KR102633168B1 - Apparatus and method for analyzing golf swing - Google Patents

Abstract

A swing analysis device and a swing analysis method are presented, and the swing analysis device according to one embodiment includes an input/output unit for acquiring a swing image of a golfer, a memory storing a program for performing machine learning, and the above. It includes a control unit that performs machine learning on the swing image by executing a program, wherein the control unit detects the golfer's swing error from the swing image using a learned machine learning model and generates a drill image based on the swing error. We recommend the mission that corresponds to the drill video above.

Description

Swing analysis device and swing analysis method {APPARATUS AND METHOD FOR ANALYZING GOLF SWING}

Embodiments disclosed in this specification relate to a swing analysis device and a swing analysis method. More specifically, a device for analyzing a swing image captured of a golfer's swing and providing a drill image and mission according to the analysis results. and methods.

As the popularity of golf increases, the needs of golfers who want to improve their golf skills are increasing.

Meanwhile, the golfer's swing posture is important in order to improve distance, etc., and in order to find and correct problems in the swing posture, golf professionals are on site when the golfer plays golf, so they can receive separate lessons, or watch a swing video recorded of the golf swing. It is possible to convey this to professional golfers and receive correction accordingly.

However, the conventional technique for performing posture correction as described above has the problem that it is difficult to perform correction when there is no professional golfer willing to coach in that it requires a professional golfer, and also, due to the difference in skill of the professional golfers, different questions are asked for each professional golfer. There is a problem with receiving coaching.

In relation to this, Korea Patent Publication No. 10-2001-0102751, which is a prior art document, takes a video of one's swing form and sends the data to the Internet site creator through the Internet or mail, and the Internet site creator who receives the video transfers the video to the professional golfer with whom he or she has a contract. It discloses a method of analyzing and requesting data and uploading the analyzed content to a professional golfer's site.

The above published patent only involves having a golfer film his or her swing and sending the filmed video to a professional golfer for analysis, making it difficult to contact a professional golfer or the cost of receiving posture correction from a professional golfer is burdensome. Golfers have a problem in that it is difficult to correct their posture. Furthermore, there is a problem in that it is difficult to know what effort the golfer should make to correct posture, as it only provides posture correction.

Therefore, the development of technology to solve the above problems became necessary.

Meanwhile, the above-described background technology is technical information that the inventor possessed for deriving the present invention or acquired in the process of deriving the present invention, and cannot necessarily be said to be known technology disclosed to the general public before filing the application for the present invention. .

The purpose of the embodiments disclosed in this specification is to present a swing analysis device and a swing analysis method.

Additionally, the embodiments disclosed in this specification are aimed at presenting an apparatus and method for analyzing swing images using artificial intelligence.

In addition, the embodiments disclosed in this specification are aimed at presenting an apparatus and method for improving a golfer's swing skills by providing a drill image based on swing image analysis results.

In addition, the embodiments disclosed in this specification are aimed at presenting an apparatus and method for improving a golfer's swing skills by providing a mission based on swing image analysis results.

As a technical means for achieving the above-described technical problem, according to an embodiment described in the present specification, a swing analysis device includes an input/output unit for acquiring a swing image of a golfer, and a program for performing machine learning. It includes a stored memory, and a control unit that performs machine learning on the swing image by executing the program, wherein the control unit detects a golfer's swing error from the swing image using a learned machine learning model, and detects the swing error of the golfer. Based on this, a drill video is recommended, and a mission corresponding to the drill video can be recommended.

In addition, as a technical means for achieving the above-described technical problem, according to an embodiment described in the present specification, a swing analysis device is a method of analyzing a golfer's swing, using a learned machine learning model to determine the golfer's swing from the swing image. It may include detecting a swing error, recommending a drill image based on the swing error, and recommending a mission corresponding to the drill image.

According to one of the means for solving the above-described problem, a swing analysis device and a swing analysis method can be provided.

According to one of the means for solving the above-described problem, a device and method for analyzing swing images using artificial intelligence can be proposed.

According to one of the means for solving the above-described problem, a device and method that can improve a golfer's swing skills can be proposed by providing a drill image based on the results of swing image analysis.

According to one of the above-described means of solving the problem, a device and method that can improve a golfer's swing skills can be proposed by providing a mission based on the swing image analysis results.

The effects that can be obtained from the disclosed embodiments are not limited to the effects mentioned above, and other effects not mentioned are clear to those skilled in the art to which the disclosed embodiments belong from the description below. It will be understandable.

1 is a configuration diagram for explaining a swing analysis device according to an embodiment disclosed herein.
2 to 3 are block diagrams for explaining a swing analysis device according to an embodiment disclosed herein.
Figure 4 is an exemplary diagram for explaining a swing analysis device according to an embodiment disclosed herein.
Figure 5 is a flowchart for explaining a swing analysis method according to an embodiment disclosed herein.

Below, various embodiments will be described in detail with reference to the attached drawings. The embodiments described below may be modified and implemented in various different forms. In order to more clearly explain the characteristics of the embodiments, detailed descriptions of matters widely known to those skilled in the art to which the following embodiments belong have been omitted. In addition, in the drawings, parts that are not related to the description of the embodiments are omitted, and similar parts are given similar reference numerals throughout the specification. Also in the following description, like components may be indicated by like reference numerals even if they are shown in different drawings.

