A virtual guqin practice system

Abstract

The application discloses a virtual qin practice system and belongs to the technical field of virtual simulation training, and solves the technical problems of sensing error interference, lack of pertinence and individualization in existing virtual qin practice. The system comprises a data acquisition module, a data processing module and a feedback output module. The hand action sensing data is collected through a data glove, consistent deviation caused by equipment factors and posture deviation caused by user operation are extracted, the difference degree of a performance effect and a standard effect is combined, user types are divided according to posture deviation characteristics, consistent deviation compensation or compensation plus posture correction is executed according to the types, and individualized standard hand shape adaptation, dynamic adjustment of a judgment threshold, deviation pre-judgment and reverse optimization of collection sensitivity can also be realized. The feedback module synchronously outputs virtual images and sound feedback. The application improves sensing data accuracy, realizes individualized and real-time qin practice guidance, and effectively optimizes qin practice experience and practice efficiency.

Description

Technical Field

[0001] This invention relates to the field of musical instrument teaching technology, and in particular to a virtual guqin practice system. Background Technology

[0002] With the deep integration of virtual reality technology, smart wearable devices, and traditional musical instrument teaching, virtual guqin practice solutions based on smart sensing gloves have become an important supplement to guqin teaching and daily practice. Traditional physical guqin practice suffers from pain points such as high space requirements, expensive equipment, difficulty for beginners to obtain real-time posture feedback, and a lack of convenient and precise practice scenarios for professional performers. Virtual guqin practice devices, through built-in posture and pressure sensors, collect information such as hand joint movement angles and finger contact force, and combine this with image interaction technology to construct a virtual playing scene. Based on the motion data, corresponding sounds are generated, effectively solving many limitations of traditional practice modes and meeting the needs of different groups to practice anytime, anywhere. Currently, the core working logic of these devices focuses on collecting hand motion data through smart gloves and converting it into action commands to achieve interaction with the virtual guqin scene and sound output. However, the technological development of existing devices is overly focused on motion acquisition and scene rendering, with significant flaws in the design of the crucial position calibration mechanism. Specifically, these devices fail to effectively identify and separate systematic errors caused by the combined effects of data glove wearing position offset and inherent sensor characteristics, while ignoring individual differences in the user's actual posture deviation. The calibration scheme lacks the ability to dynamically adapt to the motion deviation characteristics of different user groups. For example, beginners are prone to developing incorrect playing habits due to uncorrected postural deviations, while professional performers struggle to achieve high-precision performance requirements because system errors are not adequately compensated. This imperfection in the calibration mechanism leads to a continuous gap between the motion sensing accuracy of the device in practical applications and user needs, severely restricting its promotion and application in diverse scenarios such as professional practice and beginner teaching. Summary of the Invention

[0003] The purpose of this application is to provide a virtual guqin practice system that can effectively identify and separate systematic errors caused by the offset of the data glove wearing position and the inherent characteristics of the sensor, while taking into account the individual differences in user posture deviation and providing a dynamically adapted calibration scheme, thereby improving the accuracy of motion sensing and meeting the practice needs of different user groups.

[0004] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a virtual guqin practice system, comprising a data acquisition module, a data processing module, and a feedback output module;

[0005] The data acquisition module is used to collect hand motion sensing data of the user in a virtual guqin performance scenario using a data glove worn on the user's hand. The data processing module is communicatively connected to the data acquisition module and includes a deviation extraction unit, a user type determination unit, a calibration execution unit, and a standard repertoire library. The standard repertoire library stores standard performance data for at least one target piece, and the standard performance data includes at least the standard note characteristics of the piece. The deviation extraction unit is used to identify and extract systematic errors that exhibit regularity in continuous movements from the hand motion sensing data, caused by the combined effects of the data glove's wearing position offset and the inherent characteristics of the sensors. This systematic error is defined as a consistency deviation. Simultaneously, portions of the hand motion sensing data not classified as consistency deviations but caused by deviations in the user's hand posture or movement trajectory from standard requirements are identified as posture deviations. The user type determination unit is communicatively connected to the deviation extraction unit and accesses the standard repertoire library to compare and analyze the degree of difference between the user's performance effect and the standard performance effect, and to combine this with the regularity characteristics of the posture deviations. Users are categorized into at least a first type of user or a second type of user. The first type of user's posture deviation has a lower impact on the performance than a preset level, while the second type of user's posture deviation has an impact on the performance higher than or equal to the preset level. The calibration execution unit is communicatively connected to the user type determination unit and the deviation extraction unit. When the user type determination unit determines the user to be the first type, the calibration execution unit is configured to: use a preset standard playing hand shape as a reference to compensate for the consistency deviation in the hand motion sensing data; when the user type determination unit determines the user to be the second type, the calibration execution unit is configured to: first compensate for the consistency deviation in the hand motion sensing data, then compare the compensated hand motion data with the standard playing hand shape to generate a correction instruction describing the posture deviation. The feedback output module is communicatively connected to the calibration execution unit and is used to generate and present corresponding virtual image and sound feedback based on the compensated hand motion data or the correction instruction output by the calibration execution unit.

