A dance posture correction method and system based on artificial intelligence

Abstract

The present application relates to the technical field of dance posture correction, and provides a dance posture correction method and system based on artificial intelligence, which comprises the following steps: processing audio data of a dance video to be corrected, segmenting the dance video according to beat time points of music segment division, and identifying joint point coordinate strokes of a human body before and after the beat time points by using an artificial intelligence model, wherein the coordinate strokes reflect dance action conditions; determining whether dance action is stuck at the beat time points by identifying whether the joint point strokes of the human body in two video segments are continuous strokes, judging whether dance posture at the corresponding beat time points needs to be corrected, and quantifying the corresponding relationship between dance action and music rhythm, so that a user can complete the correction based on his / her own dance video, and thus more intelligent and more accurate dance teaching and evaluation are realized.

Description

Technical Field

[0001] This invention relates to the field of dance posture correction technology, and in particular to a dance posture correction method and system based on artificial intelligence. Background Technology

[0002] The rhythmic characteristics of dance music are its soul. It guides and controls the dance through time signature, tempo, rhythmic pattern, percussion timbre and paragraph changes. In the process of learning dance, dance teachers need to correct and adjust the dancer's dance posture. Generally, it is necessary to adjust the dancer's movements from head to toe section by section, which is very labor-intensive. Moreover, human eye recognition is also prone to errors and may not be able to accurately point out the problems in the dance movements.

[0003] In existing technologies, such as the published patent CN119600691A, a method and system for dance posture and movement recognition and interaction based on visual computing, it is necessary to use a standard dance posture for comparison after obtaining the dancer's dance video. Although the comparison results are accurate and standard, it requires high computing power and is inefficient. Summary of the Invention

[0004] This invention provides an artificial intelligence-based method for correcting dance postures, which addresses the problem of low efficiency in existing dance posture correction techniques.

[0005] The first aspect of this invention provides a dance posture correction method based on artificial intelligence, comprising:

[0006] Acquire the dance video to be corrected and extract the audio data from the video; identify the beat time points of music segmentation based on the spectral characteristics of the audio data;

[0007] The dance video to be corrected is segmented based on the beat time point. The video between the current beat time point and the previous beat time point is taken as the first video segment, and the video between the current beat time point and the next beat time point is taken as the second video segment. The coordinates and travel of each human joint point in the first video segment and the second video segment are identified by an artificial intelligence model.

[0008] Determine whether the movement of human joints in the first video segment is continuous with that in the second video segment. If so, determine whether the dance posture at the corresponding beat time point needs to be corrected.

[0009] Optionally, after determining whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels, the method further includes:

[0010] The spectral characteristics of the audio data of the first video segment and the second video segment are compared to obtain the spectral ratio relationship. The velocity characteristics of the human joints in the first video segment and the second video segment are obtained respectively, and the velocity relationship is compared to obtain the velocity ratio relationship. It is determined whether the velocity ratio relationship conforms to the spectral ratio relationship. If not, it is determined that the dance posture at the corresponding video segment needs to be corrected.

[0011] Optionally, before identifying the beat time points of music segment segmentation based on the spectral characteristics of the audio data, the method further includes:

[0012] The beat segmentation model is pre-trained based on a pre-set music segmentation sample set to learn the beat timing features of music segmentation; the audio data of the dance video to be corrected is input into the beat segmentation model to obtain the corresponding beat timing correction value, and the beat timing is corrected.

[0013] Optionally, determining whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels specifically involves:

[0014] Based on the human joint travel in the first and second video segments, establish the relationship between the joint velocity and time. Identify the acceleration direction at the end of the first video segment with a preset time length and the acceleration direction at the beginning of the second video segment with a preset time length, and determine whether the angle between the acceleration directions is within a preset angle.

[0015] The second aspect of this application provides an artificial intelligence-based dance posture correction system, comprising:

[0016] The beat segmentation module is used to acquire the dance video to be corrected and separate the audio data from the video; it identifies the beat time points of music segmentation based on the spectral characteristics of the audio data.

