CN114821803A - Human motion scoring method, device, computer equipment and storage medium - Google Patents

Abstract

The invention discloses a scoring method and device for human body movement, computer equipment and a storage medium, relates to the technical field of movement assessment, and is used for effectively monitoring the movement of a user and improving the accuracy and efficiency of scoring the human body movement. The main technical scheme of the invention is as follows: acquiring a skeleton sequence to be identified and a standard template skeleton sequence of a user, wherein the skeleton sequence to be identified comprises a plurality of human skeletons arranged according to a time sequence; the standard template skeleton sequence comprises a plurality of standard human skeletons arranged according to a time sequence; calculating difference values of corresponding human skeletons in the skeleton sequence to be identified and the standard template skeleton sequence; inputting the difference values into an action mode recognition model to obtain the values of all human skeletons in the skeleton sequence to be recognized; and determining the motion score of the user according to the score of each human body skeleton in the skeleton sequence to be recognized.

Description

Human motion scoring method, device, computer equipment and storage medium

technical field

The present invention relates to the technical field of motion evaluation, and in particular, to a method, device, computer equipment and storage medium for scoring human motion.

Background technique

Human Movement Science is a science that studies how the Human Movement System (HMS) functions in interdependent and interrelated mechanisms. The results of this research are widely used in sports medicine-related diagnosis and treatment. The main research content of HMS is the static and dynamic analysis of human movements to achieve a series of pathological assessments. The main contents include static posture assessment, movement patterns, joint range of motion assessment, muscle strength, etc. Among them, human motion score is the current research difficulty. The study of physical activity score is not only used for the diagnosis of motor dysfunction, but also in the exercise health assessment of some special populations, such as the exercise prescription design for the assessment of motor function status of adolescents, pregnant women and the elderly.

The existing human motion score also relies on human subjective judgment, and cannot solve the problem of objectivity, that is, the accuracy and efficiency of the existing human motion score are low.

SUMMARY OF THE INVENTION

The present invention provides a human motion scoring method, device, computer equipment and storage medium, which are used to effectively monitor user motion and improve the accuracy and efficiency of human motion scoring.

An embodiment of the present invention provides a method for scoring human motion, the method comprising:

Obtain the user's skeleton sequence to be identified and a standard template skeleton sequence, the to-be-identified skeleton sequence includes a plurality of human skeletons arranged in chronological order; the standard template skeleton sequence includes a plurality of standard human skeletons arranged in chronological order ;

Calculate the difference value of the corresponding human skeleton in the to-be-identified skeleton sequence and the standard template skeleton sequence;

The difference value is input into the action pattern recognition model, and the score of each human skeleton in the skeleton sequence to be recognized is obtained;

The user's motion score is determined according to the scores of each human skeleton in the to-be-identified skeleton sequence.

An embodiment of the present invention provides a human motion scoring device, the device comprising:

The acquisition module is used to acquire the user's skeleton sequence to be identified and a standard template skeleton sequence, the to-be-identified skeleton sequence includes a plurality of human skeletons arranged in chronological order; the standard template skeleton sequence includes a plurality of human skeletons arranged in chronological order the standard human skeleton;

A calculation module, used to calculate the difference value of the corresponding human skeleton in the skeleton sequence to be identified and the standard template skeleton sequence;

The computing module is further configured to input the difference value into the action pattern recognition model to obtain the score of each human skeleton in the skeleton sequence to be identified;

A determination module, configured to determine the motion score of the user according to the scores of each human skeleton in the to-be-identified skeleton sequence.

A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor implementing the above-mentioned method for scoring human motion when the computer program is executed.

A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, implements the above-mentioned method for scoring human motion.

The method, device, computer equipment and storage medium for scoring human motion provided by the present invention firstly obtain the user's skeleton sequence to be identified and a standard template skeleton sequence, and then calculate the corresponding human skeleton in the to-be-identified skeleton sequence and the standard template skeleton sequence. The difference value is then input into the action pattern recognition model to obtain the score of each human skeleton in the skeleton sequence to be recognized; finally, the user's motion score is determined according to the score of each human skeleton in the skeleton sequence to be recognized. Wherein, the to-be-identified skeleton sequence includes multiple human skeletons arranged in chronological order; the standard template skeleton sequence includes multiple standard human skeletons arranged in chronological order. Compared with the existing human body motion scoring which relies on human subjective judgment, the present invention does not require manual participation in the judgment in the process of human body motion scoring. The user's motion score is obtained, so that the present invention can effectively monitor the user's motion, and effectively improve the accuracy and efficiency of the human body motion score.