[0006] The present invention is further configured such that: the deviation extraction unit is specifically used to: perform statistical analysis on the hand motion sensing data of continuous frames, identify and quantify error components that recur in the continuous frames and have a fixed direction or amplitude, so as to obtain the consistency deviation; the calibration execution unit has a pre-stored calibration parameter set, the parameter values ​​in the calibration parameter set are associated with different guqin playing fingering types, and the calibration execution unit is configured to: identify the current fingering type according to the hand motion sensing data, and call the calibration parameter value associated with the current fingering type to perform the compensation of the consistency deviation.

[0007] The present invention is further configured such that: the user type determination unit is specifically used to: calculate the matching degree in pitch and timing between the continuous note sequence generated by the user and the standard note sequence of the target piece after the compensation for the consistency deviation;

[0008] If the matching degree is higher than the preset threshold, the user is identified as the first type of user; if the matching degree is lower than or equal to the preset threshold, the user is identified as the second type of user; the process of the calibration execution unit compensating for the consistency deviation and generating the correction instruction when the user is identified as the second type of user is performed in real time during the user's playing.

[0009] The present invention is further configured such that when the user type determination unit determines that the user is the second type, the calibration execution unit is further configured to: acquire the user's individual hand physiological parameters through the data glove acquisition or manual input by the user; based on the individual hand physiological parameters, adaptively adjust the standard playing hand shape to generate a personalized standard hand shape; and generate a correction instruction describing the posture deviation by comparing the hand movement data after compensating for the consistency deviation with the personalized standard hand shape.

[0010] The present invention is further configured such that: the feedback output module includes a graphics rendering unit, which is used to indicate the position and direction of the posture deviation corresponding to the correction command in real time in a virtual scene by superimposing visual markers on the user's virtual hand model.

[0011] The present invention is further configured such that: the preset threshold for determining the user type in the user type determination unit is dynamically adjusted based on the user's historical playing data; the logic of this dynamic adjustment is as follows:

[0012] When the difference between the user's performance and the standard performance, as well as the regularity of the posture deviation, meet the criteria for the first type of user, the current preset threshold is maintained. When the difference between the user's performance and the standard performance, as well as the regularity of the posture deviation, shows a trend towards meeting the criteria for the second type of user, the preset threshold is adjusted by a preset adjustment step size until the difference between the user's performance and the standard performance, as well as the regularity of the posture deviation, re-stabilize and meet the criteria for either the first type of user or the second type of user.

[0013] The present invention is further configured such that: the deviation extraction unit is also configured to perform deviation trend analysis; the deviation trend analysis includes: in the time dimension, counting the frequency of the same fingering in different playing periods of the user; in the spatial dimension, counting the overlap range of the direction and amplitude of the consistency deviation in each playing session; determining whether the consistency deviation is an irregular fluctuation based on the above trend analysis results, wherein the basis for determining it as an irregular fluctuation is that the change in the direction or amplitude of the currently detected deviation does not follow the change pattern obtained from the statistical analysis of historical action data. If it is determined to be an irregular fluctuation, then this part of the error is attributed to posture deviation.

[0014] The present invention is further configured such that: when the deviation extraction unit extracts the consistency deviation, it simultaneously combines the continuity features of the hand movements for auxiliary determination; the continuity features include the connection state of the trajectories between adjacent movements and the stability of the time interval; when a deviation is detected accompanied by an abnormal connection state or an unstable time interval, the deviation is directly determined to be a posture deviation.

[0015] The present invention is further configured such that: the data processing module further includes a deviation prediction unit; the deviation prediction unit performs trend fitting based on the consistency deviation direction, deviation amplitude and action interval of consecutive identical fingering movements in the user's historical playing data, generates a prediction result for subsequent consistency deviations, and sends the prediction result to the calibration execution unit for performing consistency deviation pre-compensation on the subsequently collected hand movement sensing data in advance; when the difference between the actual consistency deviation calculated from the newly collected hand movement sensing data and the prediction result exceeds the preset tolerance range, the deviation prediction unit automatically triggers model parameter updates.

