KR20230032454A - Home training system using artificial intelligence - Google Patents

Abstract

A system for providing home training through analysis of a user's exercise motion using artificial intelligence, comprising: a joint measurement unit for measuring at least one joint of a user's body; Using an artificial intelligence algorithm, motion classifying unit for recognizing the user's exercise motion from the data by the joint measurement unit and classifying it into any one motion that is preset; and a coincidence rate calculation unit that compares the unit motion of the exercise, which is counted as 1 time, with a standard motion to calculate a motion coincidence rate.

Description

Home training system using artificial intelligence {HOME TRAINING SYSTEM USING ARTIFICIAL INTELLIGENCE}

The present invention relates to a home training system using artificial intelligence that recognizes and analyzes a user's exercise motion using a camera and provides feedback thereto.

As the use of gyms and sports facilities is restricted due to the influence of COVID-19, the number of 'homets' (people who exercise at home) who do personal exercise and home training is increasing. In the early days, I started home training to lose weight due to Corona-19, but now more people do home training for their health, and home training is becoming a new exercise culture. Due to this, at the same time, the smart healthcare industry is also in the limelight, and related contents are currently being developed.

However, home training has a problem in that, in the absence of professional knowledge about exercise, repetition of wrong motions may give a strain to joints and eventually lead to injury. In addition, the method of imitating it through a conventional video is limited to the cross-sectional shape of the trainer, and there is a problem in that feedback on the individual posture is difficult.

Korea Registered Patent Publication 102031243 (2019.10.04.) Apparatus for recognizing free weight training motion and method thereof Korean Patent Publication No. 1020170119333 (2017.10.27) Apparatus and Method for Processing of Exercise Information Korean Registered Patent Publication 102039616 (2019.10.28) Apparatus and method for managing exercise information (APPARATUS AND METHOD FOR MANAGING EXERCISE INFORMATION)

Embodiments of the present invention have been made to solve the above problems, and to provide a home training system using artificial intelligence that can maximize the exercise effect of home training without injury or safety accidents.

We want to accurately recognize the user's exercise motion and calculate the motion agreement rate compared to the standard motion. It is intended to visualize and provide exercise data including the user's exercise count, exercise time, and motion matching rate.

In addition, it is intended to further improve user convenience for this system by developing a dedicated application installed on the user's smart device. In addition, it is intended to accurately identify the user's body and provide feedback on incorrect posture during exercise.

In addition, it facilitates the coupling relationship between each component of the present system to make maintenance and management convenient.

An embodiment of the present invention is a system for providing home training through analysis of a user's exercise motion using artificial intelligence in order to solve the above problems, joint measurement for measuring at least one joint of the user's body. wealth; Using an artificial intelligence algorithm, motion classifying unit for recognizing the user's exercise motion from the data by the joint measurement unit and classifying it into any one motion that is preset; and a coincidence rate calculation unit that compares the unit motion of the exercise, which is counted as 1 time, with a standard motion to calculate a motion coincidence rate.

The joint measurement unit includes a first measurement unit disposed in front of the user and a second measurement unit disposed on one side of the user, and according to the motion classified by the motion classifying unit, the first measurement unit and the first measurement unit. It is desirable to vary the weight of the data by the two measuring units.

Preferably, the joint measuring unit measures the spatial position of the joint joint.

The coincidence rate calculation unit calculates the motion coincidence rate using a first feature group, which is a set of angle data formed between a plurality of previously set joint joints, and a second feature group, which is a set of distance data formed between the joint joints. It is desirable to do

The standard motion is composed of a plurality of standard postures, and the standard posture is preferably determined as a combination of at least one or more of the first characteristic group and at least one or more of the second characteristic group.

It is preferable to further include; a display unit that outputs a statistical graph for the number of times of exercise, exercise time, and motion agreement rate of the user.

The sensor unit is worn on both wrists of the user and measures the amount of change in the angle of the elbow over time, and the exercise preferably includes at least a push-up.

According to the problem solving means of the present invention as described above, various effects including the following can be expected. However, the present invention is not established when all of the following effects are exhibited.