Description of drawings

In order to illustrate the technical solutions of the embodiments of the present invention more clearly, the following briefly introduces the drawings that are used in the description of the embodiments of the present invention. Obviously, the drawings in the following description are only some embodiments of the present invention. , for those of ordinary skill in the art, other drawings can also be obtained from these drawings without creative labor.

1 is a schematic structural diagram of a scoring system for human motion in an embodiment of the present invention;

2 is a flowchart of a method for scoring human motion in an embodiment of the present invention;

3 is a flowchart of determining a user's exercise score in an embodiment of the present invention;

Fig. 4 is the principle block diagram of the scoring apparatus of human body movement in one embodiment of the present invention;

FIG. 5 is a schematic diagram of a computer device in an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

The following embodiments can be applied to the scoring system for human movement shown in FIG. 1 . The scoring system for human movement shown in FIG. 1 includes a monitoring device, an interactive device, and a computing device, and the computer device communicates with the interactive device and the monitoring device. , obtain the user's motion video stream through the interactive device and the monitoring device, and then calculate the difference value of the corresponding human skeleton in the to-be-identified skeleton sequence and the standard template skeleton sequence according to the user's to-be-identified skeleton sequence and the standard template skeleton sequence. The difference value is input into the action pattern recognition model to obtain the score of each human skeleton in the skeleton sequence to be recognized; finally, the user's motion score is determined according to the score of each human skeleton in the skeleton sequence to be recognized. Wherein, the to-be-identified skeleton sequence includes multiple human skeletons arranged in chronological order; the standard template skeleton sequence includes multiple standard human skeletons arranged in chronological order.

Among them, the monitoring device can be a customized watch that integrates heart rate monitoring and inertial sensors. The interactive device may include: display: used to display measurement information, monitoring status, action completion status and other information; card reader: used to collect basic user information; camera: real-time capture of user's motion video or action photos; printer: output motion diagnosis Reports; a voice interaction system that reminds users to correct actions based on analysis results, and can issue alarms in emergency situations. The computing device is used to obtain the user's skeleton sequence to be recognized and the standard template skeleton sequence from the motion video stream, and then calculate the difference value between the skeleton sequence to be recognized and the standard template skeleton sequence corresponding to the human skeleton, and then input the difference value to the action pattern recognition. In the model, the score of each human skeleton in the skeleton sequence to be identified is obtained; finally, the user's motion score is determined according to the score of each human skeleton in the skeleton sequence to be identified.

It should be pointed out that the structure shown in FIG. 1 does not constitute a limitation on the scoring system for human movement. In other embodiments, the scoring system for human movement may include fewer or more components than those shown in the figure, or a combination of certain components, or different component arrangements.

As shown in FIG. 2 , an embodiment of the present invention provides a method for scoring human body movement, including the following steps:

S10: Obtain the user's skeleton sequence to be identified and the standard template skeleton sequence.

Wherein, the to-be-identified skeleton sequence includes a plurality of human skeletons arranged in chronological order; the standard template skeleton sequence includes a plurality of standard human skeletons arranged in chronological order, and the human skeletons in the standard template skeleton sequence are from It is extracted from the video stream of standard motion, and the extraction method is the same as the skeleton sequence to be identified. Specifically, the human skeleton is determined according to the key feature points of the human body, and the key feature points of the human body include at least: the position points of the joints where the limbs are located, the position points of the palms and soles of the feet, the position points of the head, the position points of the torso, etc. Make specific restrictions.

