JP7596245B2 - Warm-up exercises evaluation device, warm-up exercises evaluation method, and warm-up exercises evaluation program - Google Patents

Description

The present invention relates to a warm-up exercises evaluation device, a warm-up exercises evaluation method, and a warm-up exercises evaluation program, and in particular to a warm-up exercises evaluation device, a warm-up exercises evaluation method, and a warm-up exercises evaluation program that support the physical condition management of a person to be evaluated based on images of the person's exercise movements.

At construction and transportation sites, workers perform warm-up exercises before starting work to loosen up their bodies, preventing accidents and injuries and helping to manage their physical condition. Workplace supervisors and managers can tell if their subordinates are feeling unwell or are under psychological stress by the way they perform their warm-up exercises, and can speak up if they are concerned, preventing trouble from occurring.

However, a human's judgment of whether there is anything concerning about the movements of workers' warm-up exercises is based on intuition, and does not necessarily detect abnormalities. It is also not realistic for a supervisor to simultaneously observe the movements of all workers performing the exercises and discover any abnormalities. Therefore, an evaluation method that does not rely on humans is required.

As a method for evaluating a person's body movements by comparing them with model movements, for example, Patent Document 1 proposes a technology that detects a user's golf swing with a motion sensor such as an angular velocity sensor mounted inside the housing of a mobile terminal device. Furthermore, Patent Document 2 proposes a technology that enables a practitioner to judge for himself or herself whether or not he or she is able to properly imitate the movements of a model skilled practitioner.

These technologies only measure the difference between a given movement and a model, and so have the problem of not being able to detect abnormalities in the normal movements of an individual who is performing an exercise while referring to a model.

JP 2010-68947 A JP 2014-57800 A

The present invention has been made in consideration of the above-mentioned problems with the conventional art, and the object of the present invention is to provide a warm-up exercises evaluation device, a warm-up exercises evaluation method, and a warm-up exercises evaluation program that support the physical condition management of an evaluation subject who performs warm-up exercises according to a model video.

The warm-up exercises evaluation device according to one aspect of the present invention, which has been made to achieve the above object, is a warm-up exercises evaluation device that supports the physical condition management of an evaluation subject who performs warm-up exercises in accordance with a model video, and includes a storage unit that stores video data of the warm-up exercises performed by a model performer as model video data, a model video playback unit that plays back the model video data and displays it on a predetermined display means, a body movement acquisition unit that acquires, via a predetermined imaging means, body movement images of the evaluation subject who performs the warm-up exercises in accordance with the displayed model video and recognizes the positions of each part of the body of the evaluation subject in chronological order from the acquired body movement images, comparison reference data that has been set in advance for a plurality of playback positions in the model video data, and the body positions of the evaluation subject recognized in chronological order corresponding to the plurality of playback positions. The apparatus includes a body movement judging section which compares the position data of each body part of the model performer with the position data of each body part of the model performer to calculate the degree of agreement, a body movement scoring section which tallys up the calculated degrees of agreement, scores the body movements of the person being evaluated, and causes the scoring result to be displayed on the specified display means as an exercises evaluation score, and a control section which causes the model video playback section, the body movement acquisition section, the body movement judging section, and the body movement scoring section to function, and which controls the warm-up exercises evaluation device as a whole, wherein the comparison reference data is the position data of each body part of the model performer at a specific playback position of the model video data, and the body movement scoring section stores the exercises evaluation score of the person being evaluated in the storage section as learning data for setting comparison reference data for evaluating the physical condition of the person being evaluated.

The physical movement scoring unit may store in the memory unit the position data of each part of the subject's body at the multiple playback positions used in the scoring, together with the subject's exercise evaluation score, as learning data for setting comparison standard data for evaluating the subject's individual physical condition.
The warm-up exercises evaluation device further includes an evaluation standard data generation unit that generates comparison standard data for evaluating the physical condition of the individual person to be evaluated, wherein the evaluation standard data generation unit uses the learning data and exercise evaluation scores stored in the memory unit to calculate standard positions of each part of the person to be evaluated at each of the multiple playback positions through machine learning and generates comparison standard data for evaluating the physical condition of the individual person to be evaluated based on the calculated standard positions of each part of the body, the physical movement determination unit compares the position data of each part of the person to be evaluated recognized by the physical movement acquisition unit with the comparison standard data for evaluating the physical condition of the individual person to be evaluated to calculate a degree of agreement, and the physical movement scoring unit causes the scoring result using the comparison standard data for evaluating the physical condition of the individual person to be evaluated to be displayed on the specified display means as a physical condition evaluation score.