The home training system using artificial intelligence according to an embodiment of the present invention can maximize the exercise effect of home training without injury or safety accidents. In addition, it is possible to accurately recognize the user's exercise motion and calculate a motion matching rate compared with a standard motion. In addition, it is possible to visualize and provide exercise data including the user's exercise frequency, exercise time, and motion matching rate.

In addition, user convenience can be further improved through a dedicated application installed in the user's smart device. In addition, it is possible to accurately identify the user's body and provide feedback on incorrect posture during exercise. In addition, it is possible to facilitate maintenance and management by facilitating the coupling relationship between each component of the present system.

1 is a block diagram of a home training system using artificial intelligence according to an embodiment.
Figure 2 is a conceptual diagram showing a joint for the body of Figure 1;
3 is a view showing angle data and distance data formed between joint joints of three points arbitrarily set in FIG. 2;
4 is an image of executing deep learning using an artificial intelligence algorithm to implement an exercise operation in the system of FIG. 1;
5 is an exemplary view of images from various angles for data sets in the system of FIG. 1;
FIG. 6 is an exemplary view showing a motion coincidence rate for a unit motion of an exercise through the system of FIG. 1;
7 is an exemplary view showing a statistical graph related to the system of FIG. 1;

In order to fully understand the configuration and effects of the present disclosure, preferred embodiments of the present disclosure will be described with reference to the accompanying drawings. However, the present disclosure is not limited to the embodiments disclosed below, and may be implemented in various forms and various changes may be made. Hereinafter, in the description of the present invention, detailed descriptions thereof will be omitted if it is determined that the related known functions may unnecessarily obscure the subject matter of the present invention as obvious matters to those skilled in the art.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The terms may only be used for the purpose of distinguishing one component from another. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element, without departing from the scope of the present disclosure.

In this application, the terms "include" or "have" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

Terms used in this application are only used to describe specific embodiments and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. Terms used in the embodiments of the present disclosure may be interpreted as meanings commonly known to those skilled in the art unless otherwise defined.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings. 1 is a block diagram of a home training system using artificial intelligence according to an embodiment, FIG. 2 is a conceptual diagram showing joint joints for the body of FIG. 1, and FIG. 3 is a joint joint of three points arbitrarily set in FIG. It is a diagram showing angle data and distance data formed between

4 is an image of executing deep learning using an artificial intelligence algorithm to implement exercise motions in the system of FIG. 1, and FIG. 5 is an example of images of various angles for data sets in the system of FIG. 1, FIG. 6 is an exemplary view showing a motion agreement rate for a unit motion of an exercise through the system of FIG. 1 , and FIG. 7 is an example view showing a statistical graph related to the system of FIG. 1 .

1 to 7, the home training system using artificial intelligence according to an embodiment of the present invention includes a joint measurement unit 100, a motion classification unit 1100, a coincidence rate calculation unit 1200, and a communication module 300. ), a display unit 400, a memory unit 500, a sensor unit 130, an external server 600, and the like. In this system, the joint measurement unit 100, the motion classification unit 1100, the coincidence rate calculation unit 1200, the communication module 300, the display unit 400, and the memory unit 500 are to be placed inside the user's smart device. can

This system improves the system that provides home training through analysis of the user's exercise motion using artificial intelligence. To this end, this system uses a conventional deep learning artificial intelligence algorithm. Deep learning estimates joint joints in the body using multiple data sets and calculates motion accuracy, that is, motion consistency rate, through motion classification for motion postures.

To this end, in this system, CNN is first used to predict the location of joint joints and to learn correlations between joint joints. This means recognizing the user first and extracting the main features from the body. In addition, KNN analyzes joint joints predicted by CNN and then classifies motion through a machine learning model.

The joint measuring unit 100 measures at least one joint joint of the user's body. The joint measurement unit 100 may be, for example, a camera including a 3D depth sensor. The joint measurement unit 100 may be an RGB camera in which one frame data is composed of RGB values. The joint measurement unit 100 may be a measuring device including at least one sensor for generating a signal by measuring the spatial position of the joint joint. The joint measuring unit 100 recognizes a joint in the user's body. At this time, it is preferable that the joint measuring unit 100 sets the angle of view of the camera to the maximum value in order to improve the accuracy of the posture detection.