Specifically, in this embodiment, the skeleton sequence to be identified corresponding to the video stream can be determined according to the skeleton generation algorithm, and the skeleton sequence to be identified includes a plurality of human skeletons sorted in time series, that is, the human skeleton is determined according to the video frames in the video stream , the human skeleton in the skeleton sequence to be identified corresponds to the video frame in the video stream. Among them, the skeleton generation algorithm reduces the dimensionality of the video frame sequence in the video stream, and turns the action recognition problem in a video frame into a low-dimensional skeleton action recognition problem. Through dimensionality reduction, the complexity of the algorithm can be reduced and the robustness can be improved.

It should be noted that, in this embodiment, abnormal human skeletons in the skeleton sequence to be identified can also be filtered out according to the temporal continuity of the user's motion. Specifically, various recognition beats in a very short period of time can be filtered out according to the continuity of the action. For example, a filtering algorithm can be used to filter out some obvious outliers, so as to avoid the problem of inaccurate motion scores caused by unstable human skeleton recognition. .

For the embodiment of the present invention, acquiring the user's skeleton sequence to be identified includes:

S101, acquiring a user's motion video stream; extracting video frames from the motion video stream.

Specifically, in this embodiment of the present invention, a motion video stream of the user can be captured by a camera, and then video frames are extracted from the motion video stream, so as to obtain the user's skeleton sequence to be recognized based on the video frames. It should be noted that, in this embodiment, video frames are extracted from the motion video stream according to preset time intervals, so as to reduce the number of video frame processing. frame, this embodiment does not specifically limit the interval time for extracting video frames.

S102, the video frame is input into a human body feature point recognition model to obtain a plurality of human body feature points.

Wherein, the human body feature points include at least joint feature points of human limbs, trunk feature points, and head feature points. In this embodiment, the training process of the human body feature point recognition model may be: acquiring a sample motion video stream, extracting video frames from the sample motion video stream; labeling the human body feature points on the extracted video frames, according to the Perform model training on the extracted video frames and their marked human body feature points to obtain the human body feature point recognition model.

S103: Determine a human skeleton according to the human body feature points corresponding to each video frame.

Specifically, the human skeleton can be determined by connecting lines according to the positional relationship of the feature points of the human body.

S104 , sorting the human skeletons according to the time sequence of the video frames to obtain a skeleton sequence to be identified.

In this embodiment, the human skeletons determined according to the video frames are sorted according to the time sequence of the video frames to obtain the skeleton sequence to be recognized, so as to score the user's motion according to the skeleton sequence to be recognized.

S20, calculate the difference value of the corresponding human skeleton in the skeleton sequence to be identified and the skeleton sequence of the standard template.

The difference value may specifically refer to a difference in relative position, a difference in relative angle, etc., which is not specifically limited in this embodiment.

In an optional embodiment provided by the present invention, the calculation of the difference value between the to-be-identified skeleton sequence and the standard template skeleton sequence corresponding to the human skeleton includes:

S201: Obtain the skeleton sequence to be identified and the corresponding human skeleton in the standard template skeleton sequence.

Specifically, in this embodiment, the matching human skeleton is determined as the corresponding human skeleton by comparing the human skeleton in the skeleton sequence to be identified with the human skeleton in the standard template skeleton sequence, for example, the second person in the skeleton sequence to be identified is determined by matching. If the similarity between the human skeleton and the third human skeleton in the skeleton sequence to be identified is the highest, the second human skeleton and the third human skeleton are determined as corresponding human skeletons; or directly according to the human skeleton in the skeleton sequence Determine the corresponding human skeleton in sequence, that is, determine the skeleton sequence to be identified and the human skeleton in the corresponding sequence in the skeleton sequence to be identified as the corresponding human skeleton, such as the second human skeleton in the skeleton sequence to be identified and the skeleton sequence to be identified. The second human skeleton in is determined to be the corresponding human skeleton, which is not specifically limited in this embodiment.

S202: Determine each corresponding human body feature point in the corresponding human skeleton.

It should be noted that the human skeleton in this embodiment is composed of a plurality of human body feature points, and each human body feature point corresponds to identification information, and the identification information is used to uniquely identify the human body feature points in the human body skeleton. The corresponding identification information can be expressed by numbers, and the corresponding identification information can also be expressed by the names of the position points of the joints where the limbs are located, the position points of the palms and soles of the feet, the position points of the head, and the position points of the torso, such as the identification information of the head position point. It can be "head", the identification information of the position point of the left palm can be "make palm", etc., which is not specifically limited in this embodiment.

