JP6677319B2 - Sports motion analysis support system, method and program - Google Patents

Description

  The present invention relates to a sports motion analysis support system, a sports motion analysis support method, and a sports motion analysis support program that support motion analysis in sports.

  Motion capture is known as a technique that can be used for analyzing motion in sports.

  In addition, there is application software that displays a trajectory of a specific part of a body (for example, an ankle) based on a moving image obtained by imaging a person playing sports. In addition, there is application software that displays a difference between a model form and a sports form.

  Patent Document 1 describes that a swing stage and a main position of golf are specified from a measured value of a frame difference of an image using a rule-based model. Further, Patent Document 1 describes that in this case, a hidden Markov model, a state space model, a finite state machine, a regression method, a support vector machine, a neural network, a fuzzy theory, or the like may be used.

  In Patent Document 2, motion image data of 10 trials is collected, five trials are used as learning data, parameters of a hidden Markov model are estimated, and a recognition experiment is performed using the remaining five trials as test data. Is described. Patent Literature 2 describes that it has been found that a discrete cosine transform coefficient of a relatively low frequency component is effective as a feature amount for image recognition of a human motion.

JP 2014-521139 A JP-A-10-013832

  With the various application software described above, a sporter can analyze his or her own form.

  Here, the form and the result of the operation including the form do not always correspond completely. For example, even if a person's form is good, the condition may not be good and the person may not have performed well. Also, for example, even if a person's form is good, it may happen that good results are not obtained due to external factors such as rain and wind.

  However, statistically, there is a tendency that a good result is easily obtained with such a form, and a tendency that a ball is likely to fly in a specific direction with such a form.

  In addition, the inventor of the present invention considered that a form in a certain time period during which a person is performing an operation is an important form that has a great influence on the result.

  It should be noted that an example of the result is a numerical value indicating a performance of a sport. An example of a result other than the numerical value indicating the performance is an event. Specific examples of the event include, for example, items relating to the movement of equipment used in sports, such as items in which direction the ball flew after penalty kick (PK) in soccer.

  The present invention provides a sports motion analysis support system and a sports motion analysis support method capable of supporting a user to grasp a time zone in which a form has a great effect on a result among times when a series of motions in a sport are performed. And a sports motion analysis support program.

  A sports motion analysis support system according to the present invention includes a data storage unit that stores a plurality of data in which image data of a moving image representing a series of motions in a sport and the results of the motions are associated with each other. A learning unit that learns a model representing a relationship between an operation in the time section and a result corresponding to the operation for each of a plurality of time sections determined based on the time point representing the action, An evaluation unit that calculates the prediction accuracy of the predicted result for each time section.

  Further, the sports motion analysis support system according to the present invention uses a data storage unit that stores a plurality of data in which image data of a moving image representing a series of motions in a sport and a result of the motion are associated with each other, A learning unit that learns a model representing a relationship between an operation in a time section and a result corresponding to the operation for each of a plurality of time sections determined based on a time point representing a predetermined operation; And a specifying unit that specifies a time zone capable of specifying a time zone in which the degree of improvement in the prediction accuracy of the result predicted by using the time zone is large.

  Further, the sports motion analysis support method according to the present invention includes a computer including a data storage unit that stores a plurality of data in which moving image data representing a series of motions in a sport and the results of the motions are associated with each other. Using, for each of a plurality of time segments determined based on the time point representing a predetermined operation, learning a model representing the relationship between the operation in the time segment and the result corresponding to the operation, learning the model It is characterized in that the prediction accuracy of the result predicted by using the above is calculated for each time section.

  Further, the sports motion analysis support method according to the present invention includes a computer including a data storage unit that stores a plurality of data in which moving image data representing a series of motions in a sport and the results of the motions are associated with each other. Using, for each of a plurality of time segments determined based on the time point representing a predetermined operation, learning a model representing the relationship between the operation in the time segment and the result corresponding to the operation, learning the model It is characterized in that a time section in which a time zone in which the degree of improvement in the prediction accuracy of a result predicted using the prediction result is large is specified is specified.

  Further, the sports operation analysis support program according to the present invention is a sports operation analysis program installed in a computer including a data storage unit that stores a plurality of data in which moving image data representing a series of operations in sports are associated with the results of the operations. An analysis support program that, for a computer, uses a plurality of data to correspond to an operation in a time section and the operation in each of a plurality of time sections defined based on a time point representing a predetermined operation. A learning process for learning a model representing a relationship with a result to be performed, and an evaluation process for calculating prediction accuracy of a result predicted by using the model for each time section.

  Further, the sports operation analysis support program according to the present invention is a sports operation analysis program installed in a computer including a data storage unit that stores a plurality of data in which moving image data representing a series of operations in sports are associated with the results of the operations. An analysis support program that, for a computer, uses a plurality of data to correspond to an operation in a time section and the operation in each of a plurality of time sections defined based on a time point representing a predetermined operation. To perform a learning process of learning a model representing a relationship with a result to be performed, and a specifying process of specifying a time zone capable of specifying a time zone in which a degree of improvement in prediction accuracy of a result predicted using the model is large. It is characterized by.

  ADVANTAGE OF THE INVENTION According to this invention, a user can be assisted so that the time zone which a form has a big influence on a result among the time which performs a series of operation | movements in a sport can be grasped.

1 is a block diagram illustrating an example of a sports motion analysis support system according to a first embodiment of the present invention. It is a schematic diagram which shows a series of operation | movement of the player who performs long jump. It is a schematic diagram which shows the example of a some time division. FIG. 9 is an explanatory diagram showing identification information of a set of still images obtained for each time section in each image data. FIG. 4 is an explanatory diagram illustrating an example of a graph displayed by a display unit. 6 is a flowchart illustrating an example of processing progress according to the first exemplary embodiment of the present invention. It is a flowchart which shows the example of a process progress of 2nd Embodiment of this invention. FIG. 9 is an explanatory diagram illustrating an example of a display mode of a time section in step S12. It is a block diagram showing one modification of a 2nd embodiment. It is a schematic diagram which shows the example of the screen which a display part displays. FIG. 4 is a block diagram illustrating a configuration example when a data acquisition unit is provided. FIG. 3 is a schematic block diagram illustrating a configuration example of a computer according to each embodiment of the present invention. It is a block diagram showing the outline of the present invention. It is another block diagram which shows the outline | summary of this invention.

  Hereinafter, a long jump is described as an example of sports. However, the present invention is also applicable to sports other than long jump. In the following description, an example will be described in which the result of the long jump operation is a result (jump distance in this example). In this example, the result is represented numerically.