A method for evaluating warm-up exercises according to one aspect of the present invention, which has been made to achieve the above object, is a method for evaluating warm-up exercises that supports the physical condition management of an evaluation subject who performs warm-up exercises in accordance with a model video, and includes a storage unit that stores video data of the warm-up exercises performed by a model performer as model video data, a model video playback unit that plays back the model video data and displays it on a predetermined display means, and a body movement acquisition unit that acquires body movement images of the evaluation subject who performs the warm-up exercises in accordance with the displayed model video, and recognizes the position of each part of the body of the evaluation subject in chronological order from the acquired body movement images. a physical movement judging section which compares comparison reference data previously set for a plurality of playback positions in the model video data with position data of each part of the body of the person to be evaluated which is recognized in the time series corresponding to the plurality of playback positions, and calculates a degree of agreement; a physical movement scoring section which tallies up the calculated degrees of agreement, scores the physical movements of the person to be evaluated, and causes the scoring result to be displayed on the predetermined display means as an exercises evaluation score; and a control section which controls the above-mentioned sections and the device as a whole, a step of playing back model video data and displaying it on the predetermined display means; a step of having the body movement acquisition unit acquire body movement images of the person to be evaluated performing the warm-up exercises in accordance with the displayed model video and recognize the positions of each part of the body of the person to be evaluated in time series from the acquired body movement images; and a step of having the body movement determination unit compare comparison reference data previously set for a plurality of playback positions in the model video data with the position data of each part of the body of the person to be evaluated recognized in time series corresponding to the plurality of playback positions, respectively, to calculate a degree of agreement. and causing the physical action scoring unit to tally up the calculated degrees of agreement, score the physical action of the person being evaluated, and display the scoring results as an exercise evaluation score on the display means, wherein the comparison reference data is position data of each part of the body of the model performer at a specific playback position of the model video data, and the control unit further includes a step of causing the physical action scoring unit to store the exercise evaluation score of the person being evaluated in the memory unit as learning data for setting comparison reference data for evaluating the physical condition of the person being evaluated.

The present invention provides a program for evaluating warm-up exercises, which has been developed to achieve the above-mentioned object, and includes a storage unit that stores video data of the warm-up exercises performed by a model performer as model video data, a model video playback unit that plays back the model video data and displays it on a predetermined display means, a body movement acquisition unit that acquires, via a predetermined imaging means, body movement images of a subject performing the warm-up exercises in accordance with the displayed model video, and recognizes the position of each part of the subject's body in chronological order from the acquired body movement images, and a body movement acquisition unit that acquires, in advance, a plurality of playback positions in the model video data. a physical action judging section for comparing the comparison reference data set by the user with the position data of each part of the body of the person to be evaluated that is recognized in the time series corresponding to the plurality of playback positions to calculate a degree of agreement; a physical action scoring section for aggregating the calculated degrees of agreement, scoring the physical action of the person to be evaluated, and displaying the scoring result on the predetermined display means as an exercise evaluation score; and a control section for controlling the above sections and the device as a whole, the warm-up exercises evaluation device for calculating and scoring the degree of agreement of the physical actions of the person to be evaluated who performs the warm-up exercises in accordance with the example video. A program for evaluating warm-up exercises that executes a process, the program being characterized in that the control unit of the warm-up exercises evaluation device executes the steps of causing the model video playback unit to play the model video data and display it on the specified display means, causing the body movement acquisition unit to acquire body movement images of the person being evaluated performing the warm-up exercises in accordance with the displayed model video and recognize the positions of each part of the body of the person being evaluated in time series from the acquired body movement images, causing the body movement determination unit to compare comparison reference data previously set for a plurality of playback positions in the model video data with the position data of each part of the body of the person being evaluated that corresponds to the playback positions and recognize the position data in time series to calculate the degree of agreement, causing the body movement scoring unit to tally up the calculated degrees of agreement, score the body movements of the person being evaluated, and display the scoring results as an exercise evaluation score on the display means, and causing the body movement scoring unit to store the exercise evaluation score of the person being evaluated in the storage unit as learning data for setting comparison reference data for evaluating the physical condition of the person being evaluated.

According to the present invention, the physical condition of the person to be evaluated can be confirmed based on the change in the evaluation score given by the warm-up exercises evaluation device.
The supervisor can use the evaluation score to confirm that the warm-up exercises are being performed and that the exercises being performed are the same as the usual ones. This can also be used for warm-up exercises performed remotely.
In addition, by collecting personal authentication in association with evaluation data on warm-up exercises, it is possible to obtain health management data such as the fact that people who receive high evaluations for their warm-up exercises are healthy both physically and mentally, less likely to get injured, and less likely to develop injuries.

