Dance action learning progress self-adaptive adjustment method and system

Abstract

The present application relates to the technical field of learning progress feedback, in particular to a dance action learning progress adaptive adjustment method and system, comprising the following steps: analyzing joint three-dimensional posture based on a posture capture device, obtaining joint rotation sequence structure, calculating continuous frame rotation angle change, identifying rhythm transition positioning point, extracting analysis section sequence structure center label, judging center label distance from transition point, adjusting analysis section and correcting node. The present application compares the direction and amplitude of rotation change between continuous frames to determine the jump position within the action, extracts the structure center label in the local frame section based on the relative relationship of angle change, and then obtains the offset parameter based on the interval relationship between the center label and the jump position, and adjusts the analysis section again when the offset exceeds the range, so that the node recognition is established on the basis of continuous rotation structure, the jump position has dynamic reference basis, and the problem that the static node mode cannot adapt to continuous action change is solved.

Description

Technical Field

[0001] This invention relates to the field of learning progress feedback technology, and in particular to a method and system for adaptive adjustment of learning progress in dance movements. Background Technology

[0002] The field of learning progress feedback involves collecting, analyzing, and determining the learning stage of learners' motor performance or behavioral processes in educational and training scenarios. This includes generating progress data based on learners' performance in specific skill training, determining the learning stage based on this data, and using this as a basis for adjusting training content. Traditional adaptive adjustment methods for dance movement learning progress refer to methods that determine the learning stage based on learners' physical performance in dance movement training and select training content accordingly. This typically involves recording dance movement demonstration sequences, using key movement nodes as the basis for judgment, and then using beat matching or comparison of geometric relationships in body postures to determine the learner's completion level. A preset difficulty order or a fixed demonstration movement sequence table is then used as the basis for adjustment.

[0003] Existing technologies rely on demonstration segments and fixed node markers in the action node determination process, and divide stages through beat alignment or static posture comparison. The angle transitions and rotation patterns in the continuous change process of the action are not used as the basis for nodes, resulting in nodes lacking dynamic feature support and making it difficult to reflect the real change structure of the action in a local time period. Node calibration relies on fixed time sequence or external templates, which are prone to offset when faced with individual action differences. Jump positions lack reliable references in the case of fluctuations. Progress judgment is limited by static comparison methods, node fluctuations are difficult to correct in real time, the correspondence between action structure and learning stage is easily blurred, training content has delays and deviations when matching the learning rhythm, and the overall feedback loop is not sensitive enough to changes in action structure. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and propose a method and system for adaptive adjustment of dance movement learning progress.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for adaptive adjustment of dance movement learning progress, comprising the following steps:

[0006] S1: Based on the posture capture camera equipment around the training area, analyze the three-dimensional posture information of the joints in each frame, identify the rotation state of the joints in a single frame, compare the changes in joint positions in time sequence, and associate the angle changes with the frame number to obtain the joint rotation sequence structure.

[0007] S2: Based on the joint rotation sequence structure, calculate the rotation angle change of consecutive frames, compare the change rate of adjacent frames, analyze the angle change direction and amplitude fluctuation, determine whether there is rhythm jump feature, identify abnormal fluctuation frames, and obtain the rhythm jump positioning point.

[0008] S3: Based on the rhythm transition positioning point, sort the rotation angles of each frame in the analysis segment by number, compare the angle changes frame by frame, identify frames with prominent changes or frames with gentle changes, and obtain the center label of the sequence structure.

[0009] S4: Determine the distance between the center label of the sequence structure and the frame number of the rhythm transition positioning point, compare the distance with the node fluctuation range, and if it exceeds the range, adjust the length of the analysis segment, reduce the frame number range, and correct the center number to obtain the node offset parameter.

[0010] S5: Based on the node offset parameter, input the calibration frame number, and update the node identifier in combination with the trainee's action time trajectory to maintain the correspondence between the node and the action rhythm, thereby obtaining the progress rhythm node information.

[0011] The present invention improves upon this invention by including the following improvements: the joint rotation sequence structure includes a rotation trajectory index, joint angle segments, and action timing segments; the rhythm transition positioning point includes a rhythm switching reference, an action jump node group, and a local rhythm triggering factor; the sequence structure center label includes a sequence core node, local peak and valley anchor points, and a structure alignment identifier; the node offset parameter includes an offset amplitude factor, a center difference identifier, and a segment alignment correction item; and the progress rhythm node information includes a rhythm node label, a progress association index, and a learning sequence anchoring item.

[0012] The present invention is improved in that the steps for obtaining the joint rotation sequence structure are specifically as follows:

[0013] S111: Based on the posture capture camera equipment around the training area, analyze the spatial coordinates and rotation orientation data of each frame of the collected joints, determine the correspondence between each joint and the trainee's body structure, and assign the joint parameters to different categories according to the set rules to obtain the joint type mapping group.

[0014] S112: Based on the joint type mapping group, compare the intra-frame rotation orientation parameters of each type of joint, identify the associated parameters of the joint space rotation state according to the rotation direction set by the type of each joint, and collect the joint states by number to obtain the frame-level rotation state sequence.

[0015] S113: Based on the frame-level rotation state sequence, calculate the rotation orientation change of each joint in consecutive frames, and combine the corresponding frame number to perform temporal arrangement and difference integration of the rotation state of the same joint to obtain the joint rotation sequence structure.

[0016] The present invention is improved in that the step of obtaining the rhythm transition positioning point is specifically as follows:

[0017] S211: Based on the joint rotation sequence structure, analyze the spatial rotation state of the same joint in consecutive frames, calculate the rotation direction angle of each pair of adjacent frames, arrange the rotation change data of each group in time order, and obtain the inter-frame rotation amplitude sequence.

[0018] S212: Based on the inter-frame rotation amplitude sequence, compare the rotation change rate of each joint in consecutive frames, analyze the change amplitude of data points in the time series, retrieve the time periods in which continuous change or acceleration trends occur, and obtain the set of inter-frame change trends.

[0019] S213: Based on the set of inter-frame change trends, determine the change direction and rotation amplitude fluctuation of each joint in the time series, identify the frame number in the sequence where the direction reverses and the rotation amplitude fluctuates, and mark them as key moments of rhythm jump in sequence to obtain rhythm transition positioning points.

[0020] The present invention is improved in that the step of obtaining the center index of the sequence structure is specifically as follows:

[0021] S311: Based on the rhythm transition positioning point, analyze the rotation angle data of all frames in the corresponding segment, compare the angle changes of each frame with the frames before and after, determine the trend distribution of angle changes in the sequence, and obtain the joint rotation index sequence.

[0022] S312: Based on the joint rotation index sequence, filter the angle change directions of consecutive frames, calculate the degree of angle jump of each frame relative to the preceding and following frames, using the formula:

[0023] ;

[0024] Obtain the local normalized angle jump factor, label the angle abrupt change frames and angle stable frame numbers, and obtain the angle jump label set, where... Indicates the first The local normalized angle jump factor of the frame. Indicates the first Frame joint rotation angle data, Indicates the first Frame joint rotation angle data, Indicates the first Frame joint rotation angle data, Indicates the first The average value of the angle change within three frames centered on the frame;

[0025] S313: Based on the angle jump marker set, determine the distribution of each tag in the sequence, analyze the symmetry and concentration of the distribution, identify the key tags near the center number of the sequence, and obtain the center number of the sequence structure.