[0016] The present invention is further configured such that the virtual image and sound feedback output by the feedback output module are also used to reverse optimize the acquisition sensitivity of the data acquisition module; the specific logic of the reverse optimization is as follows: when the feedback indicates that the user's posture deviation after correction is still not lower than a preset deviation threshold, the acquisition sensitivity of the data acquisition module for the key movement nodes corresponding to the posture deviation is increased; when the feedback indicates that the user's posture deviation after correction is lower than the preset deviation threshold, i.e., the correction is effective, the current acquisition sensitivity is maintained, and the correlation between the acquisition sensitivity parameter and the corresponding effective movement feature is recorded.

[0017] The present invention has significant technical effects due to the adoption of the above technical solutions: the data acquisition module collects hand movement data, the data processing module identifies consistency deviation and posture deviation, classifies user types, performs compensation according to type, and the feedback output module provides feedback, thereby effectively separating system errors and providing calibration for different users, improving the accuracy of motion sensing, realizing dynamic adaptation calibration, and meeting the practice needs of different user groups. Attached Figure Description

[0018] Figure 1 This is a structural diagram of a virtual guqin practice system. Detailed Implementation

[0019] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0020] Example: Traditional smart glove-style virtual guqin practice devices have made progress in motion capture and scene rendering, but they are imperfect in the design of the key position calibration mechanism. The device has not yet formed a calibration scheme that accurately adapts to the motion deviation characteristics of different user groups, resulting in a gap between the motion sensing accuracy of the device in actual application and user needs, which limits its promotion and application in professional guqin practice and novice teaching scenarios.

[0021] This application proposes a virtual guqin practice system, including a data acquisition module, a data processing module, and a feedback output module. The system acquires user hand movement sensor data through the data acquisition module, performs deviation extraction, user type judgment, and calibration execution on the data through the data processing module, and compares it with a standard repertoire library. The system can identify and distinguish between consistency deviation and posture deviation, execute different calibration strategies according to user type, and finally provide virtual image and sound feedback through the feedback output module.

[0022] This virtual guqin practice system is designed with a modular structure, mainly consisting of a data acquisition module, a data processing module, and a feedback output module. The data acquisition module is responsible for acquiring the user's raw motion information in the virtual performance environment; the data processing module analyzes, calibrates, and makes decisions based on this raw information; and the feedback output module presents the processing results to the user in an intuitive way.

[0023] The data acquisition module acquires hand motion sensing data by wearing a data glove on the user's hand. The data glove can have multiple small sensors built in, such as a flexion sensor at each finger joint to measure the flexion angle, and an inertial measurement unit (IMU) on the back of the palm to acquire the overall posture and motion information of the hand. These sensors convert the acquired analog signals into digital signals and transmit them to the data processing module via wired or wireless means.

[0024] The data processing module is configured to receive hand motion sensing data from the data acquisition module and perform in-depth analysis and processing. The module integrates a deviation extraction unit, a user type judgment unit, a calibration execution unit, and a standard repertoire library. The standard repertoire library pre-stores standard performance data of multiple target repertoires. These data contain at least the standard note characteristics of the repertoire as a benchmark for evaluating the user's performance effect.

[0025] The deviation extraction unit is responsible for identifying and distinguishing two types of deviations from the received hand motion sensing data. Consistency deviations may be caused by slight offsets in the position of the data glove on the user's hand, or inherent zero-point drift or linearity errors in the sensor itself. This unit can perform statistical analysis on the continuous motion data of the user under a specific preset motion sequence. For example, when the user performs a series of repetitive finger exercises, it can observe that some sensor data have a continuous and regular deviation in a specific direction or amplitude, and thus identify these regular deviations as consistency deviations. Posture deviations refer to the part of the hand motion sensing data that is not classified as consistency deviations. This deviation is caused by the user's hand posture, finger bending angle, or movement trajectory deviating from the standard requirements of guqin performance during the performance.