[0017] The video processing module is used to segment the dance video to be corrected based on the beat time point. The video between the current beat time point and the previous beat time point is taken as the first video segment, and the video between the current beat time point and the next beat time point is taken as the second video segment. The artificial intelligence model is used to identify the coordinate travel of each human joint point in the first video segment and the second video segment respectively.

[0018] The dance posture correction module is used to determine whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous. If so, it determines that the dance posture at the corresponding beat time point needs to be corrected.

[0019] Optionally, after determining whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels in the dance posture correction module, it further includes:

[0020] The spectral characteristics of the audio data of the first video segment and the second video segment are compared to obtain the spectral ratio relationship. The velocity characteristics of the human joints in the first video segment and the second video segment are obtained respectively, and the velocity relationship is compared to obtain the velocity ratio relationship. It is determined whether the velocity ratio relationship conforms to the spectral ratio relationship. If not, it is determined that the dance posture at the corresponding video segment needs to be corrected.

[0021] Optionally, before identifying the beat time points for music segmentation based on the spectral characteristics of the audio data, the beat segmentation module further includes:

[0022] The beat segmentation model is pre-trained based on a pre-set music segmentation sample set to learn the beat timing features of music segmentation; the audio data of the dance video to be corrected is input into the beat segmentation model to obtain the corresponding beat timing correction value, and the beat timing is corrected.

[0023] Optionally, in the dance posture correction module, determining whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels specifically involves:

[0024] Based on the human joint travel in the first and second video segments, establish the relationship between the joint velocity and time. Identify the acceleration direction at the end of the first video segment with a preset time length and the acceleration direction at the beginning of the second video segment with a preset time length, and determine whether the angle between the acceleration directions is within a preset angle.

[0025] A third aspect of this application provides a device for a dance posture correction method based on artificial intelligence, the device comprising a processor and a memory:

[0026] The memory is used to store program code and transmit the program code to the processor;

[0027] The processor is used to execute, according to the instructions in the program code, an artificial intelligence-based dance posture correction method as described in any one of the first aspects of the present invention.

[0028] The fourth aspect of this application provides a computer-readable storage medium for storing program code for executing an artificial intelligence-based dance posture correction method as described in any of the first aspects of the present invention.

[0029] As can be seen from the above technical solution, the present invention has the following advantages: By processing the audio data of the dance video to be corrected, the dance video is segmented using the beat time points of the music segmentation. An artificial intelligence model is used to identify the coordinate travel of human joint points before and after the beat time points, and the coordinate travel reflects the dance movement. By identifying whether the human joint point travel in two video segments is continuous, it is determined whether the dance movement is stuck at the beat time point, and whether the dance posture at the corresponding beat time point needs to be corrected. The correspondence between dance movements and music rhythm is quantified, so that users can complete the correction based only on their own dance video, thereby achieving more intelligent and accurate dance teaching and assessment. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 This is a flowchart of an artificial intelligence-based dance posture correction method;

[0032] Figure 2 This is a structural diagram of an artificial intelligence-based dance posture correction system. Detailed Implementation

[0033] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0034] This invention provides an artificial intelligence-based method for correcting dance postures, which addresses the problem of low efficiency in existing dance posture correction techniques.

[0035] Please see Figure 1 , Figure 1 The first flowchart of an artificial intelligence-based dance posture correction method provided in this embodiment of the invention.

[0036] S100: Acquire the dance video to be corrected and separate the audio data from the video; identify the beat time points of the music segment segment based on the spectral characteristics of the audio data;

[0037] It should be noted that the dance video to be corrected is generally a video of the user's actual dance process. In addition to recording the dance moves, the video also contains background dance music. Therefore, audio data can be separated from the dance video using audio separation technology.

[0038] In audio data, the peak or trough of amplitude can be identified based on amplitude changes in the spectral characteristics. Since the force points or posture changes of dance movements precisely align with the strong or weak beats of the music, it's necessary to identify these beat points. Alternatively, rhythmic features can be extracted first, such as BPM, rhythmic pattern, and percussion energy. By analyzing rhythm and energy changes, different sections of the music can be automatically divided. Strong amplitude time points matching the rhythmic characteristics can be identified, and the duration between these strong amplitude time points should conform to the rhythmic characteristics. Then, sections can be segmented. Generally, the highest amplitude point corresponds to the beginning and end of a beat segment. These highest amplitude points are then used as nodes for segmenting musical sections, yielding the corresponding beat points. In dance music with multiple strong amplitude sections, the instrumental attributes of the strong amplitude points can be determined based on the timbre characteristics of the spectral characteristics, thus identifying whether they are beat points for segmenting musical sections.