1 is a block diagram showing a configuration of a warm-up exercises evaluation device according to one embodiment of the present invention. FIG. 11 is a diagram showing an example of position data of a body part recognized by modeling the body part according to the present embodiment. 4 is a flowchart showing the operation of the warm-up exercises evaluation device according to one embodiment of the present invention. 1A to 1C are diagrams showing example images corresponding to a series of actions included in the model video data. FIG. 13 shows a data sequence of modeled body movements of an evaluation subject synchronized with an example video. 10 is a flowchart showing the operation of performing a physical condition evaluation for each evaluation subject by the warm-up exercises evaluation device according to one embodiment of the present invention. FIG. 13 is a diagram showing an example of learning data used to generate comparison reference data for evaluating a physical condition.

Specific examples of embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a warm-up exercises evaluation device according to one embodiment of the present invention. As shown in FIG. 1, the warm-up exercises evaluation device 10 according to one embodiment of the present invention includes a control unit 11 having a central processing unit (CPU) (not shown), ROM, RAM, etc., which executes a predetermined program and performs various data processing and system control, a memory unit 12 which stores various programs and data, a display unit 13 consisting of a liquid crystal display or an organic EL display, etc., an input unit 14 which accepts user commands and input data from the outside, an imaging unit 15 consisting of a camera, an output unit 16 which serves as an interface for outputting data to the outside, and a communication unit 17. The warm-up exercises evaluation device 10 may be configured as a computer, a smartphone, a tablet terminal, or other mobile information terminal equipped with the above-mentioned components.

The warm-up exercises evaluation device 10 includes a model video playback unit 111, a body movement acquisition unit 113, a body movement determination unit 115, a body movement scoring unit 117, and an evaluation criteria data generation unit 119, which function when the control unit 11 reads and executes an OS (Operating System), a warm-up exercises evaluation program, and various data stored in the ROM, RAM, or storage unit 12.

The model video playback unit 111 plays model video data of a model performer performing choreography to the music of Radio Calisthenics (registered trademark) as a warm-up exercise, for example, and displays it on the display unit 13. The model video data is data that defines the body movements to the music, and is stored in advance in the storage unit 12 in a predetermined video file format that combines the music and the model image. Examples of video file formats include, but are not limited to, AVI, MPEG2, MP4, etc.

The body motion acquisition unit 113 recognizes the position of each part of the body in time series from the image of the subject captured by the imaging unit 15, and recognizes the relative positions of other parts based on the center of gravity of the body. This makes it possible to perform scoring that is less affected by the shooting position, the height difference of the subject, and the difference in physique. In this case, the body motion acquisition unit 113 may recognize the body parts by modeling them. As a body model, a model in which the positions of the main joints are estimated and the body parts are classified may be used. For this reason, for example, motion capture can be easily performed by using "YOLO v3" as an algorithm for detecting people in an image and applying "PoseNet" that can measure the positions of each part of a person's face and body in real time. By using such technology, the movement of the body model (skeletal model) corresponding to the gymnastics movement of the subject can be obtained in real time. Note that this embodiment does not require the use of YOLO and PoseNet, and other motion capture technologies may be used.

Figure 2 is a diagram showing an example of position data of body parts recognized by modeling body parts according to this embodiment. In this embodiment, the body part acquisition unit 113 classifies and recognizes body parts into 17 locations. As shown in Table 1 of Figure 2, in this embodiment, the position data of the modeled body parts is composed of the body part and position information on the X, Y, and Z coordinates (with the body's center of gravity as the origin) for each row number (ID).

The body movement determination unit 115 chronologically compares the positions of each body part acquired by the body movement acquisition unit 113 with "data specifying model reference movements (hereinafter referred to as comparison reference data)" previously stored in the storage unit 12, and determines the degree of agreement between them. The comparison reference data is data that specifies the positional relationship between the body's center of gravity and a part (predetermined part) that has a model pose for a model performer's pose (called a model pose) at a specific playback position (called a judgment point) selected from a series of movements included in the model video data, and is set by selecting multiple points in time (playback positions) from the entire model video data.
The body movement determination unit 115 determines the degree of match by comparing the comparison reference data with the positions of the body parts recognized by the body movement acquisition unit 113. Details of the method of determining the degree of match will be described later with reference to Fig. 4. Note that the determination of the degree of match is not limited to the method based on the comparison of positional relationships, and a method of calculation based on the tendency (feature amount) of the position data of each body part using machine learning or deep learning may be applied.

