recording medium. Computer-readable recording media include all types of recording devices that store data that can be read by a computer system.

Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, and optical data storage devices. Additionally, computer-readable recording media are distributed in computer systems connected to a network. , computer-readable code can be stored and executed in a distributed manner. And functional programs, codes, and code segments for implementing the present invention can be easily deduced by programmers in the technical field to which the present invention pertains.

Those skilled in the art related to this embodiment will understand that the above-described substrate can be implemented in a modified form without departing from its essential characteristics. Therefore, the disclosed methods should be considered from an explanatory rather than a restrictive perspective. The scope of the present invention is indicated in the claims rather than the foregoing description, and all differences within the equivalent scope should be construed as being included in the present invention.

Claims

At least one sensor unit that can be worn on the user's body and senses the user's posture;

An imaging unit that photographs a user wearing the sensor unit; and

A posture estimation device comprising a posture estimation unit that estimates the user's posture using sensing data sensed by the sensor unit and image processing data obtained by analyzing images captured by the imaging unit.
According to paragraph 1,

The posture estimation device includes a user terminal including the imaging unit,

A posture estimation device that displays the estimated user's posture on a display of the user terminal.
According to paragraph 2,

The posture estimation unit,

Big data is generated from the sensing data and image processing data, the user's posture is analyzed in a 360-degree direction from the generated big data using a learned artificial intelligence (AI) neural network, and the results of analyzing the big data A posture estimation device that displays on the display unit.