[0039] S200 segments the dance video to be corrected based on beat time points, taking the video between the current beat time point and the previous beat time point as the first video segment, and the video between the current beat time point and the next beat time point as the second video segment; and uses an artificial intelligence model to identify the coordinate travel of each human joint point in the first video segment and the second video segment respectively.

[0040] It should be noted that the dance video to be corrected is segmented according to the beat time points, and the resulting video segments correspond to the human dance movements in the corresponding music segment. When processing a certain beat time point, the earlier and later beat time points adjacent to it can be identified. The video between the current beat time point and the previous beat time point is taken as the first video segment, and the video between the current beat time point and the next beat time point is taken as the second video segment. These two video segments respectively reflect the dance movements before the current beat time point after the music segment segmentation, and the dance movements after the current beat time point after the music segment segmentation.

[0041] The dance video to be corrected is captured by a camera or depth sensor. Therefore, the three-dimensional coordinates of human joints can be identified in the video using an artificial intelligence model. In this embodiment, the artificial intelligence model can use a lightweight convolutional neural network or computer vision to identify human joints. The main human joints to be identified in the dance movements include elbow joints, knee joints, shoulder joints, and wrist joints. The coordinate changes of each human joint are identified in the first and second video segments respectively. The travel distance can be obtained by connecting the continuous coordinates. The human joint coordinate travel distance reflects the changes of the human body's movements over time in the dance. In addition to the changes of specific joint coordinates over time, the movement speed and acceleration of the joints can also be obtained based on the speed of coordinate changes.

[0042] S300, determine whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels. If so, determine that the dance posture at the corresponding beat time point needs to be corrected.

[0043] It's important to note that dance postures require movements to match the rhythm. Therefore, it's crucial to judge whether the movements are on time. Generally, dance movements such as waving arms, swinging arms, and kicking legs should be performed between two beat points, while the beat points should be the timing of movement transitions. For example, in the first video segment, the corresponding arm swing should occur when the arm is raised, and stop swinging at the beat point. In the second video segment, the corresponding arm swing should occur when the arm is lowered. This ensures that the movement of the corresponding joint points of the arm is in rhythm. However, if the arm does not stop swinging at the beat point, the beat point may fall within the arm's upward or downward movement, resulting in the same segment of movement occurring in different parts of the movement. In the first and second video segments, the dance movements are generally smooth and fluid. Therefore, if the end of the path corresponding to the human joint point in the first video segment can connect with the beginning of the path corresponding to the human joint point in the second video segment, the continuity of the movement can be judged by calculating whether the arc of the two paths is continuous, or by judging whether the two paths are on the same straight line when the joint point moves in a straight line. When a smooth and continuous movement is identified between the first and second video segments, it reflects whether the segmented beat time point is in the process of the dance movement. In this case, the dance posture at that time point needs to be corrected, and the dancer needs to more accurately control the timing of the movement at that beat time point so that the dance movement is more accurately timed.

[0044] In this embodiment, the audio data of the dance video to be corrected is processed, and the dance video is segmented using the beat time points of the music segments. An artificial intelligence model is used to identify the coordinate travel of human joints before and after the beat time points. The coordinate travel reflects the dance movements. By identifying whether the human joint travel in two video segments is continuous, it is determined whether the dance movements are out of sync at the beat time points. It is also determined whether the dance posture at the corresponding beat time points needs to be corrected. The correspondence between dance movements and music rhythm is quantified, so that users can complete the correction based only on their own dance videos, thereby achieving more intelligent and accurate dance teaching and assessment.

[0045] The above is a detailed description of the first embodiment of an artificial intelligence-based dance posture correction method provided in this application. The following is a detailed description of the second embodiment of an artificial intelligence-based dance posture correction method provided in this application.

