Method, device, electronic device and computer storage medium for recommending fitness movements
By analyzing the user's historical fitness data and real-time fitness videos, and dynamically recommending fitness movements, the problem that existing fitness applications cannot accurately match user conditions is solved, improving fitness effects and safety.
Patent Information
- Application Number
- CN202010933676.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-09-08
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2040-09-08
AI Technical Summary
Existing fitness apps cannot accurately match the user's actual conditions, resulting in the recommended fitness movements that cannot effectively exercise the body and may cause damage to the body.
By obtaining the user's historical fitness data, including the accuracy and frequency of each fitness movement, collecting and analyzing fitness videos in real time, using image recognition algorithms to identify action parameters, calculate accuracy, and determining recommended fitness movements based on user feedback dynamics. Select matching and unexecuted actions from the preset fitness action library as recommendation.
Provide recommendations of fitness movements that are more in line with the actual situation of the user, improve fitness effects, avoid repeated execution of the same movement, reduce muscle and joint burden, enhance the fun and diversity of fitness, and ensure proper exercise of body parts.
Smart Images

Figure CN112071392B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image recognition, and in particular to a method, device, electronic equipment and computer storage medium for recommending fitness movements. Background Art
[0002] As living standards improve, people pay more and more attention to their physical health. As we all know, regular fitness is an effective way to keep healthy. Accordingly, there are many applications on the market to help people keep fit.
[0003] This type of application will select a target fitness plan from a variety of preset fitness plans based on the user's preferences, and recommend each fitness move included in the fitness plan to the user one by one when the user needs to exercise, thereby guiding the user to complete the exercise according to the recommended fitness moves.
[0004] The problem with the prior art is that the preset fitness plans often cannot accurately match the actual conditions of the users, so that the users cannot obtain good fitness effects by performing the recommended fitness movements. Summary of the invention
[0005] Based on the above-mentioned shortcomings of the prior art, the present application provides a fitness action recommendation method, device, electronic device and computer storage medium to provide a fitness action recommendation solution that better matches the actual conditions of the user.
[0006] The first aspect of the present application provides a method for recommending fitness movements, comprising:
[0007] Acquire the user's historical fitness data; wherein the historical fitness data includes the accuracy and frequency of each fitness action performed by the user within a preset time period;
[0008] According to the historical fitness data, a recommended fitness action matching the historical fitness data is determined, and the recommended fitness action is displayed to the user.
[0009] Optionally, determining, based on the historical fitness data, a recommended fitness action that matches the historical fitness data includes:
[0010] According to the historical fitness data, a fitness action having an occurrence frequency lower than a first threshold and a correct rate lower than a second threshold is selected from the preset time period as a reference fitness action;
[0011] A fitness action that matches the reference fitness action and has not been performed by the user within the preset time period is selected from a pre-built fitness action library as a recommended fitness action.
[0012] Optionally, it is characterized in that the obtaining of the user's historical fitness data includes:
[0013] For each fitness action performed by the user within the preset time period, the accuracy of the fitness action is calculated according to the accuracy of the user performing the fitness action each time, and the frequency of occurrence of the fitness action is calculated according to the number of times the fitness action is performed within the preset time period.
[0014] Optionally, also include:
[0015] When the user is exercising, the user's exercise video is collected in real time, and the accuracy of each exercise performed by the user is obtained based on the exercise video analysis.
[0016] Optionally, obtaining the accuracy of each fitness action performed by the user according to the fitness video analysis includes:
[0017] For each fitness action performed by the user, the action parameters of the user when performing this fitness action are identified from the fitness video using an image recognition algorithm, and the accuracy of this fitness action is calculated based on the similarity between the action parameters and the standard action parameters of the fitness action.
[0018] Optionally, obtaining the accuracy of each fitness action performed by the user according to the fitness video analysis includes:
[0019] For each fitness action performed by the user, the matching degree between the fitness video of the user performing this fitness action and the teaching video of the fitness action is calculated, and the matching degree is used as the accuracy of the user performing this fitness action.
[0020] Optionally, also include:
[0021] Determining, based on the historical fitness data, a plurality of fitness movements with accuracy rates higher than a second threshold;
[0022] According to the target fitness intensity specified by the user, selecting a corresponding substitute fitness action from a preset fitness action library for each fitness action with a correct rate higher than a second threshold;
[0023] A recommended fitness plan is generated using the selected alternative fitness movements.
[0024] A second aspect of the present application provides a fitness action recommendation device, comprising:
[0025] An acquisition unit, configured to acquire historical fitness data of a user; wherein the historical fitness data includes the accuracy and frequency of each fitness action performed by the user within a preset time period;
[0026] A recommendation unit is used to determine, based on the historical fitness data, a recommended fitness action that matches the historical fitness data, and display the recommended fitness action to the user.