The joint measurement unit 100 may include a first measurement unit 110 disposed in front of the user and a second measurement unit 120 disposed on either side of the user. The first measuring unit may be a camera module already formed by a manufacturer in the user's smart device. Alternatively, the first measuring unit may be a sensor module formed in the smart device. The second measurement unit may be a separate and independent RGB camera from the first measurement unit. Alternatively, the second measuring unit may be a measuring device including a sensor for measuring the spatial position of the joint joint.

Through this, the joint measurement unit 100 can more accurately measure and recognize the user's body. The joint measuring unit 100 is difficult to determine the posture from the front during exercise, such as lunge or squat, or when a part of the user's body is covered during exercise, the position of the joint joint through the second measuring unit can measure

At this time, an operation for calibrating the data measured by the first measurement unit and the second measurement unit is first performed. This may use the spatial position of the user's joint joint measured while the user moves freely for a certain period of time.

Meanwhile, according to the motion classified by the motion classifying unit 1100 in the present system, it is preferable that weights of data by the first measurement unit and the second measurement unit are different. Depending on the exercise classification, for example, the weight of data by the first measuring unit may be set higher than that of the second measuring unit. Depending on the exercise, the utilization of data according to the measurement location may be evaluated differently. In addition, measurement of some of the joint joints may be omitted during the exercise process depending on the measurement location. As a result, according to the exercise classification, for example, the first measuring unit may change the weight of the data in the form of K% and the second measuring unit (100-K)%.

In addition, the joint measuring unit 100 may measure the spatial position of the joint joint. Here, the joint measurement unit 100 may measure in the form of three-dimensional space coordinates.

The analysis module 1000 may use artificial intelligence as a statistical model for recognizing and classifying a user's exercise motion and calculating a motion coincidence rate for a standard motion. Artificial intelligence learns knowledge based on data and outputs results, and the analysis module 1000 can calculate the motion matching rate based on the input user's gender, body information, exercise type, and exercise motion. The analysis module 1000 may convert data measured through the joint measuring unit 100 into digital data through an AD converter. The analysis module 1000 includes an operation classification unit 1100 and a coincidence rate calculation unit 1200, and is electrically connected to each of them.

The motion classification unit 1100 uses an artificial intelligence algorithm to recognize a user's motion motion from data obtained by the joint measurement unit 100 and classify it into one of preset motions. This system can provide a high recognition rate for exercise motions in any direction by learning motion motions through data sets of various angles. Here, the exercise includes push-ups, squats, and lunges included in home training. The classification technology of the motion classification unit 1100 uses a general artificial intelligence algorithm, and thus a detailed description thereof will be omitted.

The coincidence rate calculation unit 1200 compares the unit motion of the exercise counted as 1 time with the standard motion to calculate the motion coincidence rate. The coincidence rate calculation unit 1200 calculates the motion coincidence rate using a first feature group, which is a set of angle data formed between a plurality of preset joint joints, and a second feature group, which is a set of distance data formed between joint joints. It is desirable to do

In addition, the present system may include at least one or more standard motions according to the classified exercise. Also, the standard operation may be performed in a plurality of standard postures. Also, the standard posture may be determined as a combination of at least one or more of the first feature group and at least one or more of the second feature group.

The calculation of the motion agreement rate may use the joint joint measured through the joint measurement unit 100. Standard motions and standard postures constituting them are set in advance according to the exercise. Meanwhile, the motion coincidence rate for standard motions may be, for example, an average value of motion coincidence rates for each standard posture.

The standard posture may be composed of the same number of images of a normal posture and an image of an incorrect posture, which are composed of at least one or more pre-input. Here, the normal posture image and the incorrect posture image are images of the same body part corresponding to each other. Such a posture image and an incorrect posture image may be limited to a part of the body. Meanwhile, different weights may be set for the normal posture image and the incorrect posture image in the process of determining whether or not they are standard postures according to the images.

Using the data on the joint joint measured by the joint measurement unit 100, it is possible to calculate the motion agreement rate between the user's exercise posture and the standard posture. At this time, the correct posture matching probability may be calculated by comparing the exercise posture and the static posture image. Here, the motion coincidence rate may be a straight posture coincidence probability. In addition, the same exercise posture can be compared with the wrong posture image to calculate the wrong posture matching probability. Meanwhile, unlike a normal posture image, an incorrect posture image may be composed of a plurality of incorrect posture images for the same exercise posture. As a result, a plurality of incorrect posture coincidence probabilities may be calculated.