Throughout the specification, when a configuration is said to be “connected” to another configuration, this includes not only cases where it is “directly connected,” but also cases where it is “connected with another configuration in between.” In addition, when a configuration “includes” a configuration, this means that other configurations may be further included rather than excluding other configurations, unless specifically stated to the contrary.

Hereinafter, embodiments will be described in detail with reference to the attached drawings.

FIG. 1 is a configuration diagram for explaining a swing analysis device according to an embodiment disclosed herein, and FIGS. 2 and 3 are block diagrams for explaining a swing analysis device according to an embodiment disclosed herein.

The swing analysis device 100 acquires and analyzes swing images of the golfer's swing and recommends drill images and missions accordingly.

This swing analysis device 100 may be implemented in the form of either a user terminal or a server-client system, and when implemented as a server, the components constituting the swing analysis device 100 may be implemented as a plurality of physically separated servers. It can be performed in or on one server.

According to one embodiment, as shown in FIG. 1, the swing analysis device 100 may be implemented as a server-client system, and for this purpose, it includes a user terminal 10 and a server 20, and the user terminal 10 and the server 20 may communicate through the network N.

The user terminal 10 may be implemented as a computer, portable terminal, television, wearable device, etc. that can connect to a remote server through a network (N) or connect to other terminals and servers. Here, the computer includes, for example, a laptop, desktop, laptop, etc. equipped with a web browser, and the portable terminal is, for example, a wireless communication device that guarantees portability and mobility. , PCS (Personal Communication System), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), GSM (Global System for Mobile communications), IMT (International Mobile Telecommunication)-2000, CDMA (Code) All types of handhelds such as Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet), Smart Phone, Mobile WiMAX (Mobile Worldwide Interoperability for Microwave Access), etc. (Handheld)-based wireless communication device may be included. Additionally, television may include IPTV (Internet Protocol Television), Internet TV (Internet Television), terrestrial TV, cable TV, etc. Furthermore, a wearable device is a type of information processing device that can be worn directly on the human body, such as a watch, glasses, accessories, clothing, or shoes, and can connect to a remote server or other terminal via a network directly or through another information processing device. can be connected to

The server 20 may be implemented as a computer capable of communicating over a network with a user terminal 10 installed with an application for interaction with a golfer or a web browser, or may be implemented as a cloud computing server. Additionally, the server 20 may include a storage device capable of storing data or may store data through a third server.

Additionally, the swing analysis device 100 may further include a kiosk (not shown). The kiosk (not shown) can communicate with the server 20 through the network N.

According to the embodiment, such a kiosk (not shown) provides information about the recommended mission, determines whether the mission is performed, whether the mission is successful, and may raise the golfer's level and provide compensation accordingly. Additionally, the kiosk (not shown) can output swing errors by analyzing the swing video, and also output ball data (e.g. spin value, distance, etc.) by analyzing the movement of the detected golf ball or golf club, or Club data regarding the golf club used by the golfer (for example, type of golf club, etc.) can be output.

As shown in FIG. 2, the swing analysis device 100 may include an input/output unit 210, a control unit 220, a communication unit 230, and a memory 240.

The input/output unit 210 may include an input unit for receiving input from the golfer and an output unit for displaying information such as a task performance result or the status of the swing analysis device 100. For example, the input/output unit 210 may include an operation panel that receives golfer input and a display panel that displays a screen.

Specifically, the input unit may include devices that can receive various types of user input, such as a keyboard, physical button, touch screen, camera, or microphone. Additionally, the output unit may include a display panel or a speaker. However, the input/output unit 210 is not limited to this and may include a configuration that supports various inputs and outputs.

Therefore, according to one embodiment, when a golfer's swing is captured by a camera as an input unit, the output unit may output a recommended drill image or mission by analyzing the captured swing image.

Meanwhile, the control unit 220 controls the overall operation of the swing analysis device 100 and may include a processor such as a CPU or GPU. The control unit 220 may control other components included in the swing analysis device 100 to perform operations corresponding to user input received through the input/output unit 210.

For example, the control unit 220 may execute a program stored in the memory 240, read a file stored in the memory 240, or store a new file in the memory 240.

Operations related to the control unit 220 will be described in more detail later with reference to FIG. 4, and FIG. 4 is an exemplary diagram for explaining the swing analysis device 100 according to an embodiment disclosed herein.

The communication unit 230 may perform wired or wireless communication with other devices or networks. To this end, the communication unit 230 may include a communication module that supports at least one of various wired and wireless communication methods. For example, a communication module may be implemented in the form of a chipset.

Wireless communication supported by the communication unit 230 may be, for example, Wi-Fi (Wireless Fidelity), Wi-Fi Direct, Bluetooth, UWB (Ultra Wide Band), or NFC (Near Field Communication). Additionally, wired communication supported by the communication unit 230 may be, for example, USB or HDMI (High Definition Multimedia Interface).

Various types of data, such as files, applications, and programs, can be installed and stored in the memory 240. The control unit 220 may access and use data stored in the memory 240, or may store new data in the memory 240. Additionally, the control unit 220 may execute a program installed in the memory 240. Referring to FIG. 2, for example, a program for performing a swing analysis method may be installed in the memory 240, a program for performing machine learning may be installed, or a learned machine learning model program may be installed. This can be installed.