S203: Calculate the difference values between all the corresponding human body feature points in the corresponding human skeleton, and sum up all the difference values to obtain the difference value of the corresponding human skeleton.

In this embodiment, after obtaining the corresponding human skeletons in the skeleton sequence to be identified and the standard template skeleton sequence, a matching analysis is performed on the two human skeletons to determine that the two human skeletons belong to the same human body feature point, and then determine respectively The position of the human feature point in the corresponding human skeleton and the angle relative to other human feature points, and then the difference value calculated based on the position of the corresponding human feature point and the angle relative to other feature points, and finally all the difference values are summed up to get The difference value corresponding to the human skeleton.

For example, the human skeleton consists of five human feature points: human feature point 1, human feature point 2, human feature point 3, human feature point 4, and human feature point 5. After obtaining the skeleton sequence to be recognized and the standard template After the corresponding human skeletons in the skeleton sequence, the position of the human body feature point 1 in the two human skeletons, and the angle and position of the human body feature point 1 relative to other human body feature points are obtained respectively, and then according to the corresponding The relative position and angle of the human body feature point 1, and the difference value calculation (that is, calculating the difference between the position and angle of the human body feature point 1) to obtain the difference value of 1 about the human body feature point in the two human skeletons, and finally all the human body feature points. The difference value of (human body feature point 1-human body feature point 5) is summed to obtain the difference value of the corresponding human skeleton.

S30, the difference value is input into the action pattern recognition model, and the score of each human skeleton in the skeleton sequence to be recognized is obtained.

It should be noted that, in this embodiment, before the difference value is input into the action pattern recognition model and the scores of each human skeleton in the skeleton sequence to be recognized are obtained, the model needs to be trained to obtain the motion pattern recognition model. The training method is:

S301, obtaining a sample skeleton sequence and a standard template skeleton sequence.

Wherein, the sample skeleton sequence includes a plurality of human skeletons arranged in time sequence.

S302, calculate the difference value of the corresponding human skeleton in the sample skeleton sequence and the standard template skeleton sequence.

Preferably, before calculating the difference value between the sample skeleton sequence and the corresponding human skeleton in the standard template skeleton sequence, the method further includes: adjusting the limb lengths of each human skeleton in the sample skeleton sequence to the The limb lengths of the human skeleton in the standard template skeleton sequence.

S303: Perform model training according to the motion score labels and difference values corresponding to each human skeleton in the sample skeleton sequence, to obtain the motion pattern recognition model.

Among them, the sports score label is manually marked, and the sports score label can be classified as: error (s<60), qualified (60<=s<70), up to standard (70<=s<80), excellent (80 <=s<90), perfect (90<=s<100) and other five grades, of which s=100 is scored by the standard action itself. Then, based on this set of standards, we can manually mark a large number of samples as training samples for Deep Regression Networks. Pattern Recognition Model.

Correspondingly, before calculating the difference value between the skeleton sequence to be identified and the skeleton sequence of the standard template corresponding to the human skeleton, the method further includes: adjusting the length of the limbs of each human skeleton in the skeleton sequence to be identified correspondingly to a certain value. The limb lengths of the human skeleton in the standard template skeleton sequence described above.

S40, according to the score of each human skeleton in the to-be-identified skeleton sequence, determine the motion score of the user.

Further, after the user's exercise score is obtained, an alarm is issued for the user's exercise score lower than a certain value. Specifically, an alarm may be given by a display screen, an alarm may be given by a voice broadcast, or an alarm may be given by vibrating a wristwatch worn by the user, which is not specifically limited in the embodiment of the present invention.