Specifically, the coincidence rate calculation unit 1200 uses a first feature group, which is a set of angle data formed between a plurality of preset joint joints, and a second feature group, which is a set of distance data formed between joint joints, to determine the motion coincidence rate. can be calculated. The joint joint may be composed of, for example, 25 pieces. At this time, the joints are given numbers from 1 to 25, respectively, and can be distinguished through them.

Here, the angle data means the angle (a) formed between them with respect to the intermediate joint joint (A2) when the three joint joints (A1, A2, A3) are connected to each other in a straight line. For example, the angle data may be an angle (a) formed by the user's upper arm and lower arm based on the user's elbow. Here, the joint may be expressed as information in the form of three-dimensional space coordinates (x, y, z). The angle data may be, for example, an angle formed between the thigh and the calf based on the left knee. That is, the angle data may be preset according to the number of joint joints. Each of these angle data becomes a first feature group constituting an exercise posture.

On the other hand, the distance data means the distance (d) formed between the joint joints. For example, in space, it may be a straight line distance between the joint joint (A1) and the joint joint (A3). The first feature group and the second feature group may be specified in advance according to the classified exercise motion. That is, standard motions are included according to exercise motions, and standard postures constituting standard motions can be classified into a first feature group and a second feature group.

The motion coincidence rate may be calculated using an angle coincidence rate based on angle data constituting the first feature group related to the standard posture and a distance coincidence rate based on distance data constituting the second feature group. In one standard posture, the first feature group may include n pieces of angle data corresponding to n features. In this case, the angular coincidence rate may be calculated as an arithmetic average value of the first angular coincidence rate to the n-th angular coincidence rate.

The angle coincidence rate can be calculated using a, a reference range value for angle data, a reference value b for angle data preset in the present system, and c, angle data actually measured by the joint measuring unit 100. For example, the first angle coincidence rate is obtained by using a1, which is a reference range value for angle data, a reference value b1 for angle data preset in this system, and c1, which is angle data actually measured by the joint measuring unit 100. is calculated

In this system, the angular coincidence rate can be calculated, for example, as follows. The angular coincidence rate can be calculated using k ^{(bc)/a} (where k is a constant involved in the angular coincidence rate, 0.8), that is, an exponential function. Therefore, the first angular coincidence rate is k ^{(b1-c1)/a1} . Using this, the angular coincidence rate = {(first angular coincidence rate + second angular coincidence rate + ... + nth angular coincidence rate) / n}.

The distance coincidence rate can be calculated in the same way. In one standard posture, the second feature group may include m pieces of distance data corresponding to m features. In this case, the distance coincidence rate may be calculated as an arithmetic mean value of the first distance coincidence rate to the m-th distance coincidence rate. The distance coincidence rate may be calculated using d, which is a reference range value for distance data, a reference value f, which is preset for distance data in the present system, and e, which is actually measured distance data by the joint measuring unit 100. For example, the first distance coincidence rate is determined by using d1, which is a reference range value for distance data, reference value f1 for distance data preset in this system, and e1, which is distance data actually measured by the joint measuring unit 100. is calculated

The distance coincidence rate can be calculated using g ^{(fe)/d} (where g is a constant involved in the distance coincidence rate and is 0.9), that is, an exponential function. Therefore, the first distance coincidence rate is g ^{(f1-e1)/d1} . Using this, distance matching rate = {(first distance matching rate + second distance matching rate + ... + mth distance matching rate) / m}.

Meanwhile, the display unit 400 outputs a statistical graph for the user's exercise count, exercise time, and motion agreement rate. Here, the display unit 400 may be, for example, an output unit of a user's smart device. The display unit 400 may visually provide, for example, the number of workouts and the workout time according to a date in the form of a bar graph. In addition, the motion coincidence rate in each exercise motion can be expressed as a numerical value. In this system, a user's smart device may be installed with a dedicated application that works with this system. The dedicated application can statistically provide the user's exercise data with intuitive icon placement.