  However, as described above, the result of the action in the sport may be an event such as in which direction the ball flew. When the case where the result of the operation is an event is applied to the present invention, the result is represented by a value (for example, a binary value of “0” and “1”) according to the content of the event. The case where the result of the operation is an event will be described later.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing an example of a sports motion analysis support system according to the first embodiment of the present invention. The sports motion analysis support system 1 according to the first embodiment of the present invention includes a data storage unit 2, a time-segmented image extraction unit 3, a learning unit 4, a prediction unit 5, an evaluation unit 6, a display unit 7, Is provided.

  The data storage unit 2 is a storage device that stores a plurality of data in which image data of a moving image representing a series of operations in a sport and results of the operations are associated with each other. In this example, in each data, moving image data representing a series of motions of a person performing a long jump and a result (jump distance) obtained as a result of the motion are associated with each other. The data storage unit 2 stores a plurality of such data.

  The manner of associating the image data with the results is not particularly limited. For example, the result may exist as data different from the image data, and the image data and the result may be associated with each other. Further, for example, the image data and the result may be associated with each other in such a manner that the result is included in the moving image of the image data (in other words, the result is displayed on the moving image). This is the same in the embodiment described later.

  FIG. 2 is a schematic diagram showing a series of actions of a person performing a long jump (hereinafter, referred to as a player). In FIG. 2, for simplicity of the drawing, for example, the same posture is shown as the posture of the athlete during the run, but the actual athlete performs a series of operations while moving his limbs and the like. Perform the operation. This is the same for the schematic diagram shown in FIG. The athlete makes a run, jumps at the crossing board 11, and then lands. By capturing the series of operations with a video camera, moving image data representing the series of operations can be obtained. Then, data associating the image data with the result at that time becomes one data. The data storage unit 2 stores a plurality of such data in advance.

  When performing analysis for a specific player, data that associates the image data of a moving image obtained by imaging during the practice or competition of the specific player with the performance at that time is used as the data. What is necessary is just to store a plurality of them in the storage unit 2. Hereinafter, a case where the data storage unit 2 stores a plurality of pieces of data relating to a specific player will be described as an example. The specific player may be a player who uses the sports motion analysis support system 1 of the present invention, or a player who is instructed by a coach or the like who uses the sports motion analysis support system 1. Alternatively, the above-mentioned specific player may be a competitor for a user using the sports motion analysis support system 1 of the present invention. This is the same in other embodiments described later.

  Since the animation represents a series of actions of the player, the animation also represents the form of the player at each point in time.

  Here, the time division in the present invention will be described. The time section in the present invention is a time section determined based on the time point at which the moving image represents a specific action of the player. In the example of the long jump, the time at which the moving image represents the level crossing operation is set as a reference, and the reference time is set to 0. Also, a time before the reference (time 0) is represented by a negative value, and a time after the reference is represented by a positive value. The time section can be determined by defining the start time and the end time with the time at which the moving image represents the level crossing operation as 0 as a reference time. Also, a plurality of time sections are determined in advance. Further, the lengths of the time sections may be different from each other, and a plurality of time sections may include a common time zone.

  FIG. 3 is a schematic diagram illustrating an example of a plurality of time segments. In the example shown in FIG. 3, a plurality of time segments a to g are shown. “A” to “g” illustrated in FIG. 3 are identification information of time segments. In the example shown in FIG. 3, the time segment a is in the range from time “−2.5” to time “−2.0” (see FIG. 3). The time section b is a range from the time “−2.5” to the time “−1.5” (see FIG. 3). Similarly, the start time and the end time are determined for the other time segments c, d, e,. Further, in the example illustrated in FIG. 3, the start time of each time section is set to “−2.5”, but the start time of each time section is “−2.5”. The start time of each time segment may be a time earlier than “−2.5” or a time later than “−2.5”. Further, in the example illustrated in FIG. 3, the start times of the time segments are common, but the start times of the time segments do not have to be common.

  Further, each time section may be continuously determined so as not to include a common time zone with other time sections. For example, the time segment “a” from time “−2.5” to time “−2.0” is defined as time segment “a”, the time segment from time “−2.0” to time “−1.5” is defined as time segment “b”, and the time segment “−1”. .5 "to time" -1.0 "may be defined as the time section c.

  Further, each time section may be determined such that the end time is common and the start time is different. At this time, the reference time “0” may be determined as a common end time.

  Preferably, the plurality of time segments are determined so as to be ordered based on the time. In the example shown in FIG. 3, the start time of each time segment is common, and the end time is different. Accordingly, the time segments shown in FIG. 3 can be ordered in the order of end time, a, b, c, d, e, f, and g. Hereinafter, a case will be described as an example where the time segments are determined so as to be ordered based on the time.

  The time-segmented image extracting unit 3 specifies, for each image data stored in the data storage unit 2, a range corresponding to each of a plurality of predetermined time segments. Then, the time-segmented image extracting unit 3 extracts a still image from the range corresponding to each time segment for each image data, and generates a set of still images for each time segment of each image data.

  For example, assume that the first image data is image data # 1 and the arbitrary n-th image data is image data #n. The time section image extraction unit 3 specifies a range corresponding to the time section in the image data # 1 for each time section. That is, the time section image extraction unit 3 specifies a range corresponding to the time section a, a range corresponding to the time section b, and the like in the image data # 1. Then, the time division image extraction unit 3 extracts a still image from the range corresponding to the time division a in the image data # 1, and generates a set of still images corresponding to the time division a in the image data # 1. When extracting a still image from image data of a moving image, the time-division image extracting unit 3 may extract the still image every predetermined time. For example, when extracting a still image corresponding to the time segment a in the image data # 1, the time segmented image extracting unit 3 may extract the still image every 0.1 seconds from the range corresponding to the time segment a. Here, 0.1 seconds is illustrated as an example of the above-mentioned predetermined time, but the above-mentioned predetermined time is not limited to 0.1 second.

  Similarly, the time-segmented image extracting unit 3 extracts still images from the range corresponding to each time segment in the image data # 1, and generates a set of still images for each time segment.

  The time-segmented image extracting unit 3 performs the same process on each image data other than the image data # 1, and generates a set of still images for each time segment in each image data.

  A set of still images obtained for each time section in each image data represents the movement and form of the player in the corresponding time section.

  Hereinafter, for convenience, a set of still images obtained for each time section in each image data is represented by identification information of the set. The identification information of a set of still images corresponding to an arbitrary time section in arbitrary image data is represented by a code in which the identification information of the image data and the identification information of the time section are arranged. FIG. 4 is an explanatory diagram showing identification information of a set of still images obtained for each time section in each image data. For example, a set of still images extracted from the time section a in the image data # 1 is represented as “# 1, a” (see FIG. 4). The identification information of another set of still images is represented by the same rule. FIG. 4 also shows the results corresponding to each image data.