[0026] The present invention is improved in that the step of obtaining the node offset parameter is specifically as follows:

[0027] S411: Analyze the frame number of the center index of the sequence structure and the rhythm transition positioning point on the time axis, and form a set of changes in consecutive numbers by calculating the sequential differences between the numbers, to obtain the frame number interval sequence;

[0028] S412: Based on the frame numbering interval sequence, determine the relationship with the progress node fluctuation range limit. If the numbering span exceeds the limit, optimize the start and end numbers of the analysis segment. By adjusting the segment length, calculate the coupling characteristics of the rotation angle change and numbering interval within the compressed segment, obtain the rotation trend of the compressed numbering segment, and obtain the segment repositioning numbering reference.

[0029] S413: Based on the segment relocation numbering benchmark, compare the position of the sequence structure center number in the sequence, analyze the numbering differences, filter the proportional characteristics of the differences and the span of the compressed segment, and obtain the node offset parameter.

[0030] The present invention is improved in that the step of obtaining the progress rhythm node information is specifically as follows:

[0031] S511: Based on the node offset parameter, analyze the correspondence between the nodes in the calibration frame number and the action timing trajectory, compare the mapping state formed by the paragraph alignment correction item and the center difference identifier, identify the frame number that meets the progress adjustment requirements, and obtain the number mapping reference set.

[0032] S512: Based on the numbering mapping reference set, compare the node numbering order with the action time sequence trajectory, adjust the labeling position of the node in the time sequence trajectory, rearrange the node according to the adjusted time sequence structure, and obtain the node time sequence matching sequence.

[0033] S513: Based on the node timing matching sequence, determine the matching status of the node number and the rhythm structure label in the progress feedback sequence, synchronously learn the identification content of the progress feedback system, and combine the linkage relationship between the node and the rhythm structure to obtain the progress rhythm node information.

[0034] The present invention is improved in that the three-dimensional posture information of the joint refers to the coordinate position and rotation orientation of the joint in space along three directions as recorded by the posture capture device, the amount of rotation angle change refers to the angle difference of the same joint between consecutive frames, and the rate of change between adjacent frames refers to the trend of the speed of angle change between two consecutive frames.

[0035] A dance movement learning progress adaptive adjustment system, the system comprising:

[0036] The rotation sequence module analyzes the three-dimensional pose information of the joints in each frame based on the pose capture camera equipment around the training area, identifies the rotation state of the joints in a single frame, compares the changes in joint position in time sequence, and associates the angle changes with the frame number to obtain the joint rotation sequence structure.

[0037] The transition detection module calculates the rotation angle change of consecutive frames based on the joint rotation sequence structure, compares the change rate of adjacent frames, analyzes the angle change direction and amplitude fluctuation, determines whether there is a rhythm jump feature, identifies abnormal fluctuation frames, and obtains the rhythm jump location point.

[0038] Based on the rhythm transition positioning point, the center identification module sorts the rotation angles of each frame in the analysis segment by number, compares the angle changes frame by frame, identifies frames with prominent changes or frames with gentle changes, and obtains the center label of the sequence structure.

[0039] The offset correction module determines the distance between the center index of the sequence structure and the frame number of the rhythm transition positioning point, compares the distance with the node fluctuation range, and if it exceeds the range, adjusts the length of the analysis segment, reduces the frame number range, and corrects the center number to obtain the node offset parameter.

[0040] The rhythm calibration module records the calibration frame number based on the node offset parameter, updates the node identifier in combination with the trainee's movement time trajectory, so that the node and the movement rhythm are in correspondence, and obtains the progress rhythm node information.

[0041] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0042] In this invention, the time sequence formed by the rotation of the action is continuously analyzed. The rotation trajectory of the joint in the time axis is used as the basic structure. The jump position inside the action is determined by comparing the direction and scale of the rotation change between consecutive frames. The structure center label is extracted in the local frame segment by the relative relationship of the angle change. Then, the offset parameter is obtained by the interval relationship between the center label and the jump position. When the offset exceeds the range, the analysis segment is readjusted. This makes the node recognition based on the continuous rotation structure, gives the jump position a dynamic reference, makes the node stable under the fluctuation of the action, and makes the progress rhythm node information consistent with the action structure. This solves the problem that the static node mode cannot adapt to the continuous action change. Attached Figure Description

[0043] Figure 1 This is a flowchart of the main steps of the present invention;

[0044] Figure 2 This is a flowchart illustrating the process of obtaining the joint rotation sequence structure in this invention.

[0045] Figure 3 This is a flowchart illustrating the process of obtaining the rhythm transition positioning point in this invention.

[0046] Figure 4 This is a flowchart illustrating the process of obtaining the center label of the sequence structure in this invention;

[0047] Figure 5 This is a flowchart illustrating the process of obtaining node offset parameters in this invention.

[0048] Figure 6 This is a flowchart illustrating the process of obtaining progress rhythm node information in this invention. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0050] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0051] All user-related information involved in this invention (including but not limited to biometric information, identity verification information, behavioral data, device information, and other data that can be used for identity verification and personalized services) is collected and processed with the user's full knowledge and voluntary consent. The collection, storage, and use of all information strictly comply with applicable national and regional laws and regulations, and meet relevant data protection standards and policy requirements. The use of data is limited to purposes necessary for providing the technical services of this invention, and reasonable technical and management measures will be taken to ensure the security and confidentiality of users' personal information in terms of information protection and privacy.

[0052] Example: Please refer to Figure 1 This invention provides a technical solution: a method for adaptively adjusting the learning progress of dance movements, comprising the following steps:

[0053] S1: Based on the posture capture camera equipment around the training area, analyze the three-dimensional posture information of each joint in each frame, classify the joint type according to the trainee's body structure, identify the spatial rotation state of each joint in a single frame, compare the joint position changes in chronological order, and use the frame number to associate the angle changes at each moment to obtain the joint rotation sequence structure.

[0054] S2: Based on the joint rotation sequence structure, calculate the change in rotation angle of the same joint between consecutive frames, compare the rate of change between adjacent frames, and determine whether the trend of change meets the rhythm jump condition of the training action by analyzing the direction and amplitude fluctuation of the angle change. Select the frame where abnormal fluctuation occurs as the center of the key analysis interval to obtain the rhythm jump positioning point.

[0055] S3: Based on the rhythm transition positioning point, the rotation angle of each frame in the analysis segment is arranged in the order of frame number. The change in rotation angle is compared frame by frame to identify frames with prominent angle changes or frames with gentle changes in the frame sequence. The structural distribution is determined according to the center frame number to obtain the center label of the sequence structure.