The physical movement scoring unit 117 receives the judgment result of the degree of agreement at each judgment point from the physical movement judgment unit 115, tallies up the degrees of agreement at each judgment point, and scores the physical movement (gymnastics movement) of the person to be evaluated that corresponds to the model video. For example, the scoring is performed by accumulating the degrees of agreement at each judgment point or calculating the average value.
The physical movement scoring unit 117 displays the scoring results as an exercise evaluation score on the display unit 13, and stores in the memory unit 12 the positional data of each part of the body of the person being evaluated at each judgment point in association with an identifier (e.g., a user registration number) assigned to the person being evaluated.

The evaluation reference data generating unit 119 uses the position data of each part of the body of the person being evaluated at each judgment point set in the model video of the warm-up exercises, which is stored in the memory unit 12 by the body movement scoring unit 117, as learning data to generate comparison reference data for physical condition evaluation for each person being evaluated through machine learning or deep learning. The method of generating comparison reference data for physical condition evaluation will be described in detail later.

The operation of the warm-up exercises evaluation device according to one embodiment of the present invention will be described below with reference to Figures 1 and 3.

Figure 3 is a flowchart showing the operation of a warm-up exercise evaluation device according to one embodiment of the present invention.

In step S201 of FIG. 3, the model video playback unit 111 plays a model video of the warm-up exercises to be viewed by the person to be evaluated. The person to be evaluated refers to this model video and performs corresponding body movements in accordance with the model movements (choreography). Note that the model video is composed of images in which the body movements to be performed by the person to be evaluated move in sync with the music.

In synchronization with the start of playback of the model video, the body movement acquisition unit 113 acquires image data of the body movements of the person being evaluated from the imaging unit 15 (step S203). The acquired image data of the body movements is acquired by the body movement acquisition unit 113 as, for example, motion capture data of a skeletal model. The acquired data is hereinafter referred to as body movement data. The above-mentioned "YOLO" and "PoseNet" can be used as technologies for acquiring motion capture data. Note that the means for acquiring body movement data in this embodiment is not limited to these.

In step S205, the body movement determination unit 115 compares the model video data with the acquired body movement data. In this comparison, the body movement data is compared with comparison reference data of the model pose at a determination point that is preset in the model video data, to determine the degree of agreement.

In step S207, the physical movement scoring unit 117 tallies the degree of agreement at each judgment point and scores the physical movement of the person being evaluated. Note that the physical movement scoring unit 117 can score at any time, even during playback of the video, and can display the scoring results on the display unit 13.

In step S209, the body movement scoring unit 117 stores the position data of each part of the body of the person being evaluated at each judgment point in the storage unit 12 in association with the identifier given to the person being evaluated. When storing the data, an index or serial number may be further added to distinguish between data from warm-up exercises performed in the past and data from the warm-up exercises being performed this time.

Next, we will explain the model poses that are set by the warm-up exercises evaluation device according to one embodiment of the present invention.

Figure 4 shows example images corresponding to a series of movements in the warm-up exercises included in the model video data.

The playback position counter axis 310 shown in FIG. 4 is a time axis for identifying the playback position during video playback, and the numbers displayed above the playback position counter axis 310 indicate the frame number 320 from the beginning (start) of the video being played.

The model poses (330a, 330b) are poses corresponding to specific movements of interest, selected from a series of movements included in the model video, and are set in advance for multiple playback positions (judgment points). Figure 4 shows an example of model poses corresponding to the set playback positions (judgment points). For example, a skeletal model made up of a combination of joints is used for the model poses. The playback positions corresponding to each model pose (330a, 330b) in the model video are specified by frame number 320.

Figure 5 shows a data sequence of modeled body movements of the subject being evaluated, synchronized with the example video.

The physical movement data 400 of the person being evaluated is acquired by the physical movement acquisition unit 113 at a predetermined time interval (in milliseconds). Image data of the physical movements of the person being evaluated captured by the imaging unit 15 is recognized, for example, in the form of a skeletal model. Then, to make a comparison within the time period before and after the judgment point (playback position) corresponding to each model pose (330a, 330b), multiple pieces of physical movement data acquired during the time interval Δt before and after the judgment point are selected as comparison candidates.

For example, for model pose 330b in FIG. 5, the body movement data (442, 443, 444) acquired during the interval Δt between video playback positions for comparison are selected. Then, the body movement determination unit 115 calculates the degree of agreement between the comparison reference data corresponding to model pose 330b and body movement data 442 from among the multiple acquired body movement data.