[0046] In this embodiment, an artificial intelligence-based dance posture correction method is further provided. Please refer to [link to relevant documentation]. Figure 2 In the aforementioned step S300, after determining whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels, the method further includes:

[0047] The spectral characteristics of the audio data of the first video segment and the second video segment are compared to obtain the spectral ratio relationship. The velocity characteristics of the human joints in the first video segment and the second video segment are obtained respectively, and the velocity relationship is compared to obtain the velocity ratio relationship. It is determined whether the velocity ratio relationship conforms to the spectral ratio relationship. If not, it is determined that the dance posture at the corresponding video segment needs to be corrected.

[0048] It should be noted that, in addition to the basic requirement of keeping the beat as described above, dance also needs to express the emotions, story, and atmosphere of the music. That is, the amplitude and speed of the dance movements need to match the music. For example, the tempo of the music varies. Fast music refers to a fast speed and dense rhythm, which is represented by the rapid and dense repetition of waveform peaks in the time domain of the spectrum. Correspondingly, the dance needs to make fast and agile movements, that is, the speed of the dance movements is fast. On the other hand, slow music refers to a slow speed and sparse rhythm, which is represented by the large intervals between waveform peaks in the time domain of the spectrum. Correspondingly, the dance needs to make continuous and slow movements, which are full of resistance and control.

[0049] The number of peaks is extracted from the spectral features of the first and second video segments, and the peak counts are compared to obtain the peak count ratio, i.e., the spectral ratio. This ratio reflects the tempo ratio of the dance music in the audio of the first and second video segments. Furthermore, the average movement speed of the joints in the human body's movements in the first and second video segments can be calculated and compared to obtain the speed ratio. This ratio reflects the speed ratio of the dance movements. In this embodiment, it is not necessary to target a specific sound... Instead of correcting dance postures based on the specific rhythm of a musical segment, the method involves judging whether the speed ratio of dance postures in adjacent video segments matches the tempo ratio of the music. For example, if two adjacent musical segments start fast and then slow, the dance movements should also start fast and then slow. The speed ratio can be adjusted to match the degree of difference in tempo. To judge whether the speed ratio matches the spectrum ratio, the method can be based on whether the ratio values ​​are consistent or within a certain threshold range. If it exceeds the threshold range, the speed of the dance movements does not match the tempo changes of the music, which will directly affect the visual appeal of the dance. In this case, the dance postures in that video segment need to be corrected, and the specific correction can be adjusted according to the spectrum ratio of the music.

[0050] Furthermore, the determination of whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels specifically involves:

[0051] Based on the human joint travel in the first video segment and the second video segment, establish the relationship between the joint velocity and time, identify the acceleration direction at the end of the first video segment with a preset time length, and the acceleration direction at the beginning of the second video segment with a preset time length, and determine whether the angle between the acceleration directions is within the preset angle.

[0052] It should be noted that dance movements that are time-bound to the music beats are not necessarily static or involve a change in direction, but the audience must be able to perceive the change in movement at specific points. Therefore, the change in joint speed over time can be obtained by observing the movement of the human joints, and then the specific acceleration can be extracted from this time-varying change. The end of the first video segment and the beginning of the second video segment correspond to the vicinity of the music beats, requiring a change in movement. Therefore, the corresponding acceleration directions should be opposite to reflect the change in movement, such as deceleration followed by acceleration, or acceleration followed by deceleration. In this embodiment, the acceleration direction corresponds to the direction of acceleration in space. The direction of the human joint movement will change accordingly. When the acceleration... If the directional angle exceeds 90°, there will be a deceleration effect in the original direction, and the change in movement will be obvious. Therefore, the acceleration directional angle can be used to determine whether the human joint strokes in the dance movements in the first and second video segments are continuous. In this embodiment, the preset angle can be set to 90° and adjusted according to the specific dance music type. The preset time length can be set according to the processing speed of human eyes to observe the speed of movement. The time for human eyes to observe the speed of movement is mainly affected by the speed of eye scanning and the processing speed of the brain. It can generally be set to about 0.1s, that is, the change of the dance movement points within this time range will not affect the dance viewing experience, and the audience will still have the feeling of the points.