[0056] S4: Determine the frame number distance between the center label of the sequence structure and the rhythm transition positioning point, compare the distance with the progress node fluctuation range, and if it exceeds the reasonable range, adjust the analysis segment length, reduce the frame number range, re-compare the rotation angle change, and correct the center number to obtain the node offset parameter.

[0057] S5: Based on the node offset parameter, the calibration frame number is entered into the learning progress feedback system. Combined with the trainee's action time trajectory, the identifier of the action node in the progress feedback sequence is updated to complete the linkage between the node and the trainee's rhythm structure and obtain the progress rhythm node information.

[0058] The joint rotation sequence structure includes rotation trajectory index, joint angle segments, and action timing segments. The rhythm transition positioning points include rhythm switching benchmark, action jump node group, and local rhythm triggering factor. The sequence structure center label includes sequence core node, local peak and valley anchor point, and structure alignment identifier. The node offset parameters include offset amplitude factor, center difference identifier, and paragraph alignment correction item. The progress rhythm node information includes rhythm node label, progress association index, and learning sequence anchor item.

[0059] In S1, the three-dimensional posture information of the joint refers to the coordinate position and rotation orientation of the joint in three directions in space, recorded by the posture capture device, which is used to describe the complete spatial posture of the joint at a single moment; the joint type refers to the different force-generating parts divided according to the human body structure, such as shoulder, elbow, hip, knee, etc., which is used to clarify the independent role of each joint in the action sequence; the spatial rotation state refers to the angle change of the joint around its own rotation axis, which is used to show the rotation direction and rotation area of ​​the joint in a certain frame; the joint position change refers to the position difference of the same joint between two consecutive moments, which is used to describe the movement trajectory formed by the movement over time; the frame number refers to the sequential marking of each frame of sampled data by the posture capture device according to time, which is used to locate the specific moment of the action on the time axis; the angle change at each moment refers to the difference in rotation angle generated between consecutive moments, which is used to describe the change in rotation trend during the action.

[0060] In S2, the rotation angle change refers to the angle difference of the same joint between consecutive frames, used to reflect the continuous change of the action rotation; the rate of change between adjacent frames refers to the speed trend of the angle change between two consecutive frames, used to determine whether the action change tends to be stable or intensified; directional and amplitude fluctuations refer to the positive and negative trends of angle changes in direction and the magnitude of changes in amplitude, used to identify whether there are obvious fluctuations in the action; the change trend refers to the forward and backward direction of the angle data on the time axis, such as gradually rising, gradually falling, or rapidly jumping; the rhythm jump condition refers to the sudden change in the rhythm of the action, such as an instantaneous amplitude switch after a stable rotation, used to determine whether a new rhythm segment has been formed; the frame with abnormal fluctuations refers to the frame with obvious differences from the previous and subsequent change patterns, used to identify key moments in the rhythm structure of the action; the center of the key analysis interval refers to the center position of the local time segment established around the frame with abnormal fluctuations, used as a reference point for subsequent local feature identification.

[0061] In S3, a frame sequence refers to a time segment consisting of multiple consecutive frames of data captured around a rhythmic transition positioning point, used for local motion structure analysis; a frame with a prominent angle change amplitude refers to a frame whose angle change is significantly increased compared to adjacent frames, used to reflect the peaks in the local motion structure; a frame with a gradual change refers to a frame with a relatively stable angle change, used to determine the moment when the motion approaches a transition state; the center frame number refers to the key frame number within the segment used to describe the central structure of the motion, used to determine the central reference point of the local structure; and the structural distribution refers to the overall arrangement of the angle changes of each frame within the analysis segment, used to identify the local peak-valley structure and sequence pattern of the motion.

[0062] In S4, frame number distance refers to the numbering difference between two specific frames, used to assess whether they originate from the same action structure; progress node fluctuation range refers to the upper and lower intervals within which action nodes are allowed to deviate on the time axis, used to determine whether the node is within acceptable error; reasonable interval refers to the numbering range within which node offset is acceptable, used to determine whether the analysis segment needs to be readjusted; segment length refers to the time span of the continuous frame window used for analysis, used to control the scale of local structure analysis; frame number range refers to the numbering interval formed by the minimum and maximum frame numbers within the segment, used to define the temporal boundaries of the analysis segment; and corrected center number refers to the center frame number obtained after repositioning after adjusting the analysis segment, used to update the center parameters of the action node.

[0063] In S5, the learning progress feedback system refers to the progress management system used to record, update, and display action node data, and is the output end of the entire progress adjustment process; the action sequence trajectory refers to the continuous change path formed by the trainee's actions on the time axis, used to describe the overall execution sequence of the actions; the progress feedback sequence refers to the node chain of each recorded action node arranged in time, used to display structured information about the learning progress; the rhythm structure linkage refers to the correspondence formed between newly identified nodes and existing rhythm structures, used to maintain the overall coherence of the training rhythm.

[0064] Please see Figure 2 The specific steps for obtaining the joint rotation sequence structure are as follows:

[0065] S111: Based on the posture capture camera equipment around the training area, analyze the spatial coordinates and rotation orientation data of each frame of the collected joints, determine the correspondence between each joint and the trainee's body structure, and assign the joint parameters to different categories according to the set rules to obtain the joint type mapping group.

[0066] The trainee's 3D pose data is collected sequentially in each frame. First, the raw image data is transformed based on the camera position information, unifying the joint coordinates in the local coordinate system to the global space of the training area. Then, the 3D position and orientation information of each node in this space are read, and nodes corresponding to the major joints in the trainee's body, such as the shoulder, elbow, hip, knee, and ankle, are extracted. For each node, its height, front-back, left-right position offset from the center point of the torso in space are matched with the standard human skeleton model. The matching method includes calculating the distance difference between the spatial position of the node and the point of the same part in the reference model. If a node is kept within a specified spatial distance from the reference point of a certain part in multiple consecutive frames, then the node is identified as the representative point of that joint category. For example, if a node is always 20 cm above and to the right of the center point of the torso in 5 consecutive frames, and its horizontal distance fluctuation is less than 5 cm, it can be identified as the right shoulder point. Then, all nodes are grouped according to joint type, a mapping dictionary is generated, and a one-to-one correspondence between different numbered nodes and joint types is completed, and the mapping group is updated.

[0067] S112: Based on the joint type mapping group, compare the intra-frame rotation orientation parameters of each type of joint, identify the associated parameters of the joint space rotation state according to the rotation direction set by the type of each joint, and collect the joint states by number to obtain the frame-level rotation state sequence.

[0068] Each frame of the image is processed frame by frame, and the rotation orientation data of each type of joint in each frame is extracted. For each type of joint, according to the preset requirement of the most important rotation direction in that category, the rotation orientation parameter of each joint in that frame is extracted and used as the judgment criterion. Then, the rotation values ​​of all joints in the same category in the current frame are compared pairwise, the difference in the main direction is calculated, and compared with the preset rotation consistency reference value. If the difference is within the set threshold, the joints of that type in the frame are considered to have consistent rotation state. If it exceeds the threshold, the state is inconsistent. For example, if the left and right shoulder nodes belonging to the same shoulder have a rotation direction difference of less than 10 degrees in the current frame, they are considered to be consistent. Otherwise, they are marked as different rotation states. The rotation state results of each type of joint in the current frame are recorded and classified. The state information is aggregated according to the frame number to generate a frame-level rotation state sequence covering the entire action cycle.