  The learning unit 4 learns, by machine learning, a model representing a relationship between an action in the time section and a result (jump distance) corresponding to the action for each of the plurality of time sections (in other words, generation). Do).

  More specifically, this model is a model showing a relationship between a form in a time section (a form representing a set of still images corresponding to the time section) and a grade. In addition, the learning unit 4 learns a model in which a set of still images representing a form is used as an explanatory variable and the result is an objective variable. Therefore, it is possible to calculate the predicted value of the performance using the model obtained by learning and the set of still images.

  When focusing on one time section, a set of still images corresponding to the time section is obtained for each image data, and the corresponding result is stored in the data storage unit 2 for each image data. . The learning unit 4 determines, for each image data, a combination of a set of still images and a result corresponding to the time segment of interest, and uses the combination determined for each image data as learning data to determine the time of interest. The model corresponding to the section may be learned by machine learning. The machine learning algorithm may be any algorithm that can learn the model for calculating the predicted value of the grade as described above.

  Hereinafter, description will be made with reference to FIG. For example, it is assumed that the learning unit 4 learns a model corresponding to the time section a. In this case, the learning unit 4 combines the set of still images “# 1, a” with the result “5.8 m”, the combination of the set of still images “# 2, a” and the result “6.5 m”, The combination of the still image set “# 3, a” and the result “7.1 m”, the combination of the still image set “# 4, a” and the result “6.2 m” (see FIG. 4), A model indicating the relationship between the form in time section a (a set of still images corresponding to time section a) and the performance is determined using the combination as learning data.

  Similarly, the learning unit 4 learns models corresponding to the other time segments b, c, d,...

  By the processing of the learning unit 4, a model is obtained for each time section.

  The prediction unit 5 calculates a predicted value of the performance of each image data using a model corresponding to each time section.

  For example, the prediction unit 5 calculates a predicted value of the performance of the image data # 1 using a model corresponding to the time section a. In this case, the prediction unit 5 calculates the predicted value of the score by applying the set “# 1, a” of the still images corresponding to the time section a in the image data # 1 to the model corresponding to the time section a. I do. The prediction unit 5 similarly calculates a predicted value of the performance based on the model corresponding to each of the other time segments with respect to the image data # 1. That is, the prediction unit 5 calculates a predicted value of the performance for the image data # 1 for each model corresponding to each time section.

  Therefore, the prediction unit 5 calculates the predicted value of the performance for one image data by the number of time segments.

  The prediction unit 5 similarly calculates the predicted value of the performance for each image data other than the image data # 1 by the number of time segments.

  Assuming that the number of image data is N, the prediction unit 5 calculates N prediction values corresponding to the time section a. Similarly, N prediction values corresponding to each of the other time segments are also calculated.

  The evaluation unit 6 calculates the prediction accuracy of the predicted value for each time section using a plurality of predicted values (predicted values for the number of image data) corresponding to the time sections and the true values of the results. It can be said that the evaluation unit 6 evaluates the prediction accuracy of the model for each time section.

  The true value of the score is a score value stored in the data storage unit 2 in association with the image data.

  For example, the evaluation unit 6 may calculate an average value of values obtained by dividing the predicted value of the result by the true value of the result for each time section. This value can be said to be a value representing the prediction accuracy. For example, the evaluation unit 6 divides the predicted value calculated based on the time segment a of the image data # 1 by the value of the result (true value) associated with the image data # 1. For each of the other image data, the evaluation unit 6 also divides the predicted value calculated based on the time segment a by the value of the grade (true value) associated with the image data of interest. The evaluation unit 6 calculates the average value of the division results obtained for each image data as the prediction accuracy corresponding to the time section a (more specifically, the prediction accuracy of the model corresponding to the time section a). The evaluation unit 6 similarly calculates prediction accuracy corresponding to each of the other time segments.

  The calculation method of the prediction accuracy is not limited to the above example. For example, the evaluation unit 6 may calculate an average value of values obtained by dividing the absolute value of the difference between the predicted value of the result and the true value of the result by the true value for each time section. This value can also be said to be a value representing the prediction accuracy.

  The display unit 7 displays the relationship between the time section and the prediction accuracy corresponding to the time section (more specifically, the prediction accuracy of the model corresponding to the time section). The display unit 7 may display the time segment and the prediction accuracy of the model corresponding to the time segment as text information, for example. However, it is particularly preferable that the display unit 7 graphically displays the relationship between the time segment and the prediction accuracy of the model corresponding to the time segment, from the viewpoint of easy understanding of the displayed information by a viewer.

  FIG. 5 is an explanatory diagram illustrating an example of a graph displayed by the display unit 7, and illustrates a graph indicating a relationship between a time segment and a prediction accuracy of a model corresponding to the time segment. The vertical axis of the graph shown in FIG. 5 represents the prediction accuracy of the model. The horizontal axis of the graph represents time division. In this example, the start times of the time segments are common, and the end times are different. Therefore, each time section can be ordered in the order of the end time. In this example, the display unit 7 orders the time segments in order of the end time, and displays the identification information of each time segment along the horizontal axis in that order (see FIG. 5). Then, the display unit 7 displays a graph indicating a change in prediction accuracy corresponding to each time segment, for example, as illustrated in FIG.

  The time segmented image extraction unit 3, the learning unit 4, the prediction unit 5, the evaluation unit 6, and the display unit 7 are realized by, for example, a CPU (Central Processing Unit) of a computer having a display device (not shown in FIG. 1). In this case, the CPU reads the sports motion analysis support program from a program recording medium such as a computer program storage device (not shown in FIG. 1), and according to the sports motion analysis support program, the time-segmented image extraction unit 3 and the learning unit 4 , The prediction unit 5, the evaluation unit 6, and the display unit 7. A portion of the display unit 7 for determining display contents (for example, a graph) and displaying the display contents on a display device is realized by the CPU. The part of the display unit 7 that actually performs display is realized by a display device. Further, the above computer may be a personal computer or a portable computer such as a smartphone. These points are the same in other embodiments described later.

  The sports motion analysis support system 1 may have a configuration in which two or more physically separated devices are connected by wire or wirelessly. For example, the sports motion analysis support system 1 may be realized as a system in which a portable computer such as a smartphone and a server cooperate. This is the same in other embodiments described later.