[0053] Furthermore, before identifying the beat timing points of music segment segments based on the spectral characteristics of the audio data, the method further includes: pre-training the beat segmentation model based on a preset music segmentation sample set to learn the beat timing point features of music segment segments; inputting the audio data of the dance video to be corrected into the beat segmentation model to obtain the corresponding beat timing point correction value, and correcting the beat timing points.

[0054] It should be noted that in certain music styles of dance, the dance movements are not precisely timed, but rather deliberately deviate from the expected beat, performed on weak or off-beats to create rhythmic tension. These dance styles are generally related to music genres, and the beat segmentation model can be pre-trained using a pre-defined music segmentation sample set. The music segmentation sample set contains audio data and corresponding manually divided music segment time points. The model learns the beat time point features of music segment segmentation, and can accurately identify whether the music beat time points need to be deviated and corrected after inputting the audio data of the dance video to be corrected. The beat segmentation model can be trained using a temporal convolutional network, a backpropagation neural network, or a Transformer model to learn the action-beat mapping relationship.

[0055] The above is a detailed description of an artificial intelligence-based dance posture correction method provided by the first aspect of this application. The following is a detailed description of an embodiment of an artificial intelligence-based dance posture correction system provided by the second aspect of this application.

[0056] Please see Figure 2 , Figure 2 This is a structural diagram of an artificial intelligence-based dance posture correction system. This embodiment provides an artificial intelligence-based dance posture correction system, including:

[0057] The beat segmentation module 10 is used to acquire the dance video to be corrected and separate the audio data from the video; and to identify the beat time points of the music segmentation based on the spectral characteristics of the audio data.

[0058] The video processing module 20 is used to segment the dance video to be corrected based on the beat time point, taking the video between the current beat time point and the previous beat time point as the first video segment, and the video between the current beat time point and the next beat time point as the second video segment; and using an artificial intelligence model to identify the coordinate travel of each human joint point in the first video segment and the second video segment respectively.

[0059] The dance posture correction module 30 is used to determine whether the human joint point travel in the first video segment and the human joint point travel in the second video segment are continuous. If so, it determines that the dance posture at the corresponding beat time point needs to be corrected.

[0060] Furthermore, in the dance posture correction module 30, after determining whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels, it also includes:

[0061] The spectral characteristics of the audio data of the first video segment and the second video segment are compared to obtain the spectral ratio relationship. The velocity characteristics of the human joints in the first video segment and the second video segment are obtained respectively, and the velocity relationship is compared to obtain the velocity ratio relationship. It is determined whether the velocity ratio relationship conforms to the spectral ratio relationship. If not, it is determined that the dance posture at the corresponding video segment needs to be corrected.

[0062] Furthermore, before identifying the beat time points for music segmentation based on the spectral characteristics of the audio data, the beat segmentation module 10 also includes:

[0063] The beat segmentation model is pre-trained based on a pre-set music segmentation sample set to learn the beat timing features of music segmentation; the audio data of the dance video to be corrected is input into the beat segmentation model to obtain the corresponding beat timing correction value, and the beat timing is corrected.

[0064] Furthermore, in the dance posture correction module 30, determining whether the human joint point travel in the first video segment and the human joint point travel in the second video segment are continuous travels specifically involves:

[0065] Based on the human joint travel in the first and second video segments, establish the relationship between the joint velocity and time. Identify the acceleration direction at the end of the first video segment with a preset time length and the acceleration direction at the beginning of the second video segment with a preset time length, and determine whether the angle between the acceleration directions is within a preset angle.

[0066] A third aspect of this application also provides an artificial intelligence-based dance posture correction method device, including a processor and a memory: wherein the memory is used to store program code and transmit the program code to the processor; the processor is used to execute the aforementioned artificial intelligence-based dance posture correction method according to the instructions in the program code.

[0067] A fourth aspect of this application provides a computer-readable storage medium, characterized in that the computer-readable storage medium is used to store program code for executing the above-described artificial intelligence-based dance posture correction method.