[0069] S113: Based on the frame-level rotation state sequence, calculate the rotation orientation change of each joint in consecutive frames, and combine the corresponding frame number to perform temporal arrangement and difference integration of the rotation state of the same joint to obtain the joint rotation sequence structure.

[0070] The analysis focuses on the temporal trends of each joint. For each joint, the rotational orientation change values ​​are arranged sequentially according to frame number. The change values ​​of rotational orientation between adjacent frames are read continuously and recorded sequentially. This method obtains the rotational changes of the joint throughout the entire time period and generates a time series data set. Based on this, all frame change data of the joint are segmented according to a time sliding window, and the maximum change value and average change level in each segment are extracted. Then, the maximum value and average value in each window are compared. If the change value of a certain frame is significantly higher than the jump reference range set by the average value of that segment, the frame is marked as a critical change moment of the action. For example, in a 10-frame interval, if the average rotational change amplitude is 11 degrees, and a certain frame shows an 18-degree change, the frame is marked as a critical fluctuation frame. The frame number and the corresponding joint type are added to the rotational transition index list. The rotational change data and the index marks of the critical jump frames corresponding to all joints are summarized to form a complete joint rotation sequence structure.

[0071] Please see Figure 3 The specific steps for obtaining the rhythm transition positioning point are as follows:

[0072] S211: Based on the joint rotation sequence structure, analyze the spatial rotation state of the same joint in consecutive frames, calculate the rotation direction angle between each pair of adjacent frames, arrange the rotation change data of each group in time order, and obtain the inter-frame rotation amplitude sequence.

[0073] The rotation state information corresponding to each joint number is arranged sequentially along the time axis. First, the rotation vector of each joint in consecutive frames is read. For the same joint node in two adjacent frames, the angle value under its main rotation direction is extracted, and the direction change vector of the joint between the two frames is constructed. The angle between the two vectors is used as the basis for judging the difference in rotation direction. The magnitude of the angle of rotation direction is obtained according to the angle relationship between the vectors in three-dimensional space. The adjacent rotation direction combinations of all frames are processed sequentially in time order to complete the rotation angle analysis between all adjacent frames of the joint in the entire time series. For example, for the right elbow joint in the 1st frame... Between frame 2 and frame 13, the rotation directions are downward and forward, and forward, respectively. After calculating the three-dimensional direction vector difference, the angle between the two is 42 degrees. This value is used as the rotation change amplitude between the two frames. Based on this, the 13th and 14th frames, and the 14th and 15th frames are processed sequentially. All rotation direction angle values ​​are arranged in order by time number to form the rotation change sequence of the joint. Then, the same operation is performed on all joints to record the spatial angle difference of the rotation direction of each joint node between adjacent frames. These are then integrated into multiple data linked lists according to the frame number to obtain the inter-frame rotation amplitude sequence corresponding to each joint.

[0074] S212: Based on the inter-frame rotation amplitude sequence, compare the rotation change rate of each joint in consecutive frames, analyze the change amplitude of data points in the time series, retrieve the time periods with continuous change or acceleration trend, and obtain the set of inter-frame change trends.

[0075] The rotation angle change value of each joint node between adjacent frames on the time axis is read one by one. The change value in the time series is calculated, that is, the difference in rotation amplitude between the current frame and the previous frame. This difference reflects the rate of angle change and is used to determine the rate of angle change. At the same time, the frame number and its rate change value of each frame are marked. All change results are stored in a list in sequence. Then, for each segment of data in the list, a fixed-length sliding time window is used for traversal processing. The maximum and minimum change values ​​of the segment are extracted in each window, and their difference is calculated as the change amplitude of the segment. This difference is then compared with the set amplitude reference benchmark. The values ​​are compared. If the change range of a certain segment is greater than the reference value, it is considered that the segment has a rotational trend. If the rate values ​​of multiple consecutive frames in the same segment gradually increase or decrease, and the difference in the change value of each frame exceeds the rate reference value by more than 3 degrees, it is considered that there is a continuous acceleration or deceleration trend. For example, for the left knee joint in the interval from frame 20 to frame 30, if the angle change values ​​of consecutive frames are 5, 7, 10, 14 and 19 degrees respectively, it can be identified as a continuous acceleration state. This segment is marked as a trend growth segment and added to the inter-frame change trend set. The frame number and change information of all trend segments that meet the set conditions are summarized into this set.

[0076] S213: Based on the set of inter-frame change trends, determine the change direction and rotation amplitude fluctuation of each joint in the time series, identify the frame number in the sequence where the direction reverses and the rotation amplitude fluctuates, and mark them as key moments of rhythm jump in sequence to obtain rhythm transition positioning points.

[0077] The rotation direction and angle fluctuations of each joint node in each segment are analyzed. For each frame segment in the trend set, the rotation direction change trend is read frame by frame and the direction polarity is judged. It is identified whether there is a reversal between the current frame direction and the previous frame, that is, whether the current direction vector component value has the opposite sign to the previous frame. At the same time, it is checked whether there is a sudden increase or decrease in the angle change amplitude of the frame and the two adjacent frames. If both the direction reversal and the amplitude change are greater than the amplitude jump threshold (this threshold can be set to 8 degrees), then the frame is marked as a direction jump node. For example, in frame 45, the rotation direction of the right hip joint changes from forward to backward, and the rotation angle suddenly changes from 9 degrees to 21 degrees. Then the frame satisfies the two jump judgment conditions and is marked as a key rhythm transition point. The frame segments of other joints are traversed in the trend set, and the above judgment operation is repeated. The frame numbers that meet the conditions are summarized and serialized according to the time order to construct a complete list of rhythm transition positioning points.

[0078] Please see Figure 4 The specific steps for obtaining the center index of the sequence structure are as follows:

[0079] S311: Based on the rhythm transition positioning point, analyze the rotation angle data of all frames in the corresponding segment, compare the angle changes of each frame with the frames before and after, determine the trend distribution of angle changes in the sequence, and obtain the joint rotation index sequence.

[0080] An analysis segment is constructed by extending a fixed number of frames forward and backward from each positioning point. The joint rotation angle data of all frames within this segment are read. For each joint node in a frame, the angle value in its main rotation direction is extracted and sorted by frame number. The angle difference between each frame and the previous frame is calculated. Frames with positive angle increases are marked as upward trends, frames with negative angle decreases are marked as downward trends, and frames with angle differences less than a set angle micro-motion threshold (e.g., 2 degrees) are marked as stable states. The relative proportions of upward, downward, and stable states in the entire segment are then analyzed to determine whether the motion structure of the segment is continuously increasing, decreasing, or undergoing repeated changes. Finally, continuous segment marking is performed for the angle trend states of all frames to identify each complete segment. The positional boundaries of the rising, falling, and stable segments are recorded, along with the frame numbers of those with significant changes. All boundary points of angular change trends are numbered and archived. Furthermore, a trend map of joint angle changes near each rhythm transition point is constructed. For example, if the rhythm transition point is frame 85, the analysis segment can be selected as frames 75 to 95. If the shoulder joint rotation angle continuously increases from 13 degrees in frame 75 to 38 degrees in frame 90 within this segment, then this segment exhibits an overall increasing angle trend. If the angle decreases to 25 degrees between frames 91 and 95, a complete rising-falling trend structure is formed. Frame 90 is recorded as the peak of change. All trend change frame numbers are sequentially arranged to form an index table, creating a joint rotation index sequence corresponding to the rhythm transition point.