In this embodiment, the body movement determination unit 115 calculates the degree of match between skeletal models. For example, the distance between corresponding joints and the difference in the bone angles at the joints are calculated, and the degree of match is calculated by summing the absolute values of the differences. As an example, if the reciprocal of the summed absolute values is defined as the degree of match, the greater the value of the degree of match, the higher the degree of match. As another example, a system may be used in which points are assigned to each body part according to the magnitude of the difference (small is "-1", the same is "0", and large is "+1"), and the sum of these points is used to determine the degree of match.

The physical movement determination unit 115 similarly calculates the degree of match between the remaining physical movement data (443, 444) and the comparison reference data corresponding to the model pose 330b. The physical movement determination unit 115 then adopts the highest degree of match among the degrees of match corresponding to each of the physical movement data (442, 443, 444) as the degree of match with the model pose 330b. The physical movement data that was adopted and the video playback position of the adopted physical movement data are stored in the storage unit 12 in association with an identifier (user registration number) previously assigned to the person being evaluated.

In this way, the degree of agreement is calculated for each of the multiple model poses. Note that the calculation of the degree of agreement is not limited to the above method, and a method of calculating the degree of agreement from the trends of body movement data using machine learning or deep learning may also be applied.

After starting playback of the model video, the body movement scoring unit 117 may sequentially add up the degrees of agreement corresponding to the model poses or calculate an average value, and display this on the display unit 13 as an intermediate evaluation score. Then, when playback of the video ends, the final evaluation score (gymnastics evaluation score) is calculated.

As explained above, the warm-up exercises evaluation device according to one embodiment of the present invention has a function of evaluating the degree of agreement between model poses at multiple playback positions (judgment points) preset in a model video of the warm-up exercises and the physical movements of the person being evaluated, and furthermore, it uses data on the physical movements of the person being evaluated during past warm-up exercises performed in accordance with the model video to generate comparison reference data for evaluating the physical condition of the individual person being evaluated, and compares and scores the generated comparison reference data for evaluating physical condition in chronological order with the data on the physical movements of the warm-up exercises currently being performed, thereby making it possible to check the physical condition of the person being evaluated.

Figure 6 is a flowchart showing a method for evaluating the physical condition of a person being evaluated using a warm-up exercises evaluation device according to one embodiment of the present invention.

In step S501 of FIG. 6, when a user registration number is input via the input unit 14, the control unit 11 of the warm-up exercises evaluation device 10 reads out from the memory unit 12 the comparison reference data for evaluating the physical condition of the person being evaluated that is stored in association with the user registration number (step S502). Note that the comparison reference data for evaluating the physical condition that is read out at the time of first use consists of position data of each part of the body at each judgment point of the person being evaluated that was stored when the person being evaluated registered as a user, but as described below, learning is carried out with each use, and the data becomes more reflective of the person being evaluated's own normal condition.

After that, when an instruction to start playing the model video is given via the input unit 14, the model video playback unit 111 plays the model video of the warm-up exercises to be viewed by the person to be evaluated (step S503). The person to be evaluated performs the corresponding body movements in accordance with the model while referring to the model video.

In synchronization with the start of playback of the model video, the physical movement acquisition unit 113 acquires image data of the physical movement of the person being evaluated from the imaging unit 15 (step S504). The acquired image data of the physical movement is acquired by the physical movement acquisition unit 113 as, for example, motion capture data of a skeletal model, and is transmitted to the physical movement determination unit 115 as physical movement data.

In step S505, the body movement determination section 115 compares the acquired body movement data with comparison reference data of a model pose at a determination point set in advance in the model video data, and determines the degree of agreement with the model pose.
At the same time, the body movement determination section 115 compares the acquired body movement data with the comparison reference data for evaluating the physical condition of the individual to be evaluated, and determines the degree of agreement with the comparison reference data for evaluating the physical condition.

In step S506, the body movement scoring section 117 tallies up the degree of agreement with the model pose at each judgment point, and scores the body movement that matches the warm-up exercises currently being performed by the person to be evaluated.
At the same time, the body action scoring section 117 tallyes up the degree of agreement with the comparison standard data for evaluating the physical condition of the individual to be evaluated at each judgment point, and assigns an evaluation score representing the physical condition of the individual to be evaluated.

At this time, the physical movement scoring unit 117 can score the currently executed physical movement against the model pose at any time even during video playback, and can display the scoring results on the display unit 13. Meanwhile, the physical movement scoring unit 117 scores the currently executed physical movement against the comparison reference data for physical condition evaluation at any time during video playback, but displays the scoring results on the display unit 13 after video playback has ended. This makes it possible to eliminate errors and cheating that may occur when the person being evaluated looks at the score in the middle of warm-up exercises and intentionally corrects their physical movement.