  Next, the processing progress will be described. FIG. 6 is a flowchart illustrating an example of the progress of processing according to the first embodiment of this invention. Since the details of the operation of each step described below have already been described, the detailed description is omitted here.

  It is assumed that a plurality of data are stored in the data storage unit 2 in advance.

  The time-segmented image extracting unit 3 specifies, for each image data stored in the data storage unit 2, a range corresponding to each of a plurality of predetermined time segments. Then, the time-segmented image extracting unit 3 extracts a still image from the range corresponding to each time segment for each image data, and generates a set of still images for each time segment of each image data ( Step S1).

  Next, the learning unit 4 learns, for each of the plurality of time sections, a model representing the relationship between the form and the performance in the time section by machine learning (step S2).

  Next, the prediction unit 5 calculates a predicted value of the performance of each image data using a model corresponding to each time section (step S3).

  Next, for each time section, the evaluation unit 6 calculates the prediction accuracy of the model corresponding to the time section using the plurality of predicted values corresponding to the time section and the true value of the score (step S4).

  Next, the display unit 7 displays the relationship between the time segment and the prediction accuracy of the model corresponding to the time segment (Step S5). In step S5, the display unit 7 may display the time segment as text information in association with the prediction accuracy of the model corresponding to the time segment. However, it is preferable that the display unit 7 displays the relationship between the time segment and the prediction accuracy of the model corresponding to the time segment in a graph as illustrated in FIG.

  The effect of the present embodiment will be described. In the present embodiment, as described above, a model corresponding to one time section is generated using a plurality of combinations of a set of still images and a grade. Therefore, when the predicted value of the performance is calculated by this model, if the prediction accuracy of the model is good, it can be said that the predicted value statistically expresses the tendency of the performance according to the form. Then, according to the present embodiment, the evaluation unit 6 calculates, for each time section, the prediction accuracy of the model corresponding to the time section. Then, the display unit 7 displays the relationship between the time segment and the prediction accuracy of the model corresponding to the time segment. Therefore, the user of the sports motion analysis support system 1 (the player whose image data and results are stored in the data storage unit 2 or his coach or the like) can estimate the prediction accuracy for each time section that is ordered based on the time. It is possible to confirm and, based on the change in the prediction accuracy, it is possible to grasp a time zone in which the degree of improvement in the prediction accuracy of the results predicted using the model is large. As described above, if the prediction accuracy of the model is good, it can be said that the predicted value of the performance obtained by using the model statistically expresses the tendency of the performance according to the form. Therefore, to be able to grasp the time period in which the degree of improvement in the prediction accuracy of the result predicted using the model is large means that the time period in which the form has a great influence on the result can be grasped. Therefore, according to the present embodiment, it is possible to support the user so that the user can grasp the time zone in which the form has a great influence on the result among the time period for performing a series of actions in the sport (the long jump in this example). .

  For example, assume that a graph illustrated in FIG. 5 is displayed. For the time segments a and b, the prediction accuracy of the corresponding model is slightly increased. Similarly, for the time sections c to g, the prediction accuracy of the corresponding model is slightly increased. On the other hand, the degree of improvement in the prediction accuracy of the model corresponding to the time section c is greater than that of the model corresponding to the time section b. Therefore, the user can select a time zone corresponding to the difference between time segment c and time segment b (specifically, a time zone from time “−1.5” to time “−1.0”; see FIG. 3). ) Can be recognized as a time zone in which the degree of improvement in the prediction accuracy of the results predicted using the model is large. Therefore, the user can understand that the time zone is the time zone in which the form has a great influence on the result among the time period for performing a series of long jump operations.

  If such a time zone (in this example, the time zone from time “−1.5” to time “−1.0”) can be grasped, the user focuses on the form of the time zone. This can contribute to improving the performance of the players.

Embodiment 2. FIG.
The sports motion analysis support system 1 according to the first embodiment displays, for example, a graph illustrated in FIG. 5 so that the user can select a time zone in which the degree of improvement in the prediction accuracy of the results predicted using the model is large. I was able to understand. The sports motion analysis support system according to the second embodiment of the present invention specifies a time section that can specify a time zone in which the degree of improvement in prediction accuracy of a result predicted using a model is large, and displays the time section. I do.

  The sports motion analysis support system according to the second embodiment of the present invention can be represented by the blocks shown in FIG. 1 similarly to the sports motion analysis support system 1 according to the first embodiment. A second embodiment will be described. The description of the same items as in the first embodiment will be omitted as appropriate.

  The data storage unit 2, the time division image extraction unit 3, the learning unit 4, and the prediction unit 5 are the same as the data storage unit 2, the time division image extraction unit 3, the learning unit 4, and the prediction unit 5 of the first embodiment. The description is omitted.

  The evaluation unit 6 of the second embodiment performs the same operation as the evaluation unit 6 of the first embodiment, and also performs the operation described below.

  The evaluation unit 6 specifies a time section that can specify a time zone in which the degree of improvement in the prediction accuracy of the result predicted using the model is large.

  For example, the evaluation unit 6 orders the time segments based on the time, and calculates the difference between the prediction accuracy of the models corresponding to two time segments in adjacent order. In this case, the evaluation unit 6 calculates a value obtained by subtracting the prediction accuracy of the model corresponding to the time segment a from the prediction accuracy of the model corresponding to the time segment b. The evaluation unit 6 also calculates the other sets of time sections having adjacent orders, such as a set of time sections b and c and a set of time sections c and d, from the prediction accuracy of the model corresponding to the time section having the later order. , A value obtained by subtracting the prediction accuracy of the model corresponding to the preceding time segment.

  Then, for example, in the set of time segments in which the difference in prediction accuracy is maximum, the evaluation unit 6 regards the time segments in the later order as time segments in which a time zone in which the degree of improvement in prediction accuracy is large can be specified. It should be specified. For example, it is assumed that the difference in prediction accuracy between the sets of time segments b and c is the largest among the differences in prediction accuracy. In this case, the evaluation unit 6 may specify the time section c as a time section that can specify a time zone in which the degree of improvement in the prediction accuracy of the results predicted using the model is large.

  Also, for example, in a set of time segments in which the difference in prediction accuracy is equal to or greater than a predetermined threshold, a time segment in a later order is specified as a time segment in which a time zone in which the degree of improvement in prediction accuracy is large can be specified. You may. For example, it is assumed that the difference in prediction accuracy between a set of time segments b and c is equal to or greater than a threshold. In this case, the evaluation unit 6 may specify the time section c as a time section that can specify a time zone in which the degree of improvement in the prediction accuracy of the results predicted using the model is large. Note that the threshold may be determined in advance.