[0081] S312: Based on the joint rotation index sequence, filter the angle change direction of consecutive frames, calculate the degree of angle jump of each frame relative to the preceding and following frames, using the following formula:

[0082] ;

[0083] Obtain the local normalized angle jump factor, label the angle abrupt change frames and angle stable frame numbers, and obtain the angle jump label set, where... Indicates the first The local normalized angle jump factor of the frame. Indicates the first Frame joint rotation angle data, Indicates the first Frame joint rotation angle data, Indicates the first Frame joint rotation angle data, Indicates the first The average value of the angle change within three frames centered on the frame;

[0084] The local normalized angle jump factor is an indicator used to measure the magnitude of the rotation angle change of a joint in a given frame relative to the adjacent frames before and after it. It is normalized by the average angle change within the same segment. The calculation involves adding the absolute values ​​of the angle changes in a given frame to those in the preceding and following frames, and then dividing by the average angle change over three frames centered on that frame. This reflects the relative strength of the rotation jump magnitude within that segment. Normalization eliminates the influence of differences in the overall amplitude of different movements on the determination of abnormal fluctuations, more accurately identifying abnormal jumps or rhythmically stable segments in a motion sequence. It is a metric used to mark frames with abrupt angle changes and frames with stable angles.

[0085] The process involves selecting substructures consisting of three consecutive frames in the sequence. By calling the joint rotation angle value corresponding to each frame number, the angular jump amplitude of that frame relative to its preceding and following frames is calculated. The calculation process involves calculating the difference between the current frame's angle value and the angle values ​​of adjacent frames, and summing the absolute values ​​of these two differences as the original jump amplitude. Then, the average angle difference of these three substructure frames is used as a normalization reference for local mean normalization. This removes the influence of absolute deviations in the angle amplitude and normalizes the relative intensity of angle changes in the time series. Mean normalization is used, which divides the total jump amplitude by twice the average difference to form a dimensionless relative intensity index. This index is then processed using a formula. Taking frame number 50 as an example, let the angle data be as follows:

[0086] , , ;

[0087] The original angle jump calculation result is:

[0088] , The total is 9.6°;

[0089] The average angle difference between the three frames is calculated as follows:

[0090] ;

[0091] The normalized value is ;

[0092] Taking frame number 51 as an example, let... , , ;

[0093] The corresponding original difference is , The original jump amplitude was 5.6°;

[0094] Mean:

[0095] ;

[0096] The normalized jump factor is ;

[0097] Continuing with frame number 52 as an example, the angles are respectively , , ;

[0098] The original difference is , The total is 1.6°;

[0099] mean Normalized jump factor ;

[0100] The preset normalization jump factor judgment criteria can be divided into the following three numerical ranges:

[0101] when and When the frame is defined as an angle change frame, it means that the angle difference between the current frame and the previous frame reaches a relatively drastic change on the normalized scale, which belongs to the fluctuation high point in the rhythm structure.

[0102] when and When this happens, the frame is defined as an angle-stable frame, meaning that the angle jump of the frame relative to the adjacent frames remains within a stable range and is in the transition zone of rhythmic movement;

[0103] when When the angle change is small, the frame is defined as a frame with a gradual change in angle, indicating that the angle change between the frame and the neighboring frames is small on the normalized scale, and can be regarded as a sequence edge frame outside the rhythm fluctuation area.

[0104] Based on the above interval definition, the currently calculated normalized jump factor values ​​are as follows:

[0105] Frame 50 It falls within [0.8, 1.2] and is classified as an angle change frame;

[0106] Frame 51 These are also classified as angle-change frames;

[0107] Frame 52 They are also classified as angle-change frames.

[0108] The results show that in the three-frame window centered on frame number 50, the three consecutive frames constitute abrupt change frames, indicating that this position forms a strong jump node in the action rhythm structure and has significant structural anchoring features. Therefore, the numerical results can be directly used to construct an angle jump marker set and serve as the input for the enhanced region of the rhythm transition positioning point.

[0109] S313: Based on the angle jump marker set, determine the distribution of each tag in the sequence, analyze the symmetry and concentration of the distribution, identify the main tags near the center of the sequence, and obtain the sequence structure center label;

[0110] For each frame number, its positional distribution is read, and a complete set of frame numbers is constructed sequentially along the timeline. Then, the center frame number between the minimum and maximum numbers in this set is calculated as a reference axis. Using this center position as a reference point, the distribution offset of all tag numbers is calculated to both sides. The difference between each tag and the center number is determined, and the number of frames falling into different offset intervals is counted. For example, if ±5 frames from the center point is defined as a high-concentration area, ±10 frames as a secondary-concentration area, and the remaining frames as discrete areas, the degree of tag concentration can be determined. If the number of tags in the high-concentration area exceeds 60% of the total number of tags, the tag set is considered to have a concentrated distribution characteristic. Further checks are performed on the left-right symmetry of the tag distribution. Calculate the number of tags on the left and right sides of the center. If the difference is less than or equal to 1, it is considered approximately symmetrical. If the difference is greater than 1 and less than or equal to 3, it is considered slightly off-center. If the difference is greater than 3, it is considered an off-center distribution. Combining the degree of concentration and symmetry, select the main tag that is close to the center frame number and is in a highly concentrated area as the key position reference. For example, if the current angle jump tag set is distributed between frames 120 and 140, and its center number is frame 130, there are 12 tags between frames 125 and 135, accounting for 75% of the total number. At the same time, the number of tags on the left and right frames are 6 and 6 respectively. Then this segment meets the requirements of concentration and symmetry. Frame 130 is used as the main tag frame number, and the output is the center label of the sequence structure.

[0111] Please see Figure 5 The specific steps for obtaining the node offset parameter are as follows:

[0112] S411: Analyze the frame number of the sequence structure center index and rhythm transition positioning point on the time axis, and form a set of changes in consecutive numbers by calculating the sequential differences between the numbers, to obtain the frame number interval sequence.