According to one embodiment, when receiving an input requesting swing posture analysis from a golfer through the input/output unit 210, the control unit 220 may execute a program for a swing analysis method stored in the memory 240.

Meanwhile, referring to FIG. 3, the control unit 220 may include a preprocessing unit 310, an error detection unit 320, a drill recommendation unit 330, and a mission recommendation unit 340.

The preprocessor 310 according to one embodiment may acquire a swing image to extract a golfer's posture swing error from the swing image and perform preprocessing on the acquired swing image.

For example, the preprocessor 310 may acquire a swing image through the input/output unit 110.

Also, for example, the preprocessor 310 may perform preprocessing on the swing image by correcting the brightness, resolution, size, etc. of the swing image. If golf is played outdoors, the brightness and resolution of the swing video may vary depending on the amount of sunlight, etc. each time it is filmed. Additionally, if golf is played on an indoor screen golf, the brightness of the swing video may vary depending on the brightness of the lighting or the recording device, etc. Resolution, size, etc. may vary depending on the screen golf system. Reflecting this reality, preprocessing of swing images can be performed so that the learned machine learning model can effectively output results.

Also, for example, the preprocessor 310 may perform preprocessing by extracting the area containing the golfer from the swing image. When a golfer's swing is filmed, people or geographical features other than the golfer may be filmed, so the area subject to analysis within the filmed video can be extracted as a swing image.

For this purpose, the preprocessor 310 can use a machine learning model. For example, using YOLO (You Only Look Once), it can set a bounding box for an object in the swing image and extract the image within the bounding box. Alternatively, objects can be detected using a deep learning model that learns the characteristic points of golf clubs, etc.

Meanwhile, the error detection unit 320 can detect swing errors by analyzing the swing image. At this time, the error detection unit 320 can detect swing errors by analyzing the pre-processed swing image.

The error detection unit 320 can input a swing image and obtain errors related to the input swing image using a learned machine learning model.

According to one embodiment, the error detection unit 320 may train a machine learning model to output a swing error when a swing image is input.

For example, when a swing video is input, a machine learning model can be trained to output swing errors that indicate problems with the golfer's posture when swinging.

According to another embodiment, the error detection unit 320 may train a machine learning model to output a reference image with a baseline marked when a swing image is input.

For example, when a swing image is input, a machine learning model is trained to output a reference image, but the reference image with the reference line marked is collected by crawling through the web, and when the image with the baseline deleted is input as the swing image, the reference image is You can train a machine learning model to output this.

Also, for example, when a swing image is input, a machine learning model can be trained to extract the golfer's skeleton from the swing image, and the error detection unit 320 extracts and identifies the golfer's joints and body part points accordingly. A reference line connecting joints or body parts can be displayed on the swing image. Accordingly, the error detection unit 320 can output a reference image showing a reference line for identifying problems with the golfer's posture.

In relation to this, as shown in (a) of FIG. 4, when the swing image 400 is input, the error detection unit 320, as shown in (b) of FIG. 4, determines the reference line 411 connecting both shoulders and A reference image 401 showing each of the reference lines 410 connecting the crown of the head and the pelvis can be output.

In addition, the error detection unit 320 can analyze the golfer's swing error based on the reference line, and determine the swing error that may appear by considering the angle between a plurality of reference lines, the distance between the reference lines, and the relative positions between the reference lines. . For example, the error detection unit 320 can analyze that the early extension phenomenon occurred as a swing error through the angle between the reference line 410 and the reference line 411.

The error detection unit 320 may output a swing error related to the swing image and provide a reference image 401 with reference lines 410 and 411 displayed on the swing image, as shown in FIG. 4 . By referring to the reference image 401, the golfer can confirm the basis on which the swing error was derived.

Additionally, the error detection unit 320 can obtain the golfer's feedback on the reference image, and when the feedback is obtained, the machine learning model can be re-trained based on the feedback.

For example, the golfer may determine that the reference lines 410 and 411 in the reference image 401 are displayed incorrectly and correct the reference line, and the error detection unit 320 may output a reference image reflecting the corrected reference line through a machine. The learning model can be retrained.

Meanwhile, the error detection unit 320 can extract one or more swing errors from the swing image.

That is, the error detection unit 320 can extract a plurality of swing errors from the swing image, and when a plurality of swing errors are extracted, one swing error can be selected from among the extracted swing errors.

On the other hand, if one swing error is extracted from the swing image, the error detection unit 320 can transmit the one error to the drill recommendation unit 330 so that it can be used as an error for drill image recommendation.

Hereinafter, for convenience of explanation, the swing error extracted from the swing video will be referred to as 'first swing error', and the swing error selected among the first swing errors for recommendation of the drill video will be referred to as 'second swing error'.

According to one embodiment, the error detection unit 320 may select a second swing error from one or more first swing errors based on swing errors extracted from a golfer's past swing image.