The system may further include a sensor unit 130. The sensor unit 130 is worn on both wrists of the user to measure the amount of change in angle of the elbow over time. The sensor unit 130 may be formed, for example, in a wearable device worn by a user. Such an angular change amount may be used as auxiliary data for calculating motion coincidence rate. That is, the sensor unit 130 can improve the accuracy of the motion matching rate for the standard motion through the amount of change in the angle of the elbow. At this time, the exercise preferably includes at least a push-up.

The sensor unit 130 may be, for example, a three-axis (X-axis, Y-axis, and Z-axis) acceleration sensor provided on a board having a Bluetooth 4.0 communication chip embedded therein. In addition, the sensor unit 130 may be a MEMS-based 9-axis acceleration sensor composed of a 3-axis acceleration sensor, a 3-axis gyro sensor, and a 3-axis geomagnetic sensor. The sensor unit 130 may periodically transmit and receive data with the user's smart device at regular time intervals.

The external server 600 communicates with the smart device. The external server 600 stores data defining standard motions, standard postures, and the like. Meanwhile, the external server 600 may communicate with a plurality of different smart devices. The smart device can transmit/receive data with the external server 600 while a dedicated application is running.

In addition, when the motion consistency rate for the user's exercise posture is less than 0.5, the system can provide the user with various contents related to the standard motion and standard posture. The present system can maximize exercise efficiency by providing feedback (for example, raise your arms, bend your knees more, etc.) for the corresponding standard posture when the probability of matching the wrong posture is 0.8 or higher.

The communication module 300 transmits and receives certain information, data, etc. through a wired or wireless network with at least one external server 600 set in advance. Also, the communication module 300 may receive data measured through the sensor unit 130 . The communication module 300 may transmit and receive radio signals to and from a base station on a mobile communication network established according to a technical standard or communication method for mobile communication.

The memory unit 500 may store various data related to algorithms for operation of the present system, raw data measured through the joint measuring unit 100, user interface related data, and the like. The memory unit 500 may include at least one storage medium among flash memory, card-type memory, and RAM. The memory unit 500 may store data related to the number of times of exercise, the exercise time, and the motion agreement rate of each user.

Although the above has been described with reference to the illustrated embodiments of the present invention, these are only examples, and those skilled in the art to which the present invention belongs can variously It will be apparent that other embodiments that are variations, modifications and equivalents are possible. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

100: joint measurement unit 1000: analysis module
1100: motion classification unit 1200: coincidence rate calculation unit
300: communication module 400: display unit
500: memory unit 130: sensor unit
600: external server

Claims (7)

As a system that provides home training through analysis of the user's exercise motion using artificial intelligence,
Joint measuring unit for measuring at least one joint of the user's body;
Using an artificial intelligence algorithm, motion classification unit for recognizing the user's exercise motion from the data by the joint measuring unit and classifying it into any one motion that is preset; and
A home training system using artificial intelligence comprising a; coincidence rate calculation unit that compares the unit motion of the exercise, which is counted as one number of times, with a standard motion to calculate a motion coincidence rate.
According to claim 1,
The joint measurement unit includes a first measurement unit disposed in front of the user and a second measurement unit disposed on one side of the user,
According to the exercise classified by the motion classification unit, the home training system using artificial intelligence to vary the weight of the data by the first measurement unit and the second measurement unit.
According to claim 1,
The joint measuring unit home training system using artificial intelligence to measure the spatial position of the joint joint.
According to claim 1,
The coincidence rate calculation unit calculates the motion coincidence rate using a first feature group, which is a set of angle data formed between a plurality of previously set joint joints, and a second feature group, which is a set of distance data formed between the joint joints. Home training system using artificial intelligence.
According to claim 4,
Standard motion is made up of a plurality of standard postures,
The standard posture is a home training system using artificial intelligence that is determined as a combination of at least one or more of the first feature group and at least one or more of the second feature group.
According to claim 1,
A home training system using artificial intelligence further comprising: a display unit that outputs a statistical graph of the user's exercise count, exercise time, and motion agreement rate.
According to claim 1,
A sensor unit worn on both wrists of the user to measure the amount of change in the angle of the elbow over time; further comprising;
The exercise is a home training system using artificial intelligence that includes at least a push-up.

Images