[0027] The third aspect of the present application provides a computer storage medium for storing a computer program. When the computer program is executed, it is specifically used to implement the method for recommending fitness movements provided in any one of the first aspects of the present application.
[0028] A fourth aspect of the present application provides an electronic device, including a memory and a processor;
[0029] Wherein, the memory is used to store computer programs;
[0030] The processor is used to execute the computer program, specifically to implement the method for recommending fitness actions provided in any one of the first aspects of the present application.
[0031] The present application provides a method, device, electronic device and computer storage medium for recommending fitness actions, the method comprising: obtaining the user's historical fitness data; wherein the historical fitness data includes the accuracy and frequency of each fitness action performed by the user within a preset time period; according to the historical fitness data, determining the recommended fitness action that matches the historical fitness data, and displaying the recommended fitness action to the user. The present solution can obtain the user's feedback on the fitness action by analyzing the accuracy and frequency of the user's fitness action, and dynamically determine the recommended fitness action based on the user's feedback. Therefore, compared with the existing solution of recommending according to a fixed fitness plan, the present solution can recommend fitness actions that are more in line with the user's actual situation to the user, so that the user can obtain better fitness results. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0033] Figure 1 A flowchart of a fitness action recommendation method provided in an embodiment of the present application;
[0034] Figure 2 A flowchart of a method for correcting a user's fitness movements in real time provided by an embodiment of the present application;
[0035] Figure 3 A data interaction diagram of a system for correcting a user's fitness movements in real time provided by an embodiment of the present application;
[0036] Figure 4 A flowchart of a method for generating a recommended fitness plan based on historical fitness data provided in an embodiment of the present application;
[0037] Figure 5 A schematic diagram of the structure of a fitness action recommendation device provided in an embodiment of the present application;
[0038] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0040] Fitness movements are designed by relevant professionals according to the parts of the body that need to be exercised, such as squats, weighted squats, jogging, side waist stretching, etc. People who do fitness can exercise the muscles of corresponding parts of the body and achieve specific fitness effects by repeating the movements several times.
[0041] With the popularity of fitness, there are many applications that aim to provide fitness guidance. However, as described in the background technology, such existing applications often select a target fitness plan from several pre-set fitness plans, and then guide users to exercise according to the types and times of fitness movements specified in the target fitness plan. Obviously, the limited number of preset fitness plans cannot accurately match the actual physical condition of each user of the application. To address this problem, the present application provides a fitness movement recommendation method.
[0042] First of all, it should be noted that the fitness action recommendation method provided in any embodiment of the present application can be executed by an application (i.e., a client) installed on the user's terminal device (including but not limited to networked smartphones, personal computers and other electronic devices), or can be executed through interaction between the client and a server set up in the network.
[0043] For example, the client can control the user's terminal device to collect the user's fitness movements in real time, and upload the user's fitness movements to the server. The server can analyze the fitness movements uploaded by the client to obtain the accuracy of the fitness movements and record them, and then determine the recommended fitness movements based on the recorded historical fitness data, and finally control the client to display the recommended fitness movements to the user.
[0044] Please refer to Figure 1 The present application embodiment provides a method for recommending fitness movements, which may include the following steps:
[0045] S101. Obtain historical fitness data of a user.
[0046] The historical fitness data includes the accuracy and frequency of each fitness action performed by the user within a preset time period; the preset time period refers to a period of time in the past with the current time as the end time, for example, the preset time period may be the last half hour, the last 40 minutes, etc. In other words, in a specific actual operation, step S101 may be understood as obtaining the frequency of each fitness action performed by the user within the last half hour (that is, the frequency of the user performing such fitness action within the preset time period) and the accuracy of each fitness action performed by the user within the last half hour.
[0047] The method to obtain the above historical fitness data can be:
[0048] When the user is exercising, the user's fitness video is collected in real time, and the accuracy of each fitness action performed by the user is obtained based on the fitness video analysis, and the analysis results are recorded.
[0049] It is understandable that when exercising, users often repeat the same exercise for many times, for example, 10 squats and 10 push-ups. From the exercise video, each exercise performed by the user can be identified, and the exercise performed by the user this time can be compared with the standard exercise, so as to determine the difference between the exercise performed by the user this time and the standard exercise, that is, to determine the accuracy of the exercise performed this time.
[0050] On this basis, when historical fitness data needs to be obtained, the accuracy of each fitness action performed by the user within the preset time period can be obtained first. Then, for each fitness action, the accuracy of this fitness action each time it is performed can be calculated. The frequency of occurrence of the fitness action can be calculated based on the number of times this fitness action is performed within the preset time period.