A method for scoring human motion provided by an embodiment of the present invention first obtains a user's skeleton sequence to be identified and a standard template skeleton sequence, then calculates the difference value of the corresponding human skeleton between the to-be-identified skeleton sequence and the standard template skeleton sequence, and then calculates the difference between the skeleton sequence to be identified and the standard template skeleton sequence The value is input into the action pattern recognition model, and the score of each human skeleton in the skeleton sequence to be recognized is obtained; finally, the user's motion score is determined according to the score of each human skeleton in the skeleton sequence to be recognized. Wherein, the to-be-identified skeleton sequence includes multiple human skeletons arranged in chronological order; the standard template skeleton sequence includes multiple standard human skeletons arranged in chronological order. Compared with the existing human body motion scoring which relies on human subjective judgment, the present invention does not require manual participation in the judgment in the process of human body motion scoring. The user's motion score is obtained, so that the present invention can effectively monitor the user's motion, and effectively improve the accuracy and efficiency of the human body motion score.

As shown in FIG. 2, in an optional embodiment provided by the present invention, according to the score of each human skeleton in the skeleton sequence to be identified, determining the motion score of the user includes:

S401: Determine the action attributes corresponding to each human skeleton in the skeleton sequence to be identified.

There are many movements in sports, and some movements are a combination of a series of key movements, such as: high fives, back kicks, sitting and standing, etc. Some are a hold-for-one action, such as kneeling on the opposite side of the balance, plank support, etc. Therefore, in this embodiment, actions corresponding to the human skeleton are divided into different action attributes, so as to distinguish key actions and non-critical actions in the human skeleton. That is, the action attribute in this embodiment includes a key action and an intermediate action, wherein an action between two adjacent key actions is an intermediate action.

S402, according to the weight values corresponding to the key actions and the intermediate actions respectively, perform a weighted calculation on the scores of all human skeletons in the skeleton sequence to be identified, to obtain a user's motion score.

Specifically, in this embodiment, the user's exercise score is calculated by the following formula;

为关键动作集合，

为关键动作之间的中间动作集合。α为关键动作应的权重值，β为中间动作对应的权重值，α+β＝1。in,

is a collection of key actions,

A collection of intermediate actions between key actions. α is the weight value corresponding to the key action, β is the weight value corresponding to the intermediate action, α+β=1.

It should be understood that the size of the sequence numbers of the steps in the above embodiments does not mean the sequence of execution, and the execution sequence of each process should be determined by its functions and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

In one embodiment, a human body motion scoring device is provided, and the human body motion scoring device is in one-to-one correspondence with the human body motion scoring method in the above embodiment. As shown in FIG. 4 , the human body motion scoring apparatus includes: an acquisition module 10 , a calculation module 20 , and a determination module 30 . The detailed description of each functional module is as follows:

The acquisition module 10 is used to acquire the user's skeleton sequence to be identified and a standard template skeleton sequence, the to-be-identified skeleton sequence includes a plurality of human skeletons arranged in chronological order; the standard template skeleton sequence includes a plurality of human skeletons arranged in chronological order Arranged standard human skeleton;

The calculation module 20 is used to calculate the difference value of the corresponding human skeleton in the skeleton sequence to be identified and the standard template skeleton sequence;

The computing module 20 is further configured to input the difference value into the action pattern recognition model to obtain the score of each human skeleton in the skeleton sequence to be identified;

The determining module 30 is configured to determine the motion score of the user according to the score of each human skeleton in the skeleton sequence to be identified.

Further, the apparatus further includes: a training module 40;

The acquisition module 10 is further configured to acquire a sample skeleton sequence and a standard template skeleton sequence; the sample skeleton sequence includes a plurality of human skeletons arranged in chronological order;

The calculation module 20 is further configured to calculate the difference value of the corresponding human skeleton in the sample skeleton sequence and the standard template skeleton sequence;

The training module 40 is configured to perform model training according to the motion score labels and difference values corresponding to each human skeleton in the sample skeleton sequence to obtain the motion pattern recognition model.

Further, the device further includes: an adjustment module 50;

The adjustment module 50 is used for correspondingly adjusting the limb lengths of each human skeleton in the to-be-identified skeleton sequence to the limb lengths of the human skeleton in the standard template skeleton sequence;

The adjustment module 50 is configured to adjust the limb lengths of each human skeleton in the sample skeleton sequence to correspond to the limb lengths of the human skeleton in the standard template skeleton sequence.