[0113] Frame numbers are extracted from the constructed set of rhythm transition location points and the list of center labels of the sequence structure, respectively. A time series index corresponding to each frame number is established. Each pair of "center label frame numbers" and their corresponding "rhythm transition frame numbers" is paired according to time sequence. The difference between the two numbers is read group by group. The rhythm transition frame number is subtracted from the center label frame number to obtain the frame number difference value, which is added to the difference sequence. The difference sequence is then organized, and all number difference values ​​are arranged in ascending order. Simultaneously, a number direction marker is generated, distinguishing between positive and negative differences. A positive value indicates that the center label appears after the rhythm transition frame, and a negative value indicates that it appears before. The difference values ​​are then divided into intervals, with three threshold intervals set, such as frame difference. Frames ±0 to ±5 are designated as the "near distance segment," frames ±6 to ±15 as the "medium distance segment," and frames exceeding ±16 as the "far distance segment." The attribution of each set of differences within the interval is statistically analyzed. Groups with frame number differences exceeding ±20 frames are marked for subsequent scaling adjustments. Combining the continuous distribution characteristics of each set of frame number differences, the set of frame number differences between all center markers and rhythm transition points is recorded, and a number interval list is generated. For example, if the rhythm transition point is frame 100 and the corresponding center marker is frame 111, the number difference is +11. If the next frame pair is the rhythm transition point frame 150 and the center frame is frame 138, the difference is -12. The two differences are combined into a continuously changing dataset, thus forming a frame number interval sequence.

[0114] S412: Based on the frame number spacing sequence, determine the relationship with the progress node fluctuation range limit. If the number span exceeds the limit, optimize the start and end numbers of the analysis segment. By adjusting the segment length, calculate the coupling characteristics between the rotation angle change within the compressed segment and the number spacing, using the formula:

[0115] ;

[0116] Obtain the rotation trend of the compressed numbering segment to obtain the segment repositioning numbering reference, where, This indicates the rotation trend linkage amount corresponding to the segment repositioning number reference. This indicates the total number of frames within the compressed numbering segment. Indicates the first segment in the compressed section The range of rotation angle changes between the previous and current frames. Indicates the first The frame number spacing between frames and rhythm transition positioning points. Indicates the first The numbering span of the compressed segment containing the frame;

[0117] The rotation trend linkage refers to the coupling relationship between the rotation angle change of each frame action within a compressed numbering segment and the numbering distance of that frame relative to the rhythm transition positioning point within the segment. By averaging this coupling characteristic of all frames, it reflects the overall correlation between the action rotation change and the rhythm change node in the entire analysis segment. It is the overall dynamic change intensity formed by combining the rotation amplitude change of each frame action within a specific segment with the temporal distance of the frame relative to the rhythm change point. The quantification results can be used to judge the synchronization and response characteristics between rotation change and rhythm structure node.

[0118] The frame numbers in the original segment that are far from the rhythm transition location point are shrunk towards the center to construct a compressed segment number sequence. Let the original number range be 145 to 168, corresponding to a total of 24 frames. The rhythm transition location point number is 159, and the detection number span is 23 frames, exceeding the allowable number fluctuation range of 20 frames. Therefore, the compressed number range is 148 to 167, with a compressed frame count of 20 frames. Then, the complete frame number sequence within this compressed number segment is obtained. And calculate the rotation angle change for each frame. The rotation angle is calculated from the changes in 3D joint coordinates measured by the attitude capture system. Let the original angle changes from frame 1 to frame 5 be 12.5°, 9.7°, 14.8°, 11.2°, and 10.3°, respectively. The corresponding numbering intervals from frame 1 to frame 5 are 11, 10, 9, 8, and 7, respectively. The compressed segment numbering span is 20. The numbering intervals are normalized using linear normalization to convert them into dimensionless ratios. The normalized numbering intervals from frame 1 to frame 5 are 0.55, 0.5, 0.45, 0.4, and 0.35. The angles are then converted from angle values ​​to radians. The normalized angle changes from frame 1 to frame 5 are 0.2182, 0.1693, 0.2583, 0.1956, and 0.1797. Subsequently, frame-by-frame calculations are performed according to the formula:

[0119] The product in frame 1 is:

[0120] ;

[0121] Frame 2 is:

[0122] ;

[0123] Frame 3 is:

[0124] ;

[0125] Frame 4 is:

[0126] ;

[0127] Frame 5 is:

[0128] ;

[0129] Taking the above 5 frames as an example, the sum is 0.462, and the number of frames is 5. The rotation trend linkage amount is then calculated as follows:

[0130] ;

[0131] Set rotation trend linkage amount The preset reference interval is distributed in three segments:

[0132] when When the angle change amplitude is small or the distribution is relatively discrete, the synchronization between the rotation change and the rhythm transition node is insufficient, indicating that the current analysis frame sequence has not effectively covered the peak area of ​​motion change. At this time, it is recommended to expand the analysis segment range or relocate the start and end frames of the segment.

[0133] when When defined as the central linkage segment, it indicates that the angle changes of some frames within the segment have a certain concentration and rhythm response characteristics in time, but there are still local rhythm offsets. It is suitable for performing number adjustment operations. The center number can be further derived by analyzing the frame number density distribution or the fluctuation range of angle changes.

[0134] when When a segment is defined as a high-linkage segment, the rotation changes of the frames in the segment have clear rhythmic clustering characteristics, high coupling with the response of rhythmic transition nodes, and clear and concentrated action change positions. At this time, the frame number corresponding to the rotation peak in the segment can be directly selected as the structure alignment number.

[0135] The calculated rotational trend linkage is If the analysis is in a low-linkage segment, it means that the current analysis frame sequence has not effectively covered the peak area of ​​action change. In this case, it is recommended to expand the analysis segment range or relocate the start and end frames of the segment.

[0136] S413: Based on the segment relocation numbering benchmark, compare the position of the sequence structure center number in the sequence, analyze the numbering differences, screen the proportional characteristics of the differences and the compressed segment span, and obtain the node offset parameter;

[0137] As a reference number input, it is matched one-to-one with the set of center labels of the sequence structure. The number difference value between the relocation reference number and the corresponding center label number is read, and the difference data is recorded in groups. Then, it is compared with the original segment compressed span corresponding to each difference value. For each segment, the start and end numbers of the corresponding frame segment after compression are read, and the total number of frames is calculated as the compression span. Then, the number difference value is proportionally calculated with the compression span to determine the proportion of the difference value in the segment. If the proportion exceeds a set threshold, for example, if the difference between the center number and the relocation number is 14 frames and the segment compression span is 40 frames, then the proportion is 0.35. If it exceeds the preset proportion benchmark of 0.25, then the number... The difference is considered an indicator of anomaly in offset. This ratio is used as the initial offset parameter. Then, all ratios are averaged and their standard deviation is obtained. Ratios that deviate from the average by more than twice the standard deviation are filtered and marked as severely offset frame groups. In addition, the number of positive and negative offset ratios is counted, and cases with a unilateral offset ratio greater than 70% are marked. This data will be used for subsequent progress correction judgment. For example, if 14 out of 20 groups of numbered differences in the sequence are positive and their offset ratios are all greater than 0.3, it is judged that the current progress control point is generally lagging behind the rhythm point, which is recorded as a positive offset trend. All numbered difference values, relative offset directions and ratios after proportional filtering are output to form the node offset parameter.