[0051] For example, a user does 10 push-ups within a preset time period, and the accuracy of each push-up is recorded as a1, a2, ..., a10. Based on the accuracy of these 10 push-ups, the accuracy rate of the user's fitness exercise such as push-ups within the preset time period can be calculated.
[0052] Optionally, in order to improve the accuracy of the user's fitness action each time, in step S101, a teaching video of the fitness action currently performed by the user can be played in real time while the user is exercising, and the user is prompted to perform the corresponding fitness action following the currently played teaching video. The teaching video of a certain fitness action can be a video obtained by recording the process of a relevant professional (such as a professional fitness coach) performing the fitness action. The teaching video can be stored on a server in the network. When any client needs to play the teaching video of a certain fitness action, it can download the teaching video to be played from the server and play it.
[0053] In step S101, real-time collection of the user's fitness movements can be performed in the following manner:
[0054] When the user is exercising, the client controls the camera of the electronic device on which the client is installed, or other camera devices connected to the electronic device, to shoot the user, thereby recording the user's fitness process as a corresponding fitness video. Obviously, the fitness video recorded in this way shows each fitness action performed by the user during a fitness process. On this basis, the image recognition algorithm can be used to identify from the user's fitness video which specific fitness actions the user has performed during this fitness process, as well as the number of times each fitness action has been performed.
[0055] According to the collected fitness actions of the user, the accuracy of each fitness action performed by the user can be analyzed. The following provides two optional specific implementation methods for analyzing the accuracy of each fitness action performed by the user:
[0056] First, for each fitness action performed by the user, the image recognition algorithm can be used to identify the user's action parameters when performing this fitness action from the fitness video, and the similarity between the calculated user's action parameters and the standard action parameters of the fitness action can be used as the accuracy of the fitness action.
[0057] For a fitness action, its action parameters may include the displacement of various parts of the body (such as arms, waist, legs, head, etc.) and the angle relative to the ground when the user performs the action. If the fitness action performed by the user involves weight, the action parameters may also include the weight borne by the user when performing the action, that is, the user's weight.
[0058] After obtaining the user's fitness video, the image recognition algorithm can be used to identify the location of the user's body parts in each frame of the fitness video, and then calculate the movement parameters such as the displacement and angle of the body parts when the user performs this fitness action based on the location of the body parts.
[0059] Generally, fitness movements that require weight bearing are performed with the aid of corresponding fitness equipment, and fitness equipment is generally equipped with a number of detachable weight blocks with a certain weight. Users can adjust the weight bearing during fitness by increasing or decreasing the number of weight blocks installed on the equipment. Therefore, when executing step S101, the number of weight blocks installed on the fitness equipment can be identified from the user's fitness video, and then the weight bearing during fitness of the user can be calculated based on the number of weight blocks installed and the weight of a single weight block input in advance.
[0060] For any fitness action performed by the user, after obtaining the action parameters of the user when performing this fitness action, the user's action parameters can be compared with the standard action parameters of the fitness action uploaded in advance by relevant professionals (for example, professional fitness coaches) to obtain the similarity between the user's action parameters and the standard action parameters.
[0061] After the user performs a fitness action, the action parameters of this fitness action can be obtained through the above method. Assuming that the three items include leg displacement X1, arm displacement X2, and waist displacement X3, the standard action parameters of the fitness action include leg displacement Y1, arm displacement Y2, and waist displacement Y3. On this basis, the similarity of each action parameter is calculated to obtain the leg displacement similarity Z1, arm displacement similarity Z2, and waist displacement similarity Z3. Taking Z1 as an example, the calculation formula can be:
[0062] Z1=1-|X1-Y1|÷Y1
[0063] After obtaining the similarity of each action parameter, the similarity of each action parameter can be weighted averaged according to a certain weighting coefficient, and the final result is used as the accuracy of the fitness action performed by the user this time.
[0064] Second, for each fitness action performed by the user, the matching degree between the fitness video of the user performing the fitness action and the teaching video of the fitness action can be calculated, and the matching degree is used as the accuracy of the fitness action performed this time.
[0065] Specifically, the matching degree of each frame of the user's fitness video and each corresponding frame in the fitness action teaching video can be calculated frame by frame, and finally the average matching degree of all frames is calculated, and the obtained result is used as the matching degree between the two videos.
[0066] For a frame image in a fitness video and a frame image in a teaching video, first, the resolution of the two frames of images can be adjusted to make the resolution of the two frames of images consistent. Then, in order to avoid interference caused by the difference in the background images of the two frames of images, that is, the images other than the body of the person exercising, the background image parts in the two frames of images can be uniformly replaced with a completely black or completely white background color. In other words, the background image is removed from the two frames of images for which the matching degree needs to be calculated, and only the body image of the person exercising is retained. Then, the existing image matching algorithm can be used to calculate the matching degree between the two frames of images after the background image is removed.