For example, among the first swing errors extracted from the current swing image, one of the remaining first swing errors excluding the first swing error extracted from the past swing image may be selected as the second swing error. That is, when the error detection unit 320 diagnoses the slide and sway phenomenon as the first swing error in the current swing image, it has never extracted sway as the first swing error in the golfer's past swing image. If so, slide excluding sway can be determined as the second swing error.

Or, for example, among the first swing errors extracted from the current swing image, the number of times extracted from the past swing image is counted and the first swing errors are sorted according to the number of extractions, and the first swing error extracted with the greatest number of times is can be selected as the second swing error. When the error detection unit 320 diagnoses the reverse pivot and slide phenomenon as the first swing error in the current swing image, it detects the reverse pivot phenomenon twice and the slide phenomenon five times from the five past swing images of the golfer. If extracted, the slide can be determined as the second swing error. At this time, the past swing image selected for selection of the second swing error may be a swing image that was not used to recommend the drill image.

According to another embodiment, the error detection unit 320 may select a second swing error from one or more first swing errors based on the golfer's past mission performance information. At this time, the past mission performance information is information about the golfer's response to the mission assigned to the golfer in the past, and may include, for example, information about whether the mission was accomplished and the degree of performance when performing the mission.

Therefore, for example, when a golfer performs a mission assigned in the past, the error detection unit 320 stores information about mission performance as past mission performance information, and refers to the past mission performance information among the first swing errors. The golfer can select one of the first swing errors as the second swing error, excluding the swing error related to the mission accomplished. If the golfer receives a squat exercise mission as a past mission and performs the mission, and 'squat exercise mission completed' is stored as past mission performance information, the error detection unit 320 detects the first swing error matching the mission related to the squat exercise. You can select the second swing error from the remaining first swing errors except.

Meanwhile, the error detection unit 320 may determine the third swing error based on the second swing error. At this time, the third swing error is information indicating a physical problem that the golfer has, and may include, for example, lack of flexibility for each body part, lack of strength for each body part, lack of rotational force for each body part, lack of stability, and lack of balance.

To this end, the error detection unit 320 can store the physical problems causing the second swing error as a table. For example, in response to the second swing error of 'sway', 'pelvic distortion' and 'pelvic lack of flexibility' can be stored as physical problems.

According to one embodiment, the error detection unit 320 may provide the golfer with a physical problem that causes the second swing error and determine the problem selected by the golfer as the third swing error. For example, if 'lack of hip muscles' is selected based on the 'left and right arm imbalance', 'pelvic distortion', and 'lack of hip muscles' that are the causes of the second swing error, the error detection unit 320 will select 'hip muscles' ‘Lack’ can be determined as the third swing error.

As another embodiment, the error detection unit 320 may determine a physical problem selected by another golfer that is judged to be similar to the golfer as the third swing error among the physical problems that cause the second swing error. At this time, other golfers similar to the golfer have the same profile (e.g. gender, age, residence, etc.) as the golfer, have the same golf playing period, or are located in the geographic location where swing video analysis is requested (e.g. country club). name, screen golf point, etc.) may be the same golfer. Therefore, for example, when the physical problems matched as the cause of a golfer's second swing error are 'pelvic distortion' and 'lack of back flexibility', the problem selected by other golfers of the same age group as the golfer is 'lack of back flexibility'. In this case, the error detection unit 320 may determine 'lack of waist flexibility' as the third swing error.

As described above, the error detection unit 320 can diagnose the golfer's physical problems by amplifying the second swing error.

And the drill recommendation unit 330 can recommend a drill image corresponding to the swing error.

For this purpose, the drill recommendation unit 330 can store a drill image. At this time, the 'drill video' refers to an image that includes explanations or precautions for correcting golf posture or explanations of exercises for correcting golf posture, and the drill recommendation unit 330 is provided by the swing analysis device 100. Multiple drill videos can be saved by crawling videos posted on an external third server. At this time, the third server is a server that stores a plurality of videos, and may be, for example, a portal site server, an OTT (Over The Top) server, etc.

Additionally, the drill recommendation unit 330 can match the drill image to the swing error and store it.

According to one embodiment, the drill recommendation unit 330 may analyze which swing error the drill image matches, match the drill image to the corresponding swing error, and store it in a table.

To this end, the drill recommendation unit 330 analyzes the title of the video collected as the drill video, the frame included in the drill video, or the audio included in the drill video to determine which swing error the video is related to. . For example, if the title of the collected video includes 'slide', the video is judged to be an video for correcting the slide, and the drill recommendation unit 330 can store the video as a drill video in 'slide' as a swing error. . Or, for example, if the comment 'pelvic distortion' is collected from the audio included in the collected video, the drill recommendation unit 330 may store the corresponding image as a drill image in 'pelvic distortion' as a swing error.

Accordingly, when the golfer's swing error is obtained from the error detection unit 320, the drill image corresponding to the swing error can be searched and output.

According to another embodiment, the drill recommendation unit 330 may learn a machine learning model to output a drill image when a swing error is input.

And when the golfer's swing error is obtained from the error detection unit 320, the drill recommendation unit 330 can determine a drill image to be recommended by inputting the swing error using a learned machine learning model.

If there are a plurality of drill images whose output is determined as described above, the drill recommendation unit 330 may additionally calculate drill preference for each of the plurality of drill images, align the plurality of drill images, and provide the drill images to the golfer.