Further, the determination module 30 is configured to determine the action attributes corresponding to each human skeleton in the to-be-identified skeleton sequence, where the action attributes include key actions and intermediate actions;

The calculation module 20 is further configured to perform weighted calculation on the scores of all human skeletons in the to-be-identified skeleton sequence according to the weight values corresponding to the key actions and the intermediate actions respectively, to obtain the motion score of the user.

Further, the acquisition module 10 is specifically used for:

Obtain the user's motion video stream; extract video frames from the motion video stream;

Inputting the video frame into the human body feature point recognition model to obtain a plurality of human body feature points;

Determine the human skeleton according to the human body feature points corresponding to each video frame;

According to the time sequence of the video frames, the human skeleton is sorted to obtain the to-be-identified skeleton sequence.

Further, the calculation module 20 is specifically used for:

Obtain the corresponding human skeleton in the skeleton sequence to be identified and the standard template skeleton sequence;

Determine each corresponding human body feature point in the corresponding human skeleton;

Calculate the difference value between all the corresponding human body feature points in the corresponding human skeleton, and sum up all the difference values to obtain the difference value of the corresponding human skeleton.

Further, the obtaining module 10 is further configured to obtain a sample motion video stream, and extract video frames from the sample motion video stream; label the extracted video frames with human body feature points, where the human body feature points at least include human limbs joint feature points, torso feature points, and head feature points;

The training module 40 is configured to perform model training according to the extracted video frames and the marked human body feature points to obtain the human body feature point recognition model.

For the specific limitation of the human motion scoring apparatus, please refer to the definition of the human motion scoring method above, which will not be repeated here. Each module in the above-mentioned human motion scoring apparatus may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

In one embodiment, a computer device is provided, and the computer device may be a server, and its internal structure diagram may be as shown in FIG. 5 . The computer device includes a processor, memory, a network interface, and a database connected by a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium, an internal memory. The nonvolatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the execution of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer program, when executed by the processor, implements a method for scoring human motion.

In one embodiment, a computer device is provided, comprising a memory, a processor, and a computer program stored on the memory and running on the processor, and the processor implements the following steps when executing the computer program:

Obtain a user's skeleton sequence to be identified and a standard template skeleton sequence, the to-be-identified skeleton sequence includes a plurality of human skeletons arranged in chronological order; the standard template skeleton sequence includes a plurality of standard human skeletons arranged in chronological order ;

Calculate the difference value of the corresponding human skeleton in the to-be-identified skeleton sequence and the standard template skeleton sequence;

The difference value is input into the action pattern recognition model, and the score of each human skeleton in the skeleton sequence to be recognized is obtained;

According to the score of each human skeleton in the to-be-identified skeleton sequence, the motion score of the user is determined.

In one embodiment, a computer-readable storage medium is provided on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

Obtain a user's skeleton sequence to be identified and a standard template skeleton sequence, the to-be-identified skeleton sequence includes a plurality of human skeletons arranged in chronological order; the standard template skeleton sequence includes a plurality of standard human skeletons arranged in chronological order ;

Calculate the difference value of the corresponding human skeleton in the to-be-identified skeleton sequence and the standard template skeleton sequence;

The difference value is input into the action pattern recognition model, and the score of each human skeleton in the skeleton sequence to be recognized is obtained;

According to the score of each human skeleton in the to-be-identified skeleton sequence, the motion score of the user is determined.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage In the medium, when the computer program is executed, it may include the processes of the above-mentioned method embodiments. Wherein, any reference to memory, storage, database or other medium used in the various embodiments provided in this application may include non-volatile and/or volatile memory. Nonvolatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Road (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Those skilled in the art can clearly understand that, for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned functions can be allocated to different functional units, Module completion, that is, dividing the internal structure of the device into different functional units or modules to complete all or part of the functions described above.

The above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The recorded technical solutions are modified, or some technical features thereof are equivalently replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the present invention. within the scope of protection.