[0067] Optionally, if the user repeatedly performs the same fitness action multiple times within a preset time period, the arithmetic mean of the accuracy of each time the user performs the fitness action can be calculated, and the calculation result can be used as the accuracy rate of the user performing the fitness action. If the user performs a fitness action only once within a preset time period, the accuracy of this fitness action can be directly used as the accuracy rate of the fitness action within the preset time period.
[0068] S102: Determine recommended fitness actions that match the user's historical fitness data.
[0069] An optional way to determine the recommended fitness action may be:
[0070] According to the user's historical fitness data, a fitness action with an occurrence frequency lower than a first threshold and a correct rate lower than a second threshold is selected from a preset time period as a reference fitness action.
[0071] A fitness action that matches the reference fitness action and has not been performed by the user within a preset time period is selected from a pre-built fitness action library as a recommended fitness action.
[0072] For any two fitness movements, if the intensity (which can be understood as the physical strength consumed by performing the fitness movement) and the exercised parts of the two fitness movements are basically the same, then the two fitness movements can be considered to match each other. The intensity and exercised parts of a fitness movement can be pre-set in the system by relevant professionals.
[0073] The frequency of a certain fitness action within a preset time period may be considered to be a ratio obtained by dividing the number of times a user performs this fitness action within the preset time period by the cumulative number of all fitness actions performed by the user within the preset time period.
[0074] S103: Display recommended fitness movements to the user.
[0075] Step S103 may be to play a teaching video of the recommended fitness action on the display device, and prompt the user to follow the teaching video to perform the recommended fitness action.
[0076] Optionally, while playing the teaching video, corresponding voice prompt information can be output according to the playing progress of the teaching video, so as to prompt the user to adjust his posture when the user performs the recommended fitness movements.
[0077] The present application provides a fitness action recommendation method, including: real-time collection of user fitness actions, and analysis to obtain the accuracy of the user's fitness actions; determining recommended fitness actions that match the user's historical fitness data and displaying them to the user; the historical fitness data includes each fitness action performed by the user within a preset time period, the frequency of occurrence of the fitness action, and the accuracy of the fitness action. This solution can obtain the user's feedback on the current fitness action in real time by analyzing the accuracy of the fitness action when the user is exercising, and dynamically determine the recommended fitness action based on the user's feedback. Therefore, compared with the existing solution that recommends according to a fixed fitness plan, this solution can recommend fitness actions that are more in line with the user's actual situation to the user, so that the user can obtain better fitness results.
[0078] Specifically, when the user is exercising, if the accuracy of a certain fitness action is too low (i.e., the accuracy described in step S102 is lower than the second threshold), and the number of executions is too small (i.e., the frequency of occurrence described in step S102 is lower than the first threshold), it is obvious that the corresponding body part cannot be correctly exercised. In addition, long-term execution of fitness actions with too low accuracy is also likely to cause chronic damage to the body. In this regard, the prior art can only push fitness actions specified in the preset fitness plan to the user. Even if the accuracy of the user performing one or several fitness actions is too low, the corresponding correction or supplement cannot be made, resulting in poor fitness effect.
[0079] This solution recommends to users recommended fitness movements that can exercise the same or similar body parts as those exercised by the reference fitness movements for reference fitness movements that meet the above conditions. By executing the recommended fitness movements, users can perform supplementary exercises on the corresponding body parts, thereby making up for the shortcomings of the reference fitness movements that had previously been executed with too low an accuracy rate.
[0080] In addition, if the user's accuracy in executing the recommended fitness movements is high, that is, higher than the second threshold, the recommended fitness movements can be added to the user's fitness plan to replace the reference fitness movements in the original fitness plan. If the user's accuracy in executing a recommended fitness movement is still low, other recommended fitness movements can be determined until a fitness movement suitable for the user is found (that is, the execution accuracy is higher than the second threshold). This can ensure that the accuracy of each fitness movement in the fitness plan that the user ultimately uses is high enough so that each part of the user's body can be properly exercised.
[0081] On the other hand, among the matching fitness movements determined based on the reference fitness movements, there may be some fitness movements that have been performed by users before. If these fitness movements that have been performed are recommended to users, it will cause users to perform the same fitness movements too many times, which is not conducive to increasing the fun of fitness and the diversity of fitness movements. In addition, if the same fitness movement is repeated too many times, it will increase the burden on muscles and joints, easily cause injuries, and is not conducive to improving fitness results.
[0082] In contrast, when determining the recommended fitness action, this solution considers both the matching with the reference fitness action and the fact that it has not been performed within the preset time period, so that the recommended fitness action finally determined can achieve the desired exercise effect of the reference fitness action and avoid the user from repeatedly performing the same fitness action too much. By performing the recommended fitness action, the user can exercise the body parts that the reference fitness action with a low accuracy rate is supposed to exercise, and will not cause damage to the body due to repeated execution of the same fitness action too much.