  A method of specifying a time section that can specify a time zone in which the degree of improvement in the prediction accuracy of a result predicted using a model is large is not limited to the above method, and may be another method.

  The display unit 7 displays the time segment specified by the evaluation unit 6.

  FIG. 7 is a flowchart illustrating an example of the progress of processing according to the second embodiment of this invention. Steps S1 to S4 (see FIG. 7) in the second embodiment are the same as steps S1 to S4 (see FIG. 6) in the first embodiment, and a description thereof will be omitted.

  After step S4, the evaluation unit 6 specifies a time section that can specify a time zone in which the degree of improvement in the prediction accuracy of the results predicted using the model is large (step S11). An example of the operation of the evaluation unit 6 has already been described, and the description is omitted here.

  After step S11, the display unit 7 displays the time segment specified in step S11 (a time segment that can specify a time zone in which the degree of improvement in the prediction accuracy of the results predicted using the model is large) (step S11). S12).

  FIG. 8 is an explanatory diagram illustrating an example of a display mode of the time section in step S12. Here, a description will be given assuming that the time section c is specified in step S11. The display unit 7 displays an icon 21 representing the identification information of the time section for each time section. The display unit 7 also displays the start time and the end time of the time section. Then, the display unit 7 displays the icon 21 of the time segment specified in step S11 (in this example, the icon 21 of the time segment c) in a manner different from the icons 21 of the other time segments. The example illustrated in FIG. 8 illustrates a case where the frame of the icon 21 of the time section c is displayed thicker than the frame of the other icons 21. That is, the display unit 7 displays the frame of the icon 21 of the time section c thicker than the frame of each of the other icons 21 to confirm that the time section c is the time section specified in step S11. Represents.

  FIG. 8 is an example of the display mode in step S12, and the display mode in step S12 is not limited to the example illustrated in FIG.

  In the present embodiment, the evaluation unit 6 specifies a time section that can specify a time zone in which the degree of improvement in the prediction accuracy of the result predicted using the model is large. Then, the display unit 7 displays the time segment. Therefore, as in the first embodiment, the user can grasp the time zone in which the form has a great influence on the result among the time for performing a series of long jump operations. For example, it is assumed that the evaluation unit 6 specifies the time segment c in step S11, and the display unit 7 displays the time segment c. In this case, the user selects a time zone corresponding to the difference between the time segment c and the immediately preceding time segment b (specifically, the time zone from the time “−1.5” to the time “−1.0”). Time zone (see FIGS. 3 and 8)) can be recognized as a time zone in which the degree of improvement in the prediction accuracy of the results predicted using the model is large. Therefore, the user can understand that the time zone is the time zone in which the form has a great influence on the result among the time period for performing a series of long jump operations.

  Next, various modifications of the second embodiment will be described.

  The display unit 7 may perform the operation of Step S5 in the first embodiment in Step S12. That is, the display unit 7 may display the time section specified in step S11 and also display the relationship between the time section and the prediction accuracy of the model corresponding to the time section. For example, the display unit 7 may display the time section c in the mode illustrated in FIG. 8 and also display the graph illustrated in FIG.

  In addition, when a time segment is designated from outside (for example, a user), the sports motion analysis support system may further display information on the time segment. FIG. 9 is a block diagram showing a modification of the second embodiment. The sports motion analysis support system 1 shown in FIG. 9 further includes an operation unit 8 in addition to the components shown in FIG. The operation up to step S12 is the same as the operation described in the second embodiment or the operation in the above-described modification. In the following description, a case will be described as an example where the display unit 7 displays the icons 21 of the respective time segments in the mode illustrated in FIG. 8 in step S12.

  The operation unit 8 is a user interface for a user to specify a time section, and is realized by, for example, a mouse or the like. The user operates the operation unit 8 to specify the icon 21 displayed as illustrated in FIG. The display unit 7 accepts the designation of the time section according to the operation.

  Here, a description will be given assuming that the time segment specified in step S11 is specified. For example, among the icons 21 shown in FIG. 8, a description will be given assuming that an operation such as a click has been performed on the icon 21 of the time section c.

  When receiving the designation of the time section c from the outside, the display unit 7 displays the prediction accuracy of the model corresponding to the time section c.

  Further, the display unit 7 displays the image data in which the predicted value of the result calculated using the model (the model corresponding to the time section c) in step S3 is the maximum, and calculates the image data using the model in step S3. The image data with the smallest predicted value of the grade is specified. In the example of long jump, the higher the grade (jump distance) value, the better the grade, and the smaller the grade value, the worse the grade. Depending on the type of sport, the higher the score, the worse the score, and the lower the score, the better the score. In any case, the display unit 7 displays the image data in which the predicted value of the performance calculated using the model corresponding to the designated time section is the maximum, and the image data calculated using the model. The image data with the minimum predicted value of the result is specified. By specifying the image data in this manner, the display unit 7 specifies the image data that is predicted to have the best performance and the image data that is predicted to have the worst performance. Here, the maximum means the maximum among the predicted values calculated using the model corresponding to the specified time section c, and the minimum means the model corresponding to the specified time section c. Means the minimum of the predicted values calculated using

  The display unit 7 displays the moving image in the range corresponding to the designated time section c in the image data with the largest predicted value and the designated time section c in the image data with the smallest predicted value. The video in the corresponding range is displayed.

  FIG. 10 is a schematic diagram showing an example of a screen displayed on the display unit 7 when a time section is designated after step S12. The prediction accuracy display column 31 is a column for displaying the prediction accuracy of the model corresponding to the designated time section. The first image display column 32 is a column for displaying a moving image in a range corresponding to the designated time section in the image data having the maximum predicted value. The second image display column 32 is a column for displaying a moving image in a range corresponding to the designated time section in the image data having the minimum predicted value.

  When the time section is designated, the display unit 7 displays a screen illustrated in FIG.

  In the present modified example, the case where the display unit 7 specifies the image data having the maximum predicted value and the image data having the minimum predicted value has been described. The display unit 7 displays each of the image data in which the predicted value of the result calculated using the model corresponding to the designated time section corresponds to the first to the top predetermined, and the image data calculated using the model. Each of the image data whose predicted value of the grade corresponds to the first to the lower predetermined order may be specified. Then, the display unit 7 displays a moving image in a range corresponding to the designated time section in each image data whose predicted value corresponds to the first to upper predetermined positions, and a moving image whose predicted value corresponds to the lower first to lower predetermined values. The moving image in the range corresponding to the designated time section in each of the image data corresponding to the first may be displayed. The values representing the upper predetermined number and the lower predetermined number may be predetermined. Alternatively, the sports motion analysis support system 1 may be provided with an input device (user interface) for allowing the user to input values representing the upper predetermined number and the lower predetermined number to the sports motion analysis support system 1.