‘Drill preference’ is a score that indicates the golfer’s preference for drill videos.

According to one embodiment, drill preference may be determined when collecting a drill video, for example, based on at least one of the number of views, recommendations, and likes on a third server platform where the video collected as a drill video is posted. It can be determined, and for example, the drill preference can be increased in proportion to the number of views, recommendations, or likes. Through this, we can provide popular videos to golfers online.

According to another embodiment, drill preference may be determined based on feedback from other golfers who have watched the drill video.

For example, when a drill video is provided to a golfer, the number of views can be counted and the drill preference of the drill video can be calculated according to the number of views. Or, for example, when a drill video is provided to a golfer, the viewing time can be counted and the drill preference of the drill video in question can be calculated according to the viewing time. Or, for example, when a drill video is provided to a golfer, the drill preference of the drill video can be calculated by counting the number of times 'likes' are obtained from the golfer or recommendations to other golfers are entered.

Or, for example, the swing error derived by acquiring and analyzing the swing video of a golfer who received the drill video after a certain period of time has elapsed after the drill video was provided is unrelated to the provided drill video. If this is determined, it is determined that the golfer's posture has been corrected or the swing skill has improved due to the drill video, and the drill preference can be set high. Or, for example, when a drill video is provided and a mission matching the drill video is provided, drill preference can be calculated according to the mission completion rate. For example, the higher the mission achievement rate, the higher the drill preference can be calculated.

As the drill preference is calculated as described above, the drill recommendation unit 330 can sort and recommend drill images. At this time, if a predetermined condition is satisfied, some of the drill images may not be recommended.

For example, as a predetermined condition, among the drill videos sorted based on information about the model of the golfer's user terminal registered in the golfer's account and the usage plan of the golfer's user terminal (for example, whether it is a pay-as-you-go system or a flat-rate plan) , drill images with a capacity exceeding a certain amount can be excluded and only the remaining drill images can be provided.

Or, for example, based on the golfer's disease record registered in the golfer account as a predetermined condition, drill images that may worsen the disease may be excluded from the sorted drill images and the remaining drill images may be provided, and accordingly, the lumbar disc disease may be provided. For golfers, drill videos containing bending of the waist can be excluded. At this time, the determination of whether it is a drill video that can worsen the disease record is made by analyzing comments, comments, audio, images, etc. included in the drill video when collecting drill videos related to a certain movement, and then analyzing the contents of the drill video. The drill video containing this can be classified into positive diseases (i.e., diseases that improve related to the content in the video) and negative diseases (i.e., diseases that worsen related to the content in the video), and can be matched to the corresponding drill video and stored. For example, if we analyze that in a drill video to strengthen back muscles, the movement demonstrator said in audio, 'If you have a herniated disc, please do not do this movement', the drill video can be analyzed as a drill to be provided to golfers with a herniated disc. You can exclude and recommend videos.

Drill preference can be used to retrain a machine learning model. For example, drill images are sorted according to drill preference, and when a swing error is input, machine learning is used to output drill images within a certain rank according to the sorted order. The model can be retrained.

Meanwhile, the mission recommendation unit 340 may recommend a mission based on the drill video.

At this time, the mission is information that requests the golfer to perform a certain action, and may include information about a specific action, number of actions, action performance period, etc. Additionally, the mission may be additionally stored, for example, by matching one or more levels, and even if the actions are the same, the intensity of the action may vary as the number of actions or the operation performance period varies for each level. Therefore, if the level of the same mission increases, the number of repetitions can increase even if the same body part is trained. For example, if 10 ‘basic squats’ are level 1, 20 ‘basic squats’ can be set to level 2. .

Therefore, for example, the mission recommendation unit 340 recommends as a mission the 'squat' movement, 3 sets of 20 times, for 2 weeks, or the basic squat as a level 1 squat, 3 sets of 20 times, 2 You can recommend something to do during the week as a mission.

These missions may be preset in the swing analysis device 100 or may be collected by crawling through the web. When a mission is saved by being preset or collected, the intensity may be stored differently depending on the level of each mission.

Therefore, for example, a mission may be matched to a drill image and stored as a table, and the mission recommendation unit 340 may recommend a mission matched to the drill image recommended by the drill recommendation unit 330.

Or, for example, using a machine learning model learned to output a mission when a drill image is input, the mission recommendation unit 340 may recommend a mission that matches the drill image recommended by the drill recommendation unit 330. You can. Therefore, the drill recommendation unit 330 can output a mission to be recommended by inputting the drill image into the learned machine learning model.

The mission recommendation unit 340 may calculate mission preference for each of the plurality of missions for which recommendation has been determined as described above, arrange the plurality of missions in order of high mission preference, and provide the mission preference to the golfer.

At this time, mission preference can be calculated based on missions recommended to the golfer in the past, and the mission recommendation unit 340 can calculate mission preference according to the degree to which the mission is related to missions recommended in the past. For example, if the mission is related to a mission that was recommended in the past, the mission preference can be calculated low, and if the mission is unrelated to a mission that was recommended in the past, the mission preference can be calculated high. Therefore, for example, if a mission is related to lower body training and a previously recommended mission is related to upper body training or flexibility, the mission preference for that mission can be set high.