[0083] Considering that the above-mentioned embodiment obtains the accuracy of the fitness action performed by the user during the fitness process, another embodiment of the present application provides a method for correcting the user's fitness action in real time, aiming to prompt the user according to the accuracy of the user's fitness action during the fitness process to correct the wrong fitness action performed by the user. Figure 2 , the method may include the following steps:
[0084] S201. When a user is exercising, the user's fitness movements are collected in real time, and the accuracy of each fitness movement performed by the user is analyzed.
[0085] S202: Determine whether the accuracy of the user's current fitness action is greater than an accuracy threshold.
[0086] If the accuracy of the user's fitness action is greater than the accuracy threshold, it means that the user's current fitness action is relatively standard, and can achieve good fitness results without injury. Therefore, no prompt is needed, and the process returns to step S201 to continue analyzing the accuracy of the user's subsequent fitness actions.
[0087] If the accuracy of the user's fitness action is less than or equal to the accuracy threshold, it means that the user's current fitness action has serious errors, and neither a good fitness effect can be achieved, nor there is a risk of injury. Therefore, the user needs to be prompted to correct it and execute step S203.
[0088] Optionally, a corresponding accuracy threshold may be set for each fitness action, and when executing step S202, a corresponding accuracy threshold is selected for judgment according to the fitness action currently performed by the user. The accuracy threshold may be set by relevant professionals based on actual experience.
[0089] S203: Output prompt information for correcting the user's fitness movements.
[0090] Specific prompting methods include, but are not limited to, pop-up warnings, outputting voice prompt information, pausing the playing of the instructional video of the fitness action, playing the correction video corresponding to the current fitness action, etc.
[0091] The voice prompt message may be, "The current action is wrong, please stop exercising and rest for 1 minute", or "The current action is wrong, please stop exercising and watch the correction video".
[0092] Among them, the correction video of fitness movements can be understood as a video that adds explanations and guidance on the details of the movements on the basis of the teaching video.
[0093] As mentioned above, the teaching video is a video recorded by a fitness coach performing fitness movements. When playing the teaching video, the user can only imitate the fitness coach's movements based on the video they watched. The correction video can further add the fitness coach's instructions on fitness movements on the basis of the teaching video, for example, which parts of the body should be controlled to exert force when performing a certain fitness movement, which parts should be kept in a relaxed state, etc., or what posture the body should be kept in. Through these instructions, users can find the mistakes in the fitness movements they perform and correct them.
[0094] It should be noted that the process described in the above step S202 and step S203 can be similar to the process described in the above step S203. Figure 1 The fitness action recommendation method shown is executed simultaneously.
[0095] Specifically, after the user starts exercising, the user's exercise movements are collected in real time and the accuracy is analyzed, and it is determined in real time whether the accuracy of the user's current exercise movement is greater than the accuracy threshold. If it is detected at any time during the user's exercise process that the accuracy of the user's current exercise movement is lower than the accuracy threshold, the user is prompted to make corrections. In other words, the above steps S202 and S203 are executed in real time during the user's exercise process. After the user's exercise process ends, for example, after the user has executed all the exercise movements in the exercise plan formulated by the client, the client (or server) can then execute the following steps: Figure 1 Steps S102 and S103 in the illustrated embodiment display the recommended fitness movements to the user, and the user can choose to continue to perform the recommended fitness movements, that is, continue fitness, or can choose to add the recommended fitness movements to the next fitness plan. When the user starts fitness again, the client can formulate a fitness plan for the user that includes the recommended fitness movements.
[0096] The method provided in the embodiments of the present application can be performed as follows: Figure 3 Please refer to the system implementation shown in Figure 3 The system includes a server and a client running on a user's terminal device. As mentioned above, a fitness coach can record the process of performing fitness movements, obtain a teaching video of the corresponding fitness movements, and set corresponding standard movement parameters for each fitness movement. Finally, the teaching video and standard movement parameters are uploaded to the server as teaching files and stored by the server.
[0097] When the user needs to exercise, he opens the client, and the client can download the teaching video of the target fitness action specified by the user from the server, and play the teaching video, prompting the user to follow the teaching video to perform the corresponding fitness action. Specifically, the user can select several fitness actions from a variety of alternative fitness actions as the target fitness action, or select a fitness plan from a variety of preset fitness plans, and use the fitness actions included in the selected fitness plan as the target fitness action.
[0098] While the user is exercising following the instructional video, the client calls the local camera device to shoot the user's exercise process, obtains the exercise video, and uploads the exercise video to the server, which is equivalent to the server executing Figure 1 In the illustrated embodiment, step S101 is a step of collecting user's fitness movements.