[0138] Please see Figure 6 The specific steps for obtaining progress rhythm node information are as follows:

[0139] S511: Based on the node offset parameter, analyze the correspondence between the nodes in the calibration frame number and the action timing trajectory, compare the mapping state formed by the paragraph alignment correction item and the center difference identifier, identify the frame number that meets the progress adjustment requirements, and obtain the number mapping reference set.

[0140] Extract all center number differences, compressed paragraph span, and offset ratio information. Read each set of calibration frame numbers and sort them in ascending order. Then, perform mapping and matching with the node numbers recorded in the action sequence trajectory. Take each calibration frame number as a candidate node and find the original node number that is closest to it in the trajectory. Calculate the number difference between the two. If the difference is less than the node matching threshold set range (this threshold can be set to ±5 frames according to the frame frequency), the calibration number is considered to be mappable. Continue to analyze the relative positional relationship between the start and end frame numbers marked in the corresponding paragraph alignment correction item and the calibration number. Determine whether the number is within the 25% interval of the paragraph center. If it meets the condition... If a condition is set, the center difference identifier corresponding to the node is extracted for difference judgment. If the center difference value is less than 10% of the compressed paragraph span, it is considered that the number offset is acceptable. If it is greater than 20%, it is considered that the mapping is abnormal. The middle interval is marked as a critical matching node. All node offset data are traversed and filtered item by item in the above manner, abnormal offset nodes are removed and the calibration frame numbers that meet the mapping conditions are retained. The frame numbers that meet the conditions of temporal position, proportional difference and center correspondence are collected into the number mapping reference set. For example, in the interval with a paragraph compression span of 40 frames, the calibration frame number is frame 123, the center difference value is 3 frames, and the relative difference ratio is 7.5%. Then the number meets the conditions and is included in the reference set.

[0141] S512: Based on the numbering mapping reference set, compare the node numbering order with the node numbering order in the action time sequence trajectory, adjust the node's labeling position in the time sequence trajectory, rearrange the nodes according to the adjusted time sequence structure, and obtain the node time sequence matching sequence.

[0142] Each frame number is read one by one, and the original recorded node is located in the action timing trajectory. If there is a node number that is completely consistent with the reference set number, the correspondence is directly confirmed. If there is a positive or negative frame difference between the node number and the reference number that is less than the preset maximum tolerance range (e.g., set to ±6 frames), the nearest node is selected as the matching point according to the minimum distance matching strategy between the two, and the old and new nodes are compared one by one. Then, all nodes are arranged in ascending order of frame number in the reference set, and the difference between the new position and the number position in the original timing trajectory is recorded. The trend of all number changes is statistically analyzed. For node groups with consecutive offset directions, intra-segment uniform offset correction is performed. For example, if 5 consecutive node numbers in a group are more than 3 frames later than the original node, the entire node group is moved forward 3 frames for uniform adjustment. Then, all node numbers in the entire trajectory are rearranged to keep the number order monotonically unchanged, avoiding repetition and reverse order. The newly generated node time series is recorded, and the adjusted node number order is output. This order is the node timing matching sequence.

[0143] S513: Based on the node time sequence matching sequence, determine the matching of node number with rhythm structure label in the progress feedback sequence, synchronously learn the identification content of the progress feedback system, and combine the linkage relationship between nodes and rhythm structure to obtain progress rhythm node information.

[0144] Each adjusted node number is compared with the rhythm structure label list in the progress feedback sequence. Each node number is read and the label node with the closest number in the rhythm label is found. If the difference in number is within the rhythm matching interval threshold (this threshold is set to within 5 frames according to the frame rate), the node is considered to have successfully matched the rhythm label. If the difference in number exceeds this range, but there is still a record in the node mapping reference set, its rhythm label is determined according to the mapping relationship in the reference set. If the difference between the number and any rhythm label exceeds the maximum threshold and is not in the reference set, it is temporarily marked as a non-matching node and does not participate in rhythm linkage synchronization. After the matching relationship is determined, the rhythm label of all successfully matched nodes is synchronized to the record table in the learning progress feedback system. The label field of each node is updated to its corresponding rhythm label name. At the same time, its mapping source is recorded as "direct matching" or "reference matching". A synchronization table is generated and the synchronization information is linked with the node trajectory to form a complete information set containing rhythm label, corresponding number and synchronization status field. The output is the progress rhythm node information.

[0145] A dance movement learning progress adaptive adjustment system, comprising:

[0146] The rotation sequence module analyzes the three-dimensional pose information of the joints in each frame based on the pose capture camera equipment around the training area, identifies the rotation state of the joints in a single frame, compares the changes in joint position in time sequence, and associates the angle changes with the frame number to obtain the joint rotation sequence structure.

[0147] The transition detection module is based on the joint rotation sequence structure. It calculates the rotation angle change of consecutive frames, compares the change rate of adjacent frames, analyzes the angle change direction and amplitude fluctuation, determines whether there is rhythm jump characteristics, identifies abnormal fluctuation frames, and obtains the rhythm jump location point.

[0148] The center identification module, based on the rhythm transition positioning point, sorts the rotation angles of each frame in the analysis segment by number, compares the angle changes frame by frame, identifies frames with prominent changes or frames with gentle changes, and obtains the center label of the sequence structure.

[0149] The offset correction module determines the distance between the center index of the sequence structure and the frame number of the rhythm transition positioning point, compares the distance with the node fluctuation range, and if it exceeds the range, adjusts the length of the analysis segment, reduces the frame number range, and corrects the center number to obtain the node offset parameter.

[0150] The rhythm calibration module is based on the node offset parameter, inputs the calibration frame number, and updates the node identifier in combination with the trainee's movement time trajectory, so that the node and the movement rhythm are in correspondence, and obtains the progress rhythm node information.

[0151] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for adaptive adjustment of dance movement learning progress, characterized in that, Includes the following steps: S1: Based on the posture capture camera equipment around the training area, analyze the three-dimensional posture information of the joints in each frame, identify the rotation state of the joints in a single frame, compare the changes in joint positions in time sequence, and associate the angle changes with the frame number to obtain the joint rotation sequence structure. S2: Based on the joint rotation sequence structure, calculate the rotation angle change of consecutive frames, compare the change rate of adjacent frames, analyze the angle change direction and amplitude fluctuation, determine whether there is rhythm jump feature, identify abnormal fluctuation frames, and obtain the rhythm jump positioning point. S3: Based on the rhythm transition positioning point, sort the rotation angles of each frame in the analysis segment by number, compare the angle changes frame by frame, identify frames with prominent changes or frames with gentle changes, and obtain the center label of the sequence structure. The specific steps for obtaining the center index of the sequence structure are as follows: S311: Based on the rhythm transition positioning point, analyze the rotation angle data of all frames in the corresponding segment, compare the angle changes of each frame with the frames before and after, determine the trend distribution of angle changes in the sequence, and obtain the joint rotation index sequence. S312: Based on the joint rotation index sequence, filter the angle change directions of consecutive frames, calculate the degree of angle jump of each frame relative to the preceding and following frames, using the formula: ； Obtain the local normalized angle jump factor, label the angle abrupt change frames and angle stable frame numbers, and obtain the angle jump label set, where... This represents the local normalized angle jump factor of the i-th frame. This represents the joint rotation angle data in the i-th frame. This represents the joint rotation angle data for the (i+1)th frame. This represents the joint rotation angle data for the (i-1)th frame. This represents the average value of the angle change within three frames centered at frame i; S313: Based on the angle jump marker set, determine the distribution of each tag in the sequence, analyze the symmetry and concentration of the distribution, identify the key tags near the center number of the sequence, and obtain the center number of the sequence structure. S4: Determine the distance between the center label of the sequence structure and the frame number of the rhythm transition positioning point, compare the distance with the node fluctuation range, and if it exceeds the range, adjust the length of the analysis segment, reduce the frame number range, and correct the center number to obtain the node offset parameter. S5: Based on the node offset parameter, input the calibration frame number, and update the node identifier in combination with the trainee's action time trajectory to maintain the correspondence between the node and the action rhythm, thereby obtaining the progress rhythm node information.