Claims (10)

1. a scoring method of human body movement, is characterized in that, described method comprises: Obtain a user's skeleton sequence to be identified and a standard template skeleton sequence, the to-be-identified skeleton sequence includes a plurality of human skeletons arranged in chronological order; the standard template skeleton sequence includes a plurality of standard human skeletons arranged in chronological order ; Calculate the difference value of the corresponding human skeleton in the to-be-identified skeleton sequence and the standard template skeleton sequence; The difference value is input into the action pattern recognition model, and the score of each human skeleton in the skeleton sequence to be recognized is obtained; According to the score of each human skeleton in the to-be-identified skeleton sequence, the motion score of the user is determined. 2. The method according to claim 1, wherein the method further comprises: Obtaining a sample skeleton sequence and a standard template skeleton sequence; the sample skeleton sequence includes a plurality of human skeletons arranged in chronological order; Calculate the difference value of the corresponding human skeleton in the sample skeleton sequence and the standard template skeleton sequence; Model training is performed according to the motion score labels and difference values corresponding to each human skeleton in the sample skeleton sequence to obtain the motion pattern recognition model. 3. The method according to claim 2, wherein before calculating the difference value of the corresponding human skeleton in the skeleton sequence to be identified and the standard template skeleton sequence, the method further comprises: The limb lengths of each human skeleton in the to-be-identified skeleton sequence are correspondingly adjusted to the limb lengths of the human skeleton in the standard template skeleton sequence; Before calculating the difference value of the corresponding human skeleton in the sample skeleton sequence and the standard template skeleton sequence, the method further includes: The limb lengths of each human skeleton in the sample skeleton sequence are correspondingly adjusted to the limb lengths of the human skeleton in the standard template skeleton sequence. 4. The method according to any one of claims 1-3, wherein after determining the user's motion score according to the score of each human skeleton in the to-be-identified skeleton sequence, the method further comprises: include: Determine the action attributes corresponding to each human skeleton in the to-be-identified skeleton sequence, where the action attributes include key actions and intermediate actions; According to the weight values corresponding to the key actions and the intermediate actions respectively, the scores of all human skeletons in the to-be-identified skeleton sequence are weighted and calculated to obtain the user's motion score. 5. The method according to claim 1, wherein the acquiring the user's skeleton sequence to be identified comprises: Obtain the user's motion video stream; extract video frames from the motion video stream; Inputting the video frame into the human body feature point recognition model to obtain a plurality of human body feature points; Determine the human skeleton according to the human body feature points corresponding to each video frame; According to the time sequence of the video frames, the human skeletons are sorted to obtain the to-be-identified skeleton sequence. 6. The scoring method of human motion according to claim 5, wherein the calculating the difference value of the corresponding human skeleton in the skeleton sequence to be identified and the standard template skeleton sequence, comprising: Obtain the corresponding human skeleton in the skeleton sequence to be identified and the standard template skeleton sequence; Determine each corresponding human body feature point in the corresponding human skeleton; Calculate the difference value between all the corresponding human body feature points in the corresponding human skeleton, and sum up all the difference values to obtain the difference value of the corresponding human skeleton. 7. The method according to claim 5, wherein the method further comprises: Obtain a sample motion video stream, and extract video frames from the sample motion video stream; Labeling the extracted video frames with human body feature points, where the human body feature points at least include joint feature points of human limbs, torso feature points, and head feature points; Model training is performed on the extracted video frames and the marked human body feature points to obtain the human body feature point recognition model. 8. A scoring device for human movement, wherein the device comprises: The acquisition module is used to acquire the user's skeleton sequence to be identified and a standard template skeleton sequence, the to-be-identified skeleton sequence includes a plurality of human skeletons arranged in chronological order; the standard template skeleton sequence includes a plurality of human skeletons arranged in chronological order the standard human skeleton; A calculation module, used for calculating the difference value of the corresponding human skeleton in the skeleton sequence to be identified and the standard template skeleton sequence; The computing module is further configured to input the difference value into the action pattern recognition model to obtain the score of each human skeleton in the skeleton sequence to be identified; A determination module, configured to determine the motion score of the user according to the score of each human skeleton in the to-be-identified skeleton sequence. 9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the computer program as claimed in the claims The scoring method of any one of 1 to 7 for human motion. 10. A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program is executed by a processor to realize the movement of the human body according to any one of claims 1 to 7. scoring method.

Images