[0099] After the server obtains the user's fitness video, it executes the method described in the above embodiment to analyze the accuracy of each fitness action of the user. The server records each fitness action performed by the user during fitness and the corresponding accuracy, and determines in real time whether the accuracy of the user's fitness action is greater than the accuracy threshold. When it is detected that the accuracy of any fitness action performed by the user is less than or equal to the accuracy threshold, the server can send a notification to the client, thereby triggering the client to output a prompt message to correct the user's fitness action.
[0100] After the user completes the current fitness process, the server can analyze the user's historical fitness data during this period of time, that is, the frequency and accuracy of the user performing each fitness action during this period, and determine the recommended fitness action from the server's database based on the historical fitness data. After determining the recommended fitness action, the server can send the teaching video of the recommended fitness action to the client, so that the client can play the teaching video of the recommended fitness action.
[0101] Optionally, the fitness action recommendation method and correction method provided in the embodiment of the present application can also be implemented by the client in an offline mode, provided that the terminal device has sufficient computing power. Specifically, the server can send multiple fitness action teaching videos and standard action parameters to the client in advance. When the user needs to exercise, the client directly plays the locally stored fitness action teaching video to the user, and determines whether to correct the user's fitness action based on the accuracy of each fitness action performed by the user. After the user's fitness is finished, the recommended fitness action is determined from the multiple locally stored fitness actions based on the user's historical fitness data, and the locally stored recommended fitness action teaching video is played.
[0102] The present application also provides a method for determining a recommended fitness plan that meets the needs of a user based on historical fitness data. Figure 4 , the method may include the following steps:
[0103] S401. Determine, based on historical fitness data, a plurality of fitness movements with accuracy rates higher than a second threshold.
[0104] The method for determining the accuracy of each fitness action performed by the user can refer to the corresponding steps in the previous text and will not be described in detail here.
[0105] S402: According to the target fitness intensity specified by the user, a corresponding replacement fitness action is selected from a preset fitness action library for each fitness action with a correct rate higher than a second threshold.
[0106] The above-mentioned alternative fitness action refers to a fitness action that exercises the same or similar body part as the body part exercised by the fitness action with a correctness higher than the second threshold determined in step S402, and the fitness intensity of the action is the target fitness intensity specified by the user.
[0107] Fitness intensity can be divided into three levels: high intensity, medium intensity and low intensity. Accordingly, users can specify one of the high intensity, medium intensity and low intensity as the target fitness intensity.
[0108] Generally, there are many kinds of fitness exercises for exercising a certain body part, and the fitness intensities corresponding to different fitness exercises may be different to a certain extent, that is, the fitness intensity of some fitness exercises may be greater than that of other fitness exercises.
[0109] The determination of the fitness intensity of a fitness action can be achieved by integrating the evaluations of a large number of users on various fitness actions. For example, for two fitness actions H and K that exercise the same body part, the software can collect the evaluations of users who perform fitness action H and the evaluations of users who perform fitness action K respectively. If a large proportion of users in the user group performing fitness action H evaluate that the fitness action is more strenuous to perform (for example, if users evaluate that the action can be performed continuously for a small number of times, then the action can be considered strenuous), and among the user group performing fitness action K, most users evaluate that the fitness action is relatively easy to perform, then the fitness intensity of fitness action H can be set to high intensity, and the fitness intensity of fitness action K can be set to low intensity.
[0110] S403: Generate a recommended fitness plan using the selected alternative fitness movements.
[0111] A fitness plan mainly includes the following data: the types of fitness movements that need to be performed, the order in which each fitness movement is performed, that is, doing one movement first and then another, and the number of times each movement is performed during a fitness process, which may specifically include how many sets are performed, how many times each set is performed, etc.
[0112] In the recommended fitness plan, the execution order of the selected multiple replacement fitness actions may be consistent with the execution order of multiple fitness actions that the user has already performed with a correct rate higher than the second threshold.
[0113] For example, fitness actions with a correct rate higher than the second threshold include actions A, B, and C. The alternative fitness actions corresponding to these three actions are D, E, and F. In previous fitness processes, users generally performed these three fitness actions in the order of B, A, and C. Then, in the generated recommended fitness plan, the execution order of the alternative fitness actions D, E, and F can be consistent with the execution order of the corresponding fitness actions with a correct rate higher than the second threshold, that is, it is recommended to execute them in the order of E, D, and F.
[0114] The number of recommended alternative fitness movements to be performed during a fitness session (specifically including the number of sets and the number of times in each set) can be consistent with the number of corresponding fitness movements with a correct rate higher than the second threshold, or can be adjusted based on the number of corresponding fitness movements with a correct rate higher than the second threshold according to the target fitness intensity specified by the user.