2. The adaptive adjustment method for dance movement learning progress according to claim 1, characterized in that, The joint rotation sequence structure includes a rotation trajectory index, joint angle segments, and action timing segments; the rhythm transition positioning points include rhythm switching benchmarks, action jump node groups, and local rhythm triggering factors; the sequence structure center label includes sequence core nodes, local peak and valley anchor points, and structure alignment identifiers; the node offset parameters include offset amplitude factors, center difference identifiers, and paragraph alignment correction items; the progress rhythm node information includes rhythm node labels, progress association indexes, and learning sequence anchoring items.

3. The adaptive adjustment method for dance movement learning progress according to claim 1, characterized in that, The specific steps for obtaining the joint rotation sequence structure are as follows: S111: Based on the posture capture camera equipment around the training area, analyze the spatial coordinates and rotation orientation data of each frame of the collected joints, determine the correspondence between each joint and the trainee's body structure, and assign the joint parameters to different categories according to the set rules to obtain the joint type mapping group. S112: Based on the joint type mapping group, compare the intra-frame rotation orientation parameters of each type of joint, identify the associated parameters of the joint space rotation state according to the rotation direction set by the type of each joint, and collect the joint states by number to obtain the frame-level rotation state sequence. S113: Based on the frame-level rotation state sequence, calculate the rotation orientation change of each joint in consecutive frames, and combine the corresponding frame number to perform temporal arrangement and difference integration of the rotation state of the same joint to obtain the joint rotation sequence structure.

4. The adaptive adjustment method for dance movement learning progress according to claim 1, characterized in that, The specific steps for obtaining the rhythm transition positioning point are as follows: S211: Based on the joint rotation sequence structure, analyze the spatial rotation state of the same joint in consecutive frames, calculate the rotation direction angle of each pair of adjacent frames, arrange the rotation change data of each group in time order, and obtain the inter-frame rotation amplitude sequence. S212: Based on the inter-frame rotation amplitude sequence, compare the rotation change rate of each joint in consecutive frames, analyze the change amplitude of data points in the time series, retrieve the time periods in which continuous change or acceleration trends occur, and obtain the set of inter-frame change trends. S213: Based on the set of inter-frame change trends, determine the change direction and rotation amplitude fluctuation of each joint in the time series, identify the frame number in the sequence where the direction reverses and the rotation amplitude fluctuates, and mark them as key moments of rhythm jump in sequence to obtain rhythm transition positioning points.

5. The adaptive adjustment method for dance movement learning progress according to claim 1, characterized in that, The specific steps for obtaining the node offset parameter are as follows: S411: Analyze the frame number of the center index of the sequence structure and the rhythm transition positioning point on the time axis, and form a set of changes in consecutive numbers by calculating the sequential differences between the numbers, to obtain the frame number interval sequence; S412: Based on the frame numbering interval sequence, determine the relationship with the progress node fluctuation range limit. If the numbering span exceeds the limit, optimize the start and end numbers of the analysis segment. By adjusting the segment length, calculate the coupling characteristics of the rotation angle change and numbering interval within the compressed segment, obtain the rotation trend of the compressed numbering segment, and obtain the segment repositioning numbering reference. S413: Based on the segment relocation numbering benchmark, compare the position of the sequence structure center number in the sequence, analyze the numbering differences, filter the proportional characteristics of the differences and the span of the compressed segment, and obtain the node offset parameter.

6. The adaptive adjustment method for dance movement learning progress according to claim 1, characterized in that, The specific steps for obtaining the progress rhythm node information are as follows: S511: Based on the node offset parameter, analyze the correspondence between the nodes in the calibration frame number and the action timing trajectory, compare the mapping state formed by the paragraph alignment correction item and the center difference identifier, identify the frame number that meets the progress adjustment requirements, and obtain the number mapping reference set. S512: Based on the numbering mapping reference set, compare the node numbering order with the action time sequence trajectory, adjust the labeling position of the node in the time sequence trajectory, rearrange the node according to the adjusted time sequence structure, and obtain the node time sequence matching sequence. S513: Based on the node timing matching sequence, determine the matching status of the node number and the rhythm structure label in the progress feedback sequence, synchronously learn the identification content of the progress feedback system, and combine the linkage relationship between the node and the rhythm structure to obtain the progress rhythm node information.

7. The adaptive adjustment method for dance movement learning progress according to claim 1, characterized in that, The joint three-dimensional posture information refers to the coordinate position and rotation orientation of the joint in space along three directions as recorded by the posture capture device; the rotation angle change refers to the angle difference between consecutive frames of the same joint; the adjacent frame change rate refers to the trend of the speed of angle change between two consecutive frames.

8. A dance movement learning progress adaptive adjustment system, characterized in that, The system is used to implement the adaptive adjustment method for dance movement learning progress according to any one of claims 1-7, the system comprising: The rotation sequence module analyzes the three-dimensional pose information of the joints in each frame based on the pose capture camera equipment around the training area, identifies the rotation state of the joints in a single frame, compares the changes in joint position in time sequence, and associates the angle changes with the frame number to obtain the joint rotation sequence structure. The transition detection module calculates the rotation angle change of consecutive frames based on the joint rotation sequence structure, compares the change rate of adjacent frames, analyzes the angle change direction and amplitude fluctuation, determines whether there is a rhythm jump feature, identifies abnormal fluctuation frames, and obtains the rhythm jump location point. Based on the rhythm transition positioning point, the center identification module sorts the rotation angles of each frame in the analysis segment by number, compares the angle changes frame by frame, identifies frames with prominent changes or frames with gentle changes, and obtains the center label of the sequence structure. The offset correction module determines the distance between the center index of the sequence structure and the frame number of the rhythm transition positioning point, compares the distance with the node fluctuation range, and if it exceeds the range, adjusts the length of the analysis segment, reduces the frame number range, and corrects the center number to obtain the node offset parameter. The rhythm calibration module records the calibration frame number based on the node offset parameter, updates the node identifier in combination with the trainee's movement time trajectory, so that the node and the movement rhythm are in correspondence, and obtains the progress rhythm node information.

Images