Additionally, mission preference can be calculated in proportion to the number of selections made by the golfer or other golfers, and the mission recommendation unit 204 determines the number of times a mission related to the mission has been selected by the golfer, or other golfers for whom the same drill video as the golfer has been recommended. The mission preference for the corresponding mission can be calculated based on the number of times the corresponding mission is selected. For example, if the number of missions related to 'Plank' was selected by golfers in the past is higher than other missions, the highest mission preference for the 'Plank 1 minute per day for 2 weeks' mission can be determined among the plurality of missions. Or, for example, when there are multiple other golfers who have been recommended the same drill video as the golfer, the highest mission preference can be set for the mission selected by the largest number of golfers.

Additionally, mission preference can be calculated based on past mission performance information of missions related to the relevant mission. The mission recommendation unit 340 may calculate the mission preference as low if it determines that the golfer's degree of achievement of the mission related to the mission is low, and may calculate the mission preference as high if it determines that the golfer's degree of achievement in the mission related to the mission is high. Therefore, for example, for the mission 'Plank 1 minute every day for 2 weeks', if it is determined that a past golfer performed plank for 1 minute once a week for the mission 'Plank 1 minute 3 times a week for 2 weeks', the golfer for that mission It is determined that the degree of achievement is low, and the mission preference can be calculated to have a lower value than the mission preference of other missions.

In this way, the mission recommendation unit 340 can sort missions according to mission preference and select and recommend missions within a predetermined rank.

Additionally, the mission recommendation unit 340 may recommend a mission to the golfer and then determine whether the golfer performs the mission.

That is, the mission recommendation unit 340 can determine whether the golfer performs the mission selected by the golfer among the recommended missions.

Whether or not the mission is performed can be determined using the function of the electronic terminal 10 included in the swing analysis device 100 implemented as a server-client system.

For example, it can be judged by analyzing the mission performance video obtained by filming a golfer performing a mission. For example, when a golfer who was given a mission to do 20 squats took a squat exercise video and uploaded it as a mission performance video. , the mission recommendation unit 340 can count the number of squats performed by the golfer by analyzing the mission performance video. Or, for example, as a golfer carrying out a mission carries out a mission while holding a user terminal, the movement of the golfer is detected by a gyro sensor built into the user terminal, and the number of squats performed by the golfer can be counted accordingly.

Alternatively, whether or not the mission is performed can be determined using a kiosk (not shown) included in the swing analysis device 100.

Alternatively, whether or not the mission is performed can be determined through, for example, a sensor installed in the screen golf system. For example, whether the mission is performed can be determined by detecting the golfer's movement through a vision sensor.

Alternatively, whether or not the mission is performed can be determined through, for example, a sensor built into the golf club. Therefore, in order to determine whether the mission, such as barbell squats or deadlifts, which are exercised by using the golf club like a barbell, is performed, the golfer It is possible to determine whether the mission is being performed by detecting whether the golf club is held horizontally and moved.

By determining whether the golfer has performed the mission in this way, the mission recommendation unit 340 can provide compensation according to mission accomplishment.

For example, when the golfer achieves a mission, the mission recommendation unit 340 may provide virtual currency such as a certain point or an item as a reward, or may raise the golfer's level. Examples of compensation are not limited to the examples described above.

On the other hand, for example, if the golfer fails to achieve the mission, the mission recommendation unit 340 may impose a penalty, for example, demote the golfer's level, or deduct a certain amount of points held by the golfer. Also, for example, the mission recommendation unit 340 may provide comments to encourage mission performance.

Additionally, the mission recommendation unit 340 may determine whether the golfer has performed the mission and increase the mission level to be matched to the golfer.

For example, if the mission recommendation unit 340 determines that the golfer has performed level 1 of the mission, it raises the level of the mission corresponding to the golfer by one level and provides a level 2 mission the next time the same mission is recommended to the golfer. can do.

Or, for example, if the mission recommendation unit 340 determines that the golfer has performed level 1 of the mission, it immediately provides the next level mission in succession so that the golfer can perform the level 2 mission and performs the mission accordingly. You can repeatedly train the body part you want to train.

On the other hand, the mission recommendation unit 340 may additionally set a lesson curriculum suitable for the golfer based on the drill video or mission and provide lesson content according to the curriculum to the golfer. The lesson content may be, for example, a lesson video, or it may provide a virtual golf course for practicing certain shots and simulate the golfer's shots accordingly on the golf course.

The control unit 220 including the mission recommendation unit 340 may further include a chatting unit (not shown) that provides an interface for chatting with the golfer, and the chatting unit (not shown) is an artificial intelligence chatbot. It can be implemented. In other words, an artificial intelligence chatbot that can process text or voice into natural language and provide output values accordingly, for example, information about the degree to which the golfer watched the drill video, whether the golfer performed the mission, and the golfer's progress rate in the lesson curriculum. Can be provided as text or audio.

Meanwhile, the machine learning model in each component constituting the control unit 220 may be implemented as a network model such as, for example, a convolutional neural network (CNN), a recurrent neural network (RNN), or a deep neural network (DNN). . At this time, the machine learning models in each of the preprocessing 310, error detection unit 320, drill recommendation unit 330, and mission recommendation unit 350 may be implemented in different networks or may be implemented in the same network.