  According to this modification, it is possible to display a video that is predicted to have the best performance, that is, a video during a time period when the form has a significant effect on the result, and that the performance is predicted to be the worst. Video that has a significant effect on the results. Therefore, by analyzing or comparing those images, the images can be used for improving the form, discovering the habit of the form, and the like.

  The display unit 7 may perform the above-described display operation only when the time segment specified in step S11 is specified. Further, when an arbitrary time segment is designated, the display unit 7 may perform the above-described display operation according to the designated time segment.

  In the first embodiment, the second embodiment, and the modifications thereof, the case where the result of the operation is a result and is represented by a numerical value has been described as an example. The result of the action in the sport may be an event. Hereinafter, a case where a PK (penalty kick) scene in soccer is applied to the present invention will be described as an example. The description will be made assuming that the result of the PK operation is one of two types of events: "the ball has flew to the right" and "the ball has flew to the left." The description of the same items as those in the first embodiment, the second embodiment, and the modifications thereof will be omitted.

  The event “the ball flew to the right” is represented by “1”, and the event “the ball flew to the left” is represented by “0”. That is, the result (event) of the PK operation is represented by a binary value.

  The data storage unit 2 associates image data of a moving image representing a series of motions of a person performing PK (hereinafter, referred to as a player) with a result of the motion (event “1” or event “0”). Are stored in advance. The result of this operation can be, for example, a nominal measure.

  Further, in this example, a plurality of time sections may be determined based on the time point at which the moving image represents the kicking motion of the player.

  In this example, the learning unit 4 learns, for each of a plurality of time segments, a model representing the relationship between the operation in the time segment and the probability that the event “1” occurs or the probability that the event “0” occurs. The probability that the event "1" occurs or the probability that the event "0" occurs is represented by one objective variable. Here, the possible range of the objective variable is 0 to 1. If the value of the objective variable is greater than 0.5, the value represents the probability that an event "1" (i.e., the event "ball will fly to the right") will occur, and the closer the value is to 1, the more likely it is. The probability that the event “1” occurs is high, and the closer the value is to 0.5, the lower the probability that the event “1” occurs. If the value of the objective variable is smaller than 0.5, the value represents the probability that an event “0” (that is, the event “the ball flies to the left”) will occur, and the value becomes 0. The closer the value is, the higher the probability of occurrence of the event “0” is, and the closer the value is to 0.5, the lower the probability of occurrence of the event “0” is. When learning such a model, the learning unit 4 may use, for example, logistic regression analysis as an algorithm of machine learning.

  The learning unit 4 determines a combination of a set of still images corresponding to a time segment of interest and a result (“1” or “0”) for each image data, and learns a combination determined for each image data. A model corresponding to a time segment of interest may be learned by machine learning (for example, logistic regression analysis) by using the data as the application data.

  The learning unit 4 learns a model for each time section.

  The prediction unit 5 calculates a predicted value of a result for each image data using a model corresponding to each time section. The predicted value is the value of the objective variable, and represents the probability that the event “1” occurs or the probability that the event “0” occurs. The probability (value of the objective variable) calculated by the prediction unit 5 is a continuous value that can take a value in the range of 0 to 1. Therefore, it can be said that the value of the target variable calculated by the prediction unit 5 is, for example, an ordinal scale.

  When the evaluation unit 6 calculates the prediction accuracy of the model corresponding to the time section for each time section, the evaluation section 6 may perform the following processing. The true value of the result is “1” or “0” and is stored in the data storage unit 2 in association with the image data.

  Here, the case where the evaluation unit 6 calculates the prediction accuracy of the model corresponding to the time section a by focusing on the time section a will be described as an example. It is assumed that the prediction unit 5 calculates a predicted value for each image data using a model corresponding to the time section a. The number of prediction values calculated for the time section a is the number of image data. The evaluation unit 6 regards each predicted value as “1” or “0”. That is, the evaluation unit 6 regards the predicted value as “1” if one predicted value is larger than 0.5, and sets the predicted value as “0” if the predicted value is smaller than 0.5. I reckon. Then, the evaluation unit 6 determines whether or not the predicted value regarded as “1” or “0” matches the true value of the result, and counts the number of predicted values that match the true value. I do.

  For example, it is assumed that a prediction value calculated by applying the model corresponding to the time section a to the image data # 1 is “0.8”. It is also assumed that the true value of the result corresponding to image data # 1 is "1". Since 0.8> 0.5, the evaluation unit 6 regards the predicted value “0.8” as “1” and determines that it matches the true value “1”. In this case, if the true value is “0”, the evaluation unit 6 may regard the predicted value “0.8” as “1” and determine that it does not match the true value “0”.

  Further, for example, it is assumed that a predicted value calculated by applying the model corresponding to the time section a to the image data # 2 is “0.3”. It is also assumed that the true value of the result corresponding to image data # 2 is “0”. Since 0.3 <0.5, the evaluation unit 6 regards the predicted value “0.3” as “0” and determines that it matches the true value “0”. In this case, if the true value is “1”, the evaluation unit 6 may regard the predicted value “0.3” as “0” and determine that it does not match the true value “1”.

  As described above, the evaluation unit 6 regards the predicted value as “1” or “0” for each predicted value corresponding to the image data, and counts the number of predicted values whose values match the true values. . Then, the evaluation unit 6 sets the value obtained by dividing the count value by the number of predicted values calculated for the time section a as the prediction accuracy of the model corresponding to the time section a. As described above, a value obtained by dividing the above count value by the number of predicted values can also be referred to as a coincidence rate.

  The evaluation unit 6 calculates the prediction accuracy of the model corresponding to each of the other time segments.

  The other points are the same as those of the above-described first embodiment and second embodiment. Therefore, even when the result of the operation is an event, the present invention can be applied to the first embodiment and the second embodiment. Then, the same effects as in the first embodiment and the second embodiment can be obtained. In other words, the user can grasp the time zone in which the form has a great effect on the result (in this example, the traveling direction of the ball) in the time period for performing a series of PK operations.

  In particular, in this example, if data relating to competitors (players of the opposing team) is stored in the data storage unit 2, the time zone in which the form has a great influence on the result is grasped, and You can focus on the form. Then, the keeper can easily find out the habits and the like of the players of the opponent team, and can contribute to the PK rejection rate of the keeper.