[0026] The user type determination unit is configured to communicate with the deviation extraction unit and access the standard repertoire library. This unit first acquires hand movement data after the deviation extraction unit has compensated for consistency deviations, and then uses this data to generate the actual note features played by the user, representing the user's performance effect. Simultaneously, the unit retrieves the standard note features of the current target piece from the standard repertoire library, representing the standard performance effect. Subsequently, the unit evaluates the degree of difference between the user's performance effect and the standard performance effect, for example, by calculating the matching degree of pitch, timing, and dynamics. Based on this, the unit combines the regularity characteristics of posture deviations to classify users into either a first-type user or a second-type user. Specifically, when the impact of the user's posture deviation on the performance effect is lower than a preset level, the user is classified as a first-type user; conversely, when the impact of the posture deviation on the performance effect is higher than or equal to the preset level, the user is classified as a second-type user. For second-type users, the impact of posture deviations during performance on the final performance effect is higher than or equal to the preset level.

[0027] The calibration execution unit is configured to communicate with the user type determination unit and the deviation extraction unit. When the user type determination unit determines the user to be a first-type user, it indicates that the user's main problem lies in the consistency deviation caused by the device itself. In this case, the calibration execution unit uses a preset standard playing hand shape as a reference to compensate for the consistency deviation in the hand motion sensing data collected by the data acquisition module. This compensation process can use pre-calibrated compensation parameters, for example, through simple addition and subtraction operations or linear transformations, to correct the systematic offset in the sensor data. When the user type determination unit determines the user to be a second-type user, it indicates that the user, in addition to... Besides potential inconsistencies, posture deviations also affect performance. In this case, the calibration execution unit first compensates for inconsistencies in the hand motion sensor data acquired by the data acquisition module to eliminate system errors at the device level. Then, the unit compares the compensated hand motion data with a preset standard playing hand shape. This comparison can involve multiple dimensions, such as comparing the bending angle of the fingers, the tilt of the wrist, and the curvature of the palm. Through this comparison, the calibration execution unit can identify the specific differences between the user's actual hand shape and the standard hand shape, and generate correction instructions describing the posture deviation based on these differences.

[0028] The feedback output module is configured to communicate with the calibration execution unit and generate and present corresponding virtual image and sound feedback based on the compensated hand movement data or correction instructions output by the calibration execution unit. When the user is identified as a first-type user, the feedback output module mainly drives the virtual hand model to perform movement simulation in the virtual scene based on the compensated hand movement data and triggers the corresponding virtual guqin timbre. When the user is identified as a second-type user, in addition to providing basic virtual performance feedback, the feedback output module will also indicate the deviation position and correction direction of the user's hand posture in the virtual image through visual cues, such as highlighting, arrow indication, or superimposing virtual skeletons, based on the correction instructions.

[0029] This system addresses the shortcomings of existing virtual guqin practice devices by providing a refined position calibration mechanism. By accurately distinguishing between consistency deviations caused by the device and posture deviations caused by the user, and implementing targeted calibration strategies based on user type, the system can provide device error compensation for experienced users and posture correction guidance for beginners or users with posture problems. As a result, the system improves the accuracy of motion sensing and meets the needs of users in professional practice and novice teaching scenarios.

[0030] In some of the embodiments described above in this application, although a virtual guqin practice system for identifying and compensating for consistency deviations and posture deviations has been proposed, there is still room for further optimization regarding the specific method for extracting consistency deviations and how to perform refined compensation based on the complexity of guqin playing. If consistency deviations are not accurately and efficiently identified and compensated, it may affect the accuracy of subsequent posture deviation judgments, thereby reducing the effectiveness of system feedback.

[0031] This application further proposes a method for optimizing consistency deviation extraction and compensation to improve the accuracy of the system's processing of user hand motion data. Specifically, the deviation extraction unit is used to perform statistical analysis on continuous frames of hand motion sensing data, identify and quantify error components that recur in the continuous frames and have a fixed direction or amplitude, in order to obtain the consistency deviation. By performing statistical analysis on continuous frames of hand motion sensing data, the deviation extraction unit can effectively distinguish systematic errors that exhibit regularity in continuous movements, i.e., consistency deviations, caused by the offset of the data glove wearing position and the inherent characteristics of the sensors. This statistical analysis can be implemented using various methods. For example, by calculating the mean, variance, or trend of specific sensor readings in continuous frames, error patterns that persist in different performance movements and have a relatively fixed direction or amplitude can be identified. For example, sliding window averaging, Kalman filtering, or machine learning-based pattern recognition algorithms can be used to analyze time series data, thereby accurately quantifying these recurring error components. In this way, the system can effectively separate the inherent bias of the sensor itself from the actual posture deviation of the user, ensuring the accuracy of subsequent processing.