As described above, an input/output unit 110 for acquiring a swing image of a golfer, a memory 140 for storing a program for performing machine learning, and machine learning on the swing image by executing the program The swing analysis device 100, which includes a control unit 120 that performs, detects the golfer's swing error from the swing image using a learned machine learning model, recommends a drill image based on the swing error, and You can recommend missions that correspond to the drill video. Accordingly, the swing analysis device 100 can contribute to improving the golfer's golf skills by correcting the golfer's swing posture.

At this time, the swing analysis device 100 can analyze the golfer's swing by filming the golfer's golf play on the golf field, and can also analyze the golfer's swing by filming the golfer's golf play on the screen golf system. You can.

When the camera detects at least one of the golfer, the golf ball, and the golf club, a swing image may be generated as the camera performs filming of the golfer's swing motion, or the camera may continuously film the golfer and frame with the swing motion. You can create a swing image by extracting .

In this regard, the screen golf system may include a camera installed to film the at-bats at which a golfer can hit a golf ball, and the camera is installed in a simulator where all data necessary for virtual golf simulation is stored and processed, etc. , or may be installed in a sensor that communicates with the simulator, or may be implemented as a separate device. When a camera or simulator detects at least one of a golfer, a golf ball, and a golf club through a sensor, a swing image may be generated as the camera performs filming of the golfer's swing motion, or the camera may continuously film the golfer, You can create a swing video by extracting frames with swing motion.

The swing analysis device 100 according to an embodiment described in this specification is described in detail as being performed by analyzing the swing image as the golfer's swing is filmed on a screen golf system or golf field, but is not necessarily limited to this and is performed by analyzing the golfer's swing. When a swing video is obtained by filming, it can be applied to any type of system or device that analyzes the swing video.

Meanwhile, Figure 5 is a flowchart for explaining a swing analysis method according to an embodiment. The swing analysis method shown in FIG. 5 includes time-series processing steps in the swing analysis device 100 shown in FIGS. 1 to 4. Therefore, even if the content is omitted below, the content described above regarding the swing analysis device 100 can also be used in the swing analysis method according to the embodiment shown in FIG. 5.

As shown in FIG. 5, the swing analysis device 100 may receive a request for swing analysis from the golfer (S510).

For example, upon obtaining a request for analysis of a swing video uploaded by a golfer or detecting movement of at least one of the movement of the golfer, golf ball, and golf club, the swing analysis device 100 receives a request for swing analysis. You can judge.

The swing analysis device 100 can detect swing errors by analyzing the swing image (S520).

For example, the swing analysis device 100 may extract one or more first swing errors and select one of the one or more first swing errors as the second swing error. At this time, the swing analysis device 100 selects the second swing error from the one or more first swing errors based on the swing error extracted from the golfer's past swing image or selects the first swing error based on the golfer's past mission performance information. You can select the second swing error.

Also, for example, the swing analysis device 100 may determine a third swing error indicating a golfer's physical problem based on the second swing error.

As described above, when a swing error is detected, the swing analysis device 100 can recommend a drill image based on the detected swing error (S530).

For example, the swing analysis device 100 may recommend a drill image based on at least one of the first swing error, the second swing error, and the third swing error.

And based on the recommended drill image, the swing analysis device 100 can recommend a mission (S540) and monitor whether the golfer performs the recommended mission. Accordingly, if it is determined that the golfer has performed the mission (S550), the swing analysis device 100 may provide compensation to the golfer (S560).

According to the swing analysis method according to an embodiment described in this specification, a golfer can easily discover problems with his or her swing pose and make efforts to improve it.

The virtual golf simulation method described above can also be implemented in the form of a computer-readable medium that stores instructions and data executable by a computer. At this time, instructions and data can be stored in the form of program code, and when executed by a processor, they can generate a certain program module and perform a certain operation. Additionally, computer-readable media can be any available media that can be accessed by a computer and includes both volatile and non-volatile media, removable and non-removable media. Additionally, computer-readable media may be computer recording media, which are volatile and non-volatile implemented in any method or technology for storage of information such as computer-readable instructions, data structures, program modules, or other data. It can include both volatile, removable and non-removable media. For example, computer recording media may be magnetic storage media such as HDDs and SSDs, optical recording media such as CDs, DVDs, and Blu-ray discs, or memory included in servers accessible through a network.

The virtual golf simulation method described above may be implemented as a computer program (or computer program product) including instructions executable by a computer. A computer program includes programmable machine instructions processed by a processor and may be implemented in a high-level programming language, object-oriented programming language, assembly language, or machine language. . Additionally, the computer program may be recorded on a tangible computer-readable recording medium (eg, memory, hard disk, magnetic/optical medium, or solid-state drive (SSD)).

The virtual golf simulation method described above can be implemented by executing the computer program described above by a computing device. The computing device may include at least some of a processor, memory, a storage device, a high-speed interface connected to the memory and a high-speed expansion port, and a low-speed interface connected to a low-speed bus and a storage device. Each of these components is connected to one another using various buses and may be mounted on a common motherboard or installed in some other suitable manner.