  Further, as shown as one of the modifications of the second embodiment, the operation result may be an event even when a time division is externally designated after step S12. For example, it is assumed that, in step S12, the display unit 7 displays the icon 21 of each time section in a mode illustrated in FIG. 8, and thereafter, the time section c is designated. In this case, the display unit 7 displays the prediction accuracy of the model corresponding to the time section c.

  Further, the display unit 7 specifies the image data having the maximum predicted value calculated using the model and the image data having the minimum predicted value calculated using the model. In this example, the larger the value, the higher the probability that the ball will fly right after kicking, and the smaller the value, the higher the probability that the ball will fly left after kicking. That is, the display unit 7 identifies the image data of the form in which the probability that the ball will fly to the right after the kick is the highest, and the image data of the form in which the probability that the ball will fly to the left after the kick is the highest. You are doing.

  The display unit 7 displays the moving image in the range corresponding to the designated time section c in the image data with the largest predicted value and the designated time section c in the image data with the smallest predicted value. The video in the corresponding range is displayed.

  The screen displayed by the display unit 7 when the time section is designated may be the same as the screen illustrated in FIG. Note that the text information such as “Result: good” and “Result: bad” shown in FIG. 10 may be displayed as “ball traveling direction: right”, “ball traveling direction: left”, and the like, respectively.

  In this case, it is possible to display a video in which the ball is predicted to be most likely to fly to the right, and a video during a time when the form has a great influence on the result. It is possible to display a moving image that is expected to be easy and that is in a time zone in which the form has a great influence on the result. Therefore, by analyzing or comparing those images, the images can be used for improving the form, discovering the habit of the form, and the like.

  In each embodiment of the present invention and its modifications, the sports motion analysis support system 1 may include a data acquisition unit that acquires data to be stored in the data storage unit 2 from outside. FIG. 11 is a block diagram illustrating a configuration example when a data acquisition unit is provided. The data storage unit 2, the time-segmented image extraction unit 3, the learning unit 4, the prediction unit 5, the evaluation unit 6, and the display unit 7 include those elements in the first embodiment, and those in the second embodiment and its modifications. They are the same as those elements, and the description is omitted.

  The data acquisition unit 9 acquires a plurality of pieces of data in which image data of a moving image representing a series of actions in sports and the results of the actions are associated with each other, and stores the data in the data storage unit 2.

  For example, it is assumed that a plurality of the above data are stored in an external device (not shown). In this case, the data acquisition unit 9 may access the device, acquire a plurality of data from the device, and store the data in the data storage unit 2. The processing after the data acquisition unit 9 stores a plurality of data in the data storage unit 2 is the same as the processing described in the first embodiment, or the processing described in the second embodiment or its modification. is there.

  The data acquisition unit 9 is realized by, for example, a CPU of a computer that operates according to a sports motion analysis support program.

  In the above description, a long jump or PK in soccer has been described as an example. The sports operation to which the present invention is applied is not limited to these. For example, data in which image data of a swing operation in golf and results (flying distance) are associated with each other may be stored in the multiple data storage unit 2. In this case, the time when the golf club hits the ball may be set as the reference time. Further, for example, data in which the image data of the action of the toss of the volleyball setter and the result indicating whether the ball flew to the right or left may be stored in the multiple data storage unit 2. Good. In this case, the time when the setter tosses may be set as the reference time. Further, for example, the present invention can be applied to a pitching operation of a pitcher in baseball, a formation of rugby, or a formation of American football. As described above, the present invention is applicable to various sports operations.

  FIG. 12 is a schematic block diagram illustrating a configuration example of a computer according to each embodiment of the present invention. The computer 1000 includes a CPU 1001, a main storage device 1002, an auxiliary storage device 1003, an interface 1004, a display device 1005, and an input device 1006. The input device 1006 corresponds to the operation unit 8 shown in FIG.

  The sports motion analysis support system 1 according to each embodiment of the present invention is mounted on a computer 1000. The operation of the sports motion analysis support system 1 is stored in the auxiliary storage device 1003 in the form of a program (sport motion analysis support program). The CPU 1001 reads out the program from the auxiliary storage device 1003, expands the program in the main storage device 1002, and executes the above processing according to the program.

  The auxiliary storage device 1003 is an example of a non-transitory tangible medium. Other examples of the non-transitory tangible medium include a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, and a semiconductor memory connected via the interface 1004. When the program is distributed to the computer 1000 via a communication line, the computer 1000 that has received the program may load the program into the main storage device 1002 and execute the above processing.

  Further, the program may be for realizing a part of the processing described above. Furthermore, the program may be a difference program that implements the above-described processing in combination with another program already stored in the auxiliary storage device 1003.

  A part or all of the components may be realized by a general-purpose or dedicated circuit, a processor, or a combination thereof. These may be configured by a single chip, or may be configured by a plurality of chips connected via a bus. Some or all of the components may be realized by a combination of the above-described circuit and the like and a program.

  When some or all of the components are realized by a plurality of information processing devices, circuits, and the like, the plurality of information processing devices, circuits, and the like may be centrally arranged or distributed. For example, the information processing device, the circuit, and the like may be realized as a form in which each is connected via a communication network, such as a client and server system or a cloud computing system.

  Next, the outline of the present invention will be described. FIG. 13 is a block diagram showing an outline of the present invention. The sports motion analysis support system of the present invention includes a data storage unit 2, a learning unit 4, and an evaluation unit 6.

  The data storage unit 2 stores a plurality of pieces of data in which image data of a moving image representing a series of actions in sports and results of the actions are associated.

  The learning unit 4 uses a plurality of data to determine, for each of a plurality of time segments determined based on a time point representing a predetermined operation, a relationship between an operation in the time segment and a result corresponding to the operation. Learn a model that represents.

  The evaluation unit 6 calculates the prediction accuracy of the result predicted using the model for each time section.

  With such a configuration, it is possible to assist the user so that the user can grasp the time zone in which the form has a great influence on the result among the time periods for performing a series of actions in the sport.

  FIG. 14 is another block diagram showing an outline of the present invention. The sports motion analysis support system of the present invention may include the data storage unit 2, the learning unit 4, and the specifying unit 16.

  The data storage unit 2 and the learning unit 4 are the same as the data storage unit 2 and the learning unit 4 shown in FIG.

  The specifying unit 16 (for example, the evaluation unit 6 in the second embodiment) specifies a time section that can specify a time zone in which the degree of improvement in prediction accuracy of a result predicted using a model is large.

  In this case as well, it is possible to assist the user so that the user can grasp the time zone in which the form has a great influence on the result among the time periods for performing a series of actions in the sport.

  Each of the above embodiments may be described as the following supplementary notes, but is not limited thereto.