[0032] To further improve the accuracy of compensation, the calibration execution unit pre-stores a set of calibration parameters. The parameter values ​​in this set are associated with different guqin playing fingering types. The calibration execution unit is configured to identify the current fingering type based on hand movement sensor data and then call upon the calibration parameter values ​​associated with that fingering type to perform consistency deviation compensation. The pre-stored set of calibration parameters in the calibration execution unit, with parameter values ​​associated with different guqin playing fingering types, aims to provide refined compensation for the diverse fingering techniques used in guqin playing, such as hooking, plucking, stroking, picking, and tremolo. Different finger techniques involve different movement trajectories and force patterns of the hand and fingers. This can lead to different types or degrees of consistency deviations in the data glove under different finger techniques. The calibration parameter set can be a pre-built database or model containing compensation coefficients, offsets, or transformation matrices optimized for each finger technique type. When the calibration execution unit receives hand motion sensing data, it first uses the built-in finger technique recognition module to determine the type of finger technique the user is currently performing. The finger technique recognition module uses, for example, a pre-trained classifier, and then retrieves and applies parameters from the calibration parameter set that correspond to that finger technique type. The system uses the most suitable calibration parameter values ​​to compensate for consistency deviations. This adaptive compensation mechanism based on fingering type significantly improves the accuracy and applicability of compensation. Through the above technical solution, the system can more accurately identify and quantify consistency deviations caused by the data glove itself in the virtual guqin practice system. Specifically, by statistically analyzing continuous frame hand motion sensing data, the system can effectively distinguish between regular system errors and actual user posture deviations, avoiding misjudging inherent sensor errors as user technical defects. On this basis, by pre-storing a set of calibration parameters associated with different guqin playing fingering types in the calibration execution unit, and calling the corresponding calibration parameter values ​​for compensation according to the currently identified fingering type, the system can provide more refined and personalized consistency deviation compensation for the complex and varied fingering characteristics in guqin playing. This fingering type-related compensation mechanism ensures that hand motion sensing data can be accurately calibrated in different performance scenarios, thus providing more reliable and accurate basic data for subsequent user type judgment, posture deviation analysis, and correction command generation, significantly improving the effectiveness of feedback and user experience of the virtual guqin practice system.

[0033] In some of the embodiments described above in this application, although users are classified into first-type users or second-type users and differentiated according to the degree of influence of their posture deviation on the performance effect, and compensation is provided for consistency deviation, no clear real-time mechanism is provided for how to accurately quantify the "degree of difference" between the user's performance effect and the standard performance effect, and when to generate compensation and correction instructions for consistency deviation. This may result in the system's evaluation of user performance being inaccurate and the feedback being untimely in a dynamic practice environment, thereby affecting the efficiency and experience of user learning.

[0034] In response, this application further proposes an improved scheme, wherein the user type determination unit is specifically used to: calculate the matching degree of the continuous note sequence generated by the user and the standard note sequence of the target piece in terms of pitch and timing after compensating for consistency deviation; if the matching degree is higher than a preset threshold, the user is determined to be a first type of user; if the matching degree is lower than or equal to the preset threshold, the user is determined to be a second type of user; the process of the calibration execution unit compensating for consistency deviation and generating correction instructions when the user is determined to be a second type of user is performed in real time during the user's playing.

[0035] Specifically, this technical solution aims to provide a precise method for quantifying the degree of difference between a user's performance and a standard performance. After the data processing module receives hand movement data that has been compensated for by the calibration execution unit to reduce consistency deviations, the user type determination unit identifies and extracts the actual note sequence produced by the user from this data, including information such as the pitch, start time, and duration of each note. At the same time, the user type determination unit obtains the standard note sequence of the current target piece from the standard repertoire library. Subsequently, the system uses a sequence comparison algorithm, such as Dynamic Time Warping (DTW) or a feature point matching algorithm, to compare the user's note sequence with the standard note sequence. In terms of pitch, the pitch difference of the corresponding notes can be calculated; in terms of temporal sequence, the start time deviation and duration deviation of the notes can be calculated. By comprehensively considering the differences in these dimensions, a quantitative matching score can be obtained, which can objectively reflect the accuracy of the user's performance in terms of pitch and rhythm.

[0036] Based on this quantitative matching degree, the system further determines the user type. After calculating the matching degree between the user's performance and the standard performance, the user type determination unit compares it with a preset threshold. This preset threshold is set by the system based on teaching experience or a large amount of user data to distinguish users at different learning stages or skill levels. If the matching degree is higher than the threshold, it indicates that the user's performance in pitch and timing is good. Even if there are posture deviations, their impact on the final performance effect is relatively small. Therefore, the user is classified as a first-type user. Conversely, if the matching degree is lower than or equal to the threshold, it indicates that the user has significant problems in pitch or timing. This usually means that posture deviations have a significant impact on the performance effect. Therefore, the user is classified as a second-type user. This judgment method based on quantitative indicators makes user classification more objective and accurate.