Here, the processor can process instructions within the computing device, such as displaying graphical information to provide a graphic user interface (GUI) on an external input or output device, such as a display connected to a high-speed interface. These may include instructions stored in memory or a storage device. In other embodiments, multiple processors and/or multiple buses may be utilized along with multiple memories and memory types as appropriate. Additionally, the processor may be implemented as a chipset consisting of chips including multiple independent analog and/or digital processors.

Memory also stores information within a computing device. In one example, memory may be comprised of volatile memory units or sets thereof. As another example, memory may consist of non-volatile memory units or sets thereof. The memory may also be another type of computer-readable medium, such as a magnetic or optical disk.

And the storage device can provide a large amount of storage space to the computing device. A storage device may be a computer-readable medium or a configuration that includes such media, and may include, for example, devices or other components within a storage area network (SAN), such as a floppy disk device, a hard disk device, an optical disk device, Or it may be a tape device, flash memory, or other similar semiconductor memory device or device array.

The term '~unit' used in the above embodiments refers to software or hardware components such as FPGA (field programmable gate array) or ASIC, and the '~unit' performs certain roles. However, '~part' is not limited to software or hardware. The '~ part' may be configured to reside in an addressable storage medium and may be configured to reproduce on one or more processors. Therefore, as an example, '~ part' refers to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, and procedures. , subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables.

The functions provided within the components and 'parts' may be combined into a smaller number of components and 'parts' or may be separated from additional components and 'parts'.

In addition, the components and 'parts' may be implemented to regenerate one or more CPUs within the device or secure multimedia card. The above-described embodiments are for illustrative purposes, and those skilled in the art will recognize that the above-described embodiments can be easily modified into other specific forms without changing the technical idea or essential features of the above-described embodiments. You will understand. Therefore, the above-described embodiments should be understood in all respects as illustrative and not restrictive. For example, each component described as unitary may be implemented in a distributed manner, and similarly, components described as distributed may also be implemented in a combined form.

The scope sought to be protected through this specification is indicated by the patent claims described later rather than the detailed description above, and should be interpreted to include the meaning and scope of the claims and all changes or modified forms derived from the equivalent concept. .

100: swing analysis device
10: Electronic terminal 20: Server
210: input/output unit 220: control unit
230: Communication unit 240: Memory
310: preprocessing unit 320: error detection unit
330: Drill recommendation unit 340: Mission recommendation unit

Claims (16)

As a swing analysis device,
An input/output unit for acquiring a swing image of a golfer;
Memory in which programs for performing machine learning are stored; and
It includes a control unit that performs machine learning on the swing image by executing the program,
The control unit,
One or more first swing errors are extracted from the swing image using a learned machine learning model, and one of the first swing errors is selected as a second swing error, based on the golfer's past mission performance information. Select the 2nd swing error from the 1st swing error,
A swing analysis device that recommends a drill image based on the second swing error and recommends a mission corresponding to the drill image. delete According to paragraph 1,
The control unit,
A swing analysis device that selects a second swing error from the first swing error based on a swing error extracted from the golfer's past swing image. delete According to paragraph 1,
The control unit,
A swing analysis device that determines a third swing error indicating a physical problem of the golfer based on the second swing error and recommends a drill image based on the third swing error. According to paragraph 1,
The swing analysis device communicates with a third server,
The control unit,
Select a drill video to recommend from among the drill videos collected by crawling the videos stored in the third server, and select the drill video based on at least one of the number of views, recommendations, and likes for each drill video on the third server. A swing analysis device. In paragraph 1.
The control unit,
A swing analysis device that recommends a mission corresponding to the drill video, but recommends a mission selected based on a mission previously recommended to the golfer. According to paragraph 1,
The control unit,
A swing analysis device that determines whether the golfer has performed the mission and provides compensation to the golfer as a result of the determination. A method for a swing analysis device to analyze a golfer's swing, comprising:
Detecting a golfer's swing error from a swing image using a learned machine learning model;
Recommending a drill image based on the swing error; and
Including the step of recommending a mission corresponding to the drill video,
The step of detecting the swing error is,
Extracting one or more first swing errors from the swing image; and
Comprising the step of selecting a second swing error from the first swing error based on the golfer's past mission performance information,
The step of recommending the drill video is,
A swing analysis method comprising recommending a drill image based on the second swing error. delete According to clause 9,
The step of selecting the second swing error is,
A swing analysis method comprising selecting a second swing error from the first swing error based on a swing error extracted from the golfer's past swing image. delete According to clause 9,
The step of detecting the swing error is,
It further includes determining a third swing error indicating a physical problem of the golfer based on the second swing error,
The steps for recommending the drill video are:
A swing analysis method comprising recommending a drill image based on the third swing error. According to clause 9,
The swing analysis device communicates with a third server,
The step of recommending the drill video is,
Select a drill video to recommend from among the drill videos collected by crawling the videos stored in the third server, and select the drill video based on at least one of the number of views, recommendations, and likes for each drill video on the third server. A swing analysis method including the steps of: In paragraph 9.
The steps to recommend the above mission are:
A swing analysis method comprising recommending a mission corresponding to the drill video, but recommending a mission selected based on a mission previously recommended to the golfer. According to clause 9,
determining whether the golfer performs the mission; and
A swing analysis method further comprising providing compensation to the golfer as a result of the determination.