(Appendix 1)
A data storage unit that stores a plurality of image data of moving images representing a series of actions in sports, and a plurality of data in which the results of the actions are associated
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. A learning unit to learn,
A sports motion analysis support system, comprising: an evaluation unit that calculates prediction accuracy of a result predicted using a model for each time section.

(Appendix 2)
The sports motion analysis support system according to claim 1, further comprising a display unit that displays a graph indicating a relationship between the time section and a prediction accuracy of a model corresponding to the time section.

(Appendix 3)
The sports motion analysis support system according to Supplementary Note 1 or 2, wherein the evaluation unit specifies a time section that can specify a time zone in which a degree of improvement in prediction accuracy of a result predicted using the model is large.

(Appendix 4)
The display unit displays the prediction accuracy of the model corresponding to the time section specified from the outside, specifies a predetermined number of image data based on a predicted value predicted using the model, and specifies each specified image data. The sports motion analysis support system according to any one of Supplementary Notes 1 to 3, wherein a moving image of the time section is displayed.

(Appendix 5)
The result of the operation associated with the image data is a numerical value indicating the grade,
The sports motion analysis support system according to any one of Supplementary notes 1 to 4, wherein the learning unit learns a model representing a relationship between an action in the time section and a numerical value indicating a result for each time section.

(Appendix 6)
The result of the operation associated with the image data is an event,
The sports motion analysis support system according to any one of Supplementary notes 1 to 4, wherein the learning unit learns a model representing a relationship between an action in the time section and a probability of occurrence of an event for each time section.

(Appendix 7)
A data storage unit that stores a plurality of image data of a moving image representing a series of operations in sports, and a plurality of data in which the results of the operations are associated with each other;
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. A learning unit to learn,
A sports motion analysis support system, comprising: a specification unit that specifies a time zone that can specify a time zone in which a degree of improvement in prediction accuracy of a result predicted using a model is large.

(Appendix 8)
A computer including a data storage unit that stores a plurality of pieces of data in which image data of a moving image representing a series of actions in sports and results of the actions are associated with each other.
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. Learn,
A sports motion analysis support method, wherein a prediction accuracy of a result predicted using a model is calculated for each time section.

(Appendix 9)
A computer including a data storage unit that stores a plurality of pieces of data in which image data of a moving image representing a series of actions in sports and results of the actions are associated with each other.
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation Learn,
A sports motion analysis support method characterized by identifying a time zone that can identify a time zone in which the degree of improvement in prediction accuracy of a result predicted using a model is large.

(Appendix 10)
A sports motion analysis support program mounted on a computer including a data storage unit that stores a plurality of data in which moving image data representing a series of motions in a sport and a result of the motion are associated with each other,
On the computer,
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. Learning process to learn, and
A sports motion analysis support program that executes an evaluation process that calculates the prediction accuracy of the results predicted using the model for each time interval.

(Appendix 11)
A sports motion analysis support program mounted on a computer including a data storage unit that stores a plurality of data in which moving image data representing a series of motions in a sport and a result of the motion are associated with each other,
On the computer,
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. Learning process to learn, and
A sports motion analysis support program for executing a specific process of specifying a time zone that can specify a time zone in which the degree of improvement in prediction accuracy of a result predicted using a model is large.

  The present invention has been described with reference to the exemplary embodiments, but the present invention is not limited to the above exemplary embodiments. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

Industrial applicability

  INDUSTRIAL APPLICABILITY The present invention is suitably applicable to a sports motion analysis support system that supports motion analysis in sports.

Reference Signs List 1 sports motion analysis support system 2 data storage unit 3 time segmented image extraction unit 4 learning unit 5 prediction unit 6 evaluation unit 7 display unit 8 operation unit 9 data acquisition unit

Claims (10)

A data storage unit that stores a plurality of image data of a moving image representing a series of operations in sports, and a plurality of data in which the results of the operations are associated with each other;
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. A learning unit to learn,
A sports motion analysis support system, comprising: an evaluation unit that calculates prediction accuracy of a result predicted using a model for each time section. The sports motion analysis support system according to claim 1, further comprising a display unit that displays a graph indicating a relationship between a time section and a prediction accuracy of a model corresponding to the time section. The sports motion analysis support system according to claim 1 or 2, wherein the evaluation unit specifies a time section that can specify a time zone in which the degree of improvement in prediction accuracy of a result predicted using the model is large. The display unit displays the prediction accuracy of the model corresponding to the time section specified from the outside, specifies a predetermined number of image data based on a predicted value predicted using the model, and specifies each specified image data. The sports motion analysis support system according to any one of claims 1 to 3, wherein a moving image of the time section is displayed. The result of the operation associated with the image data is a numerical value indicating the grade,
The sports motion analysis support system according to any one of claims 1 to 4, wherein the learning unit learns, for each time interval, a model representing a relationship between the motion in the time interval and a numerical value indicating a score. . A data storage unit that stores a plurality of image data of moving images representing a series of actions in sports, and a plurality of data in which the results of the actions are associated
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation A learning unit to learn,
A sports motion analysis support system, comprising: a specification unit that specifies a time zone that can specify a time zone in which a degree of improvement in prediction accuracy of a result predicted using a model is large. A computer including a data storage unit that stores a plurality of pieces of data in which image data of a moving image representing a series of actions in sports and results of the actions are associated with each other.
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. Learn,
A sports motion analysis support method, wherein a prediction accuracy of a result predicted using a model is calculated for each time section. A computer including a data storage unit that stores a plurality of pieces of data in which image data of a moving image representing a series of actions in sports and results of the actions are associated with each other.
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. Learn,
A sports motion analysis support method characterized by identifying a time zone that can identify a time zone in which the degree of improvement in prediction accuracy of a result predicted using a model is large. A sports motion analysis support program mounted on a computer including a data storage unit that stores a plurality of data in which moving image data representing a series of motions in a sport and a result of the motion are associated with each other,
On the computer,
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. Learning process to learn, and
A sports motion analysis support program that executes an evaluation process that calculates the prediction accuracy of the results predicted using the model for each time interval. A sports motion analysis support program mounted on a computer including a data storage unit that stores a plurality of data in which moving image data representing a series of motions in a sport and a result of the motion are associated with each other,
On the computer,
Using a plurality of the data, for each of a plurality of time segments determined based on the time point representing a predetermined operation, a model representing the relationship between the operation in the time segment and the result corresponding to the operation. Learning process to learn, and
A sports motion analysis support program for executing a specific process of specifying a time zone that can specify a time zone in which the degree of improvement in prediction accuracy of a result predicted using a model is large.

Images