[0068] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described apparatus and equipment can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0069] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0070] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0071] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0072] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0073] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dance posture correction method based on artificial intelligence, characterized in that... include: Obtain the dance video to be corrected and extract the audio data from the video; Identify the beat timing points of music segmentation based on the spectral characteristics of audio data; The dance video to be corrected is segmented based on the beat time point. The video between the current beat time point and the previous beat time point is taken as the first video segment, and the video between the current beat time point and the next beat time point is taken as the second video segment. Artificial intelligence models were used to identify the coordinates and travel distances of each human joint in the first and second video segments, respectively. To determine whether the movement of human joints in the first video segment and the movement of human joints in the second video segment are continuous, the following steps are taken: establish the relationship between the speed of joints and time based on the movement of human joints in the first and second video segments respectively; identify the acceleration direction at the end of the first video segment with a preset time length and the acceleration direction at the beginning of the second video segment with a preset time length; determine whether the angle between the acceleration directions is within a preset angle; if so, determine that the dance posture at the corresponding beat time point needs to be corrected.

2. The dance posture correction method based on artificial intelligence according to claim 1, characterized in that, After determining whether the human joint travel in the first video segment and the human joint travel in the second video segment are continuous travels, the method further includes: The spectral characteristics of the audio data of the first video segment and the second video segment are compared to obtain the spectral ratio relationship. The velocity characteristics of the human joints in the first video segment and the second video segment are obtained respectively, and the velocity relationship is compared to obtain the velocity ratio relationship. It is determined whether the velocity ratio relationship conforms to the spectral ratio relationship. If not, it is determined that the dance posture at the corresponding video segment needs to be corrected.

3. The dance posture correction method based on artificial intelligence according to claim 1, characterized in that, Before identifying the beat timing points of music segment segments based on the spectral characteristics of the audio data, the method further includes: The beat segmentation model is pre-trained based on a pre-set music segmentation sample set to learn the beat timing features of music segmentation; the audio data of the dance video to be corrected is input into the beat segmentation model to obtain the corresponding beat timing correction value, and the beat timing is corrected.

4. A dance posture correction system based on artificial intelligence, characterized in that, include: The beat segmentation module is used to acquire the dance video to be corrected and extract the audio data from the video; Identify the beat timing points of music segmentation based on the spectral characteristics of audio data; The video processing module is used to segment the dance video to be corrected based on the beat time point, taking the video between the current beat time point and the previous beat time point as the first video segment, and the video between the current beat time point and the next beat time point as the second video segment. Artificial intelligence models were used to identify the coordinates and travel distances of each human joint in the first and second video segments, respectively. The dance posture correction module is used to determine whether the movement of human joints in the first video segment and the movement of human joints in the second video segment are continuous. Specifically, it establishes the relationship between the speed of joints and time based on the movement of human joints in the first and second video segments, identifies the acceleration direction at the end of the first video segment with a preset time length, and the acceleration direction at the beginning of the second video segment with a preset time length, and determines whether the angle between the acceleration directions is within a preset angle. If so, it determines that the dance posture at the corresponding beat time point needs to be corrected.

5. The dance posture correction system based on artificial intelligence according to claim 4, characterized in that, In the dance posture correction module, after determining whether the human joint point travel in the first video segment and the human joint point travel in the second video segment are continuous travels, it also includes: The spectral characteristics of the audio data of the first video segment and the second video segment are compared to obtain the spectral ratio relationship. The velocity characteristics of the human joints in the first video segment and the second video segment are obtained respectively, and the velocity relationship is compared to obtain the velocity ratio relationship. It is determined whether the velocity ratio relationship conforms to the spectral ratio relationship. If not, it is determined that the dance posture at the corresponding video segment needs to be corrected.

6. The dance posture correction system based on artificial intelligence according to claim 4, characterized in that, Before identifying the beat timing points for music segmentation based on the spectral characteristics of the audio data in the beat segmentation module, the module further includes: The beat segmentation model is pre-trained based on a pre-set music segmentation sample set to learn the beat timing features of music segmentation; the audio data of the dance video to be corrected is input into the beat segmentation model to obtain the corresponding beat timing correction value, and the beat timing is corrected.

7. A dance posture correction device based on artificial intelligence, characterized in that, The device includes a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute, according to the instructions in the program code, the dance posture correction method based on any one of claims 1-3.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program code for executing the dance posture correction method based on artificial intelligence as described in any one of claims 1-3.

Images