In step S507, the body movement scoring unit 117 stores in the memory unit 12 the position data of the body parts of the person being evaluated at each judgment point in association with the identifier of each judgment point (e.g., frame number) and the identifier (user registration number) assigned to the person being evaluated. When storing the data, it is preferable to further add an index or serial number to distinguish between data from the currently performed warm-up exercises and data from previously performed warm-up exercises.

The above-mentioned steps S504 to S506 are executed sequentially for a plurality of preset playback positions (judgment points), and step S507 may be executed sequentially as the scoring is performed at each judgment point, or may be executed collectively after the scoring at the final judgment point is completed. However, after the scoring at the final judgment point is completed, the position data of each part of the body of the person being evaluated at each of the plurality of playback positions (judgment points) used for the scoring is stored in the storage unit 12 as data in a format associated with the above-mentioned identifiers (user registration number and serial number of the execution time) together with the exercises evaluation score of the person being evaluated. The above-mentioned stored data is used as learning data for setting comparison standard data for the physical condition evaluation of the person being evaluated.

In this way, the warm-up exercises evaluation device and warm-up exercises evaluation method according to one embodiment of the present invention can provide data that can be used to judge the physical condition of an individual by comparing the warm-up exercises currently being performed by the individual with the normal physical movements of the individual during previous warm-up exercises and quantifying the deviation (calculating the degree of agreement), in addition to scoring the performance of the warm-up exercises currently being performed by the individual (whether the physical movements are faithful to the model).

This article describes a method for generating comparison reference data for evaluating physical condition used in a warm-up exercises evaluation device and warm-up exercises evaluation method according to one embodiment of the present invention.

The comparison reference data for evaluating physical condition is generated by using a collection of data on the physical movements of multiple warm-up exercises previously performed by the person being evaluated as learning data at multiple playback positions (judgment points) that are preset in the model video of warm-up exercises.

Figure 7 shows an example of learning data used to generate comparison reference data for health evaluation.

The warm-up exercises evaluation device may be pre-programmed to execute a step of generating or updating comparison reference data for physical condition evaluation after the series of steps related to warm-up exercises evaluation corresponding to the video playback shown in FIG. 6 is completed. At this time, the display unit 13 may be caused to display that learning is in progress, although this is not shown.

When the process of generating or updating comparison standard data for evaluating physical condition begins, the warm-up exercises evaluation device causes the evaluation standard data generation unit 119 to execute the steps of: using the learning data and exercise evaluation scores stored in the memory unit 12 to calculate the standard positions of each part of the subject's own body for each of a plurality of playback positions (judgment points) by machine learning or deep learning; and generating comparison standard data for evaluating the subject's personal physical condition based on the calculated data on the standard positions of each part of the body.

Specifically, each time the person to be evaluated performs a new warm-up exercise, the evaluation reference data generating unit 119 reads out the learning data for that person stored in the memory unit 12, calculates the standard positions of each part of the person to be evaluated for each of a plurality of judgment points combining the present and past (for example, body movement data 443, 443', 443'' shown in FIG. 7) by machine learning or deep learning, and generates comparison reference data for evaluating the physical condition of the individual person to be evaluated based on the calculated standard positions of each part of the body. At this time, the exercise evaluation score for each execution may be used for weighting. A predetermined weighting coefficient may be added to the positions of each part of the body at each judgment point in the execution execution with a high exercise evaluation score.
Therefore, the comparison standard data for evaluating the physical condition at each judgment point is the data of the physical movements when the warm-up exercises were performed well (in physical condition) from the current to the past, and this data is reflected in the comparison standard data for evaluating the physical condition of the person being evaluated.

The above describes in detail an embodiment of the present invention with reference to the drawings, but the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the technical scope of the present invention.