[0115] For example, if the target fitness intensity is high intensity, the number of executions of the corresponding fitness actions with a correct rate higher than the second threshold may be appropriately increased to obtain the number of executions of the alternative fitness actions in the recommended fitness plan.
[0116] If the target fitness intensity is low intensity, the execution times of the corresponding fitness actions with a correct rate higher than the second threshold may be appropriately adjusted downward to obtain the execution times of the replacement fitness actions in the recommended fitness plan.
[0117] By executing the method provided in the above embodiment, it is possible to select fitness movements that meet the user's fitness intensity needs and have a high expected accuracy rate for the user, and use the selected fitness movements to generate a recommended fitness plan that meets the user's personalized needs, so that each user's fitness plan can not only adapt to his or her actual physical condition, but also meet the user's needs for fitness intensity.
[0118] In combination with the fitness action recommendation method provided in the embodiment of the present application, the embodiment of the present application also provides a fitness action recommendation device, please refer to Figure 5 , the device may include the following units:
[0119] The acquisition unit 501 is used to acquire the user's historical fitness data.
[0120] The historical fitness data includes the accuracy and frequency of each fitness action performed by the user within a preset time period.
[0121] The recommendation unit 502 is used to determine recommended fitness actions that match the historical fitness data based on the historical fitness data, and display the recommended fitness actions to the user.
[0122] Optionally, when the recommendation unit 502 determines the recommended fitness action matching the historical fitness data according to the historical fitness data, it is specifically used to:
[0123] According to the historical fitness data, a fitness action with a frequency lower than a first threshold and a correct rate lower than a second threshold is selected from a preset time period as a reference fitness action;
[0124] A fitness action that matches the reference fitness action and has not been performed by the user within a preset time period is selected from a pre-built fitness action library as a recommended fitness action.
[0125] Optionally, when the recommendation unit 502 displays the recommended fitness action to the user, it is specifically used to:
[0126] An instructional video of recommended fitness movements is played on the display device.
[0127] When the acquisition unit 501 acquires the user's historical fitness data, it is specifically used to:
[0128] For each fitness action performed by the user within a preset time period, the accuracy of the fitness action is calculated according to the accuracy of each fitness action performed by the user, and the frequency of occurrence of the fitness action is calculated according to the number of times the fitness action is performed within the preset time period.
[0129] Optionally, the device further includes an analyzing unit 503, configured to:
[0130] When the user is exercising, the user's exercise video is collected in real time, and the accuracy of each exercise performed by the user is obtained based on the exercise video analysis.
[0131] When the analysis unit 503 obtains the accuracy of each fitness action performed by the user according to the fitness video analysis, it is specifically used to:
[0132] For each fitness action performed by the user, the image recognition algorithm is used to identify the user's action parameters when performing this fitness action from the fitness video, and the accuracy of this fitness action is calculated based on the similarity between the action parameters and the standard action parameters of the fitness action.
[0133] When the analysis unit 503 obtains the accuracy of each fitness action performed by the user according to the fitness video analysis, it is specifically used to:
[0134] For each fitness action performed by the user, the matching degree between the fitness video of the user performing this fitness action and the teaching video of the fitness action is calculated, and the matching degree is used as the accuracy of the user performing this fitness action.
[0135] Optionally, the device further includes a correction unit 504, configured to:
[0136] If the accuracy of any fitness action performed by the user is lower than a preset accuracy threshold, a prompt message is output; wherein the prompt message is used to prompt the user to correct the fitness action currently being performed.
[0137] Optionally, the recommendation unit 502 may also be used to:
[0138] Determining, based on the historical fitness data, a plurality of fitness movements with accuracy rates higher than a second threshold;
[0139] According to the target fitness intensity specified by the user, selecting a corresponding substitute fitness action from a preset fitness action library for each fitness action with a correct rate higher than a second threshold;
[0140] Generate a recommended fitness plan using the selected alternative fitness movements.
[0141] The present application provides a fitness action recommendation device, wherein an acquisition unit 501 is used to acquire the user's historical fitness data. The historical fitness data includes the accuracy and frequency of each fitness action performed by the user within a preset time period. The recommendation unit 502 is used to determine the recommended fitness action that matches the historical fitness data based on the historical fitness data, and display the recommended fitness action to the user. This solution can obtain the user's feedback on the fitness action by analyzing the accuracy of the user's fitness action, and dynamically determine the recommended fitness action based on the user's feedback. Therefore, compared with the existing solution that recommends according to a fixed fitness plan, this solution can recommend fitness actions that are more in line with the user's actual situation to the user, so that the user can obtain better fitness results.
[0142] The present application embodiment is an electronic device, please refer to Figure 6 , the electronic device may include a memory 601 and a processor 602.
[0143] The memory 601 is used to store computer programs.
[0144] The processor 602 is used to execute the above-mentioned computer program, specifically to implement the fitness action recommendation method provided in any embodiment of the present application.
[0145] An embodiment of the present application also provides a computer storage medium for storing a computer program. When the stored computer program is executed, it is used to implement the fitness action recommendation method provided in any embodiment of the present application.
[0146] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.
[0147] It should be noted that the concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.
[0148] The present application can be implemented or used by professionals and technicians. Various modifications to these embodiments will be apparent to professionals and technicians in the field, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for recommending fitness movements, characterized in that: include: When users are exercising, collect their exercise videos in real time; For each fitness action performed by the user, the action parameters of the user when performing this fitness action are identified from the fitness video by using an image recognition algorithm, and the accuracy of this fitness action is calculated based on the similarity between the action parameters and the standard action parameters of the fitness action, wherein the action parameters include the displacement and angle of each part of the body when the user performs this fitness action; Acquire the user's historical fitness data; wherein the historical fitness data includes the accuracy and frequency of each fitness action performed by the user within a preset time period; the accuracy of the fitness action is calculated based on the accuracy of the user performing the fitness action each time; Determine, based on the historical fitness data, recommended fitness movements that match the historical fitness data, and display the recommended fitness movements to the user; wherein the matching fitness movements refer to two fitness movements with the same intensity and exercising the same body parts; If the accuracy rate of the user performing the recommended fitness action is higher than a second threshold, the recommended fitness action is used to replace the fitness action in the fitness plan.
2. The recommendation method according to claim 1, characterized in that: The step of determining, based on the historical fitness data, a recommended fitness action that matches the historical fitness data includes: According to the historical fitness data, a fitness action having an occurrence frequency lower than a first threshold and a correct rate lower than a second threshold is selected from the preset time period as a reference fitness action; A fitness action that matches the reference fitness action and has not been performed by the user within the preset time period is selected from a pre-built fitness action library as a recommended fitness action.
3. The recommendation method according to claim 1, characterized in that: The obtaining of the user's historical fitness data includes: For each fitness action performed by the user within the preset time period, the accuracy of the fitness action is calculated according to the accuracy of the user performing the fitness action each time, and the frequency of occurrence of the fitness action is calculated according to the number of times the fitness action is performed within the preset time period.
4. The recommendation method according to claim 1, characterized in that: The obtaining the accuracy of each fitness action performed by the user according to the fitness video analysis includes: For each fitness action performed by the user, the matching degree between the fitness video of the user performing this fitness action and the teaching video of the fitness action is calculated, and the matching degree is used as the accuracy of the user performing this fitness action.
5. The recommendation method according to claim 1, characterized in that: Also includes: Determining, based on the historical fitness data, a plurality of fitness movements with accuracy rates higher than a second threshold; According to the target fitness intensity specified by the user, selecting a corresponding substitute fitness action from a preset fitness action library for each fitness action with a correct rate higher than a second threshold; A recommended fitness plan is generated using the selected alternative fitness movements.
6. A fitness action recommendation device, characterized in that: include: An analysis unit, used to collect the user's fitness video in real time when the user is exercising; and for identifying, for each fitness action performed by the user, the action parameters of the user when performing this fitness action from the fitness video using an image recognition algorithm, and calculating the accuracy of this fitness action according to the similarity between the action parameters and the standard action parameters of the fitness action, wherein the action parameters include the displacement and angle of each body part of the user when performing this fitness action; An acquisition unit, configured to acquire historical fitness data of a user; wherein the historical fitness data includes the accuracy and frequency of each fitness action performed by the user within a preset time period; the accuracy of the fitness action is calculated based on the accuracy of each fitness action performed by the user; A recommendation unit is used to determine, based on the historical fitness data, a recommended fitness action that matches the historical fitness data, and display the recommended fitness action to the user; and to use the recommended fitness action to replace the fitness action in the fitness plan if the user's accuracy in performing the recommended fitness action is higher than a second threshold; wherein matching fitness actions refer to two fitness actions with the same intensity and exercising the same parts of the body.
7. A computer storage medium, characterized in that: Used to store a computer program, which, when executed, is specifically used to implement the method for recommending fitness actions as described in any one of claims 1 to 5.
8. An electronic device, characterized in that: including memory and processor; Wherein, the memory is used to store computer programs; The processor is used to execute the computer program, specifically to implement the method for recommending fitness actions as described in any one of claims 1 to 5.
Citation Information
Patent Citations
Man-machine interaction system having bodybuilding action correcting function
CN105561536A
Method and system for correcting body building motions
CN105597282A
Digital body-building system
CN106267734A
Fitness exercise training method and device and computer readable storage medium
CN107694046A
Fitness auxiliary equipment, device and storage medium
CN108310749A