REFERENCE SIGNS LIST 10 Warm-up exercises evaluation device 11 Control unit 12 Memory unit 13 Display unit 14 Input unit 15 Imaging unit 16 Output unit 17 Communication unit 111 Model video playback unit 113 Body movement acquisition unit 115 Body movement determination unit 117 Body movement scoring unit 119 Evaluation reference data generation unit

Claims (3)

A warm-up exercises evaluation device that supports physical condition management of an evaluation subject who performs warm-up exercises in accordance with a model video,
A storage unit that stores video data of the warm-up exercises performed by a model performer as model video data;
a model video playback unit for playing back the model video data and displaying it on a predetermined display means;
a body motion acquisition unit that acquires, via a predetermined imaging means, body motion images of the person to be evaluated performing the warm-up exercises in accordance with the displayed example video, and recognizes the positions of each part of the body of the person to be evaluated in chronological order from the acquired body motion images;
a body movement determination unit that calculates a degree of agreement between comparison reference data that has been set in advance for a plurality of playback positions in the example video data and position data of each part of the body of the subject that corresponds to the plurality of playback positions and that is recognized in the time series;
a body movement scoring unit that aggregates the calculated degrees of agreement, scores the body movement of the person to be evaluated, and displays the scoring result on the predetermined display means as an exercise evaluation score for the model video;
an evaluation standard data generating unit that generates comparison standard data for evaluating the physical condition of the individual subject;
a control unit that causes the model video playback unit, the body movement acquisition unit, the body movement judging unit, the body movement scoring unit, and the evaluation reference data generation unit to function, and also controls the warm-up exercises evaluation device as a whole,
the comparison reference data is position data of each part of the body of the model performer at a specific playback position of the model video data;
the body movement scoring unit stores in the storage unit position data of each part of the body of the subject at the plurality of playback positions used in scoring the exercise evaluation score of the subject as learning data for setting comparison standard data for evaluating the physical condition of the subject;
the evaluation reference data generation unit reads out the learning data of the person to be evaluated stored in the storage unit each time the person to be evaluated newly performs the warm-up exercises, and executes machine learning using position data of each part of the body of the person to be evaluated at the multiple playback positions in each of the warm-up exercises previously performed by the person to be evaluated and the newly performed warm-up exercises, which are included in the learning data, as input, to calculate a tendency (feature amount) indicated by the positions of each part of the body of the person to be evaluated at each of the multiple playback positions, and generates comparison reference data for evaluating the personal physical condition of the person to be evaluated, which specifies a standard position of each part of the body of the person to be evaluated at each of the multiple playback positions, based on the tendency (feature amount) calculated by the machine learning ;
the physical movement determination unit further compares the position data of each part of the body of the subject recognized by the physical movement acquisition unit with comparison reference data for evaluating the physical condition of the subject, and calculates a degree of agreement;
the physical action scoring unit causes the predetermined display means to display a score of the physical condition of the evaluation subject on the predetermined display means by tallying up and scoring the degree of agreement with the comparison standard data for evaluating the physical condition of the evaluation subject ;
A warm-up exercises evaluation device , characterized in that the tendency (feature amount) is composed of body parts and positions that reflect movements unique to the individual person to be evaluated . A warm-up exercise evaluation method for supporting a physical condition management of an evaluation subject who performs warm-up exercises in accordance with a model video, comprising:
a physical movement acquisition unit that acquires physical movement images of a person to be evaluated performing the warm-up exercises in accordance with the displayed model video and recognizes the positions of each part of the person to be evaluated in time series from the acquired physical movement images; a physical movement determination unit that calculates a degree of agreement between comparison reference data previously set for a plurality of playback positions in the model video data and the position data of each part of the person to be evaluated that corresponds to the plurality of playback positions and that is recognized in time series; a physical movement scoring unit that calculates the degree of agreement to score the physical movements of the person to be evaluated and displays the scoring result on the predetermined display means as an exercises evaluation score for the model video; an evaluation reference data generation unit that generates comparison reference data for evaluating the physical condition of each person to be evaluated; and a control unit that controls the above units and the device as a whole.
The control unit:
causing the model video playback unit to play back the model video data and display it on the predetermined display means;
a step of causing the body motion acquisition unit to acquire body motion images of the person to be evaluated performing the warm-up exercises in accordance with the displayed model video, and recognizing the positions of each part of the body of the person to be evaluated in chronological order from the acquired body motion images;
a step of causing the body movement determination unit to compare comparison reference data previously set for a plurality of playback positions in the model video data with position data of each part of the body of the subject recognized in the time series corresponding to the plurality of playback positions, and to calculate a degree of agreement;
and causing the body movement scoring unit to collect the calculated degrees of agreement, score the body movement of the person to be evaluated, and display the scoring result on the display means as an exercise evaluation score for the model video.
Here, the comparison reference data is position data of each part of the body of the model performer at a specific playback position of the model video data,
The control unit:
causing the body action scoring unit to store in the storage unit position data of each part of the body of the subject at the plurality of playback positions, which was used in scoring the exercise evaluation score of the subject, as learning data for setting comparison standard data for evaluating the physical condition of the subject;
a step of causing the evaluation reference data generation unit to read out the learning data of the person to be evaluated stored in the storage unit each time the person to be evaluated newly performs the warm-up exercises, and to execute machine learning using position data of each part of the body of the person to be evaluated at the multiple playback positions in each of the warm-up exercises previously performed by the person to be evaluated and the newly performed warm-up exercises, which are included in the learning data, to calculate a tendency (feature amount) indicated by the position of each part of the body of the person to be evaluated at each of the multiple playback positions, and to generate comparison reference data for evaluating the physical condition of the person to be evaluated, which specifies a standard position of each part of the body of the person to be evaluated at each of the multiple playback positions, based on the tendency (feature amount) calculated by the machine learning ;
a step of causing the body movement determination unit to compare position data of each part of the body of the subject recognized by the body movement acquisition unit with comparison reference data for evaluating the physical condition of the subject, and calculate a degree of agreement;
and causing the physical action scoring unit to display, on the predetermined display means, a result of tallying up and scoring the degree of agreement with the comparison standard data for evaluating the physical condition of the individual of the evaluation subject as a physical condition evaluation score of the evaluation subject ,
A method for evaluating warm-up exercises , characterized in that the tendency (feature amount) is composed of body parts and positions that reflect movements unique to the individual subject to be evaluated . a storage section for storing video data of a model performer 's warm-up exercises as model video data; a model video playback section for playing back the model video data and displaying it on a predetermined display means; a body motion acquisition section for acquiring, via a predetermined imaging means, body motion images of a subject performing the warm-up exercises in accordance with the displayed model video and recognizing the positions of each part of the subject's body in time series from the acquired body motion images; comparison reference data previously set for a plurality of playback positions in the model video data and the positions of each part of the subject's body recognized in time series corresponding to the plurality of playback positions a body movement judging section which compares the body movement data of the person to be evaluated with the position data of the person to be evaluated to calculate a degree of agreement, a body movement scoring section which counts up the calculated degrees of agreement, scores the body movement of the person to be evaluated, and displays the scoring result on the predetermined display means as an exercises evaluation score corresponding to the model video, an evaluation standard data generating section which generates comparison standard data for evaluating the physical condition of each of the people to be evaluated, and a control section which controls the above sections and the device in general, said warm-up exercises evaluation program causing a warm-up exercises evaluation device to execute a process of calculating and scoring the degree of agreement of the body movements of the person to be evaluated who are performing warm-up exercises in accordance with the model video,
The control unit of the warm-up exercises evaluation device
causing the model video playback unit to play back the model video data and display it on the predetermined display means;
a step of causing the body motion acquisition unit to acquire body motion images of the person to be evaluated performing the warm-up exercises in accordance with the displayed model video, and recognizing the positions of each part of the body of the person to be evaluated in chronological order from the acquired body motion images;
a step of causing the body movement determination unit to compare comparison reference data previously set for a plurality of playback positions in the model video data with position data of each part of the body of the subject recognized in the time series corresponding to the playback positions, and calculate a degree of agreement;
and causing the physical action scoring unit to collect the calculated degrees of agreement, score the physical action of the person to be evaluated, and display the scoring result on the display means as an exercise evaluation score for the model video.
causing the body action scoring unit to store in the storage unit position data of each part of the body of the subject at the plurality of playback positions, which was used in scoring the exercise evaluation score of the subject, as learning data for setting comparison standard data for evaluating the physical condition of the subject;
a step of causing the evaluation reference data generation unit to read out the learning data of the person to be evaluated stored in the storage unit each time the person to be evaluated newly performs the warm-up exercises, and to execute machine learning using position data of each part of the body of the person to be evaluated at the multiple playback positions in each of the warm-up exercises previously performed by the person to be evaluated and the newly performed warm-up exercises, which are included in the learning data, to calculate a tendency (feature amount) indicated by the position of each part of the body of the person to be evaluated at each of the multiple playback positions, and to generate comparison reference data for evaluating the physical condition of the person to be evaluated, which specifies a standard position of each part of the body of the person to be evaluated at each of the multiple playback positions, based on the tendency (feature amount) calculated by the machine learning ;
a step of causing the body movement determination unit to compare position data of each part of the body of the subject recognized by the body movement acquisition unit with comparison reference data for evaluating the physical condition of the subject, and calculate a degree of agreement;
and causing the body action scoring unit to display, on the predetermined display means, a score of the physical condition of the subject, the score of the degree of agreement with the comparison reference data for evaluating the physical condition of the subject , and
A warm-up exercises evaluation program , characterized in that the tendency (feature amount) is composed of body parts and positions that reflect movements unique to the individual person to be evaluated .

Images