[0037] This solution also emphasizes the real-time nature of the system's core processing. This means that from data acquisition and consistency deviation compensation to user type identification and the generation of posture correction instructions for the second type of user, all these steps occur while the user is playing the guqin, rather than being analyzed offline after the performance ends. To achieve real-time performance, the system needs efficient data processing capabilities and low-latency algorithms. For example, the data acquisition module can continuously acquire data at a high frame rate, while the data processing module adopts a streaming architecture to ensure that each processing step can be completed in a very short time. This real-time processing capability ensures that the feedback output module can provide users with immediate visual and auditory feedback, allowing users to immediately perceive and correct problems in their performance, thereby significantly improving practice efficiency and learning experience. This effectively solves the problems of inaccurate evaluation of user performance and untimely feedback in virtual guqin practice systems. Specifically, by calculating the matching degree in pitch and timing between the user's continuous note sequence after consistency deviation compensation and the standard note sequence of the target piece, the system can objectively and quantitatively evaluate the user's performance. This precise matching calculation allows the user type judgment unit to more accurately classify users into either the first or second type, thus avoiding misjudgments caused by fuzzy evaluation. More importantly, the calibration execution unit's compensation for consistency deviations and the generation of correction instructions when the user is judged to be a second type are performed in real time during the user's playing. This means that during the user's playing, the system can instantly identify and compensate for system errors caused by the data glove, and judge the user's technical level and problems in real time based on the user's current performance. For second-type users who are not performing well, the system can immediately generate correction instructions describing posture deviations and present them to the user in real time through the feedback output module. This real-time and immediate feedback mechanism allows users to receive guidance and make adjustments as soon as their mistakes occur, greatly shortening the learning cycle and improving the pertinence and effectiveness of practice. Compared with the traditional method of analysis and feedback after the playing is over, the real-time processing capability of this application significantly improves the user's learning efficiency and immersion, enabling users to master guqin playing techniques more effectively.

[0038] When a user is identified as a second-type user, the system generates a posture deviation correction instruction by comparing the hand movement data after compensating for consistency deviation with a preset standard playing hand shape. However, due to differences in individual hand physiological characteristics, such as palm size, finger length, and joint flexibility, a uniform standard playing hand shape may not be suitable for all users. This may result in the posture deviation correction instruction generated by the system being difficult for users to execute effectively, thereby affecting practice results and user experience.

[0039] Furthermore, when the user type judgment unit determines that the user is a second type of user, the calibration execution unit is also used to obtain the user's individual hand physiological parameters; based on the individual hand physiological parameters, the standard playing hand shape is adaptively adjusted to generate a personalized standard hand shape; the correction instruction describing the posture deviation is generated by comparing the hand movement data after compensating for consistency deviation with the personalized standard hand shape.

[0040] Specifically, in acquiring individual hand physiological parameters, this step aims to collect data reflecting the unique physical characteristics of the user's hands. These parameters are crucial for subsequently generating personalized standard hand shapes that conform to the user's actual situation. This can be achieved by guiding the user to perform a series of hand measurements, such as palm width, finger length, and knuckle circumference. These measurement data can be manually input by the user, automatically collected using external measurement devices such as 3D scanners and depth cameras, or inferred from the initial calibration data of data gloves in a specific posture. In addition, basic information such as the user's age and gender can be combined for auxiliary judgment to obtain more comprehensive individual physiological characteristics.

[0041] The core of this solution lies in the adaptive adjustment of the standard playing hand shape based on individual hand physiological parameters to generate a personalized standard hand shape. This step utilizes the acquired individual physiological parameters of the user to customize and modify the preset general standard playing hand shape, making it more suitable for the user's actual hand conditions. This results in a more reasonable and easily achievable personalized standard hand shape for the user. This can be achieved through parametric model adjustment, where the standard playing hand shape is constructed as a parametric three-dimensional hand model. By inputting the user's individual hand physiological parameters into this model, such as finger length ratio and joint range of motion, the system can automatically adjust the model's bone length, joint position, and rotational limitations to match its geometry and kinematic characteristics with the user's hand features. Alternatively, a machine learning model can be trained by inputting the user's physiological parameters and a large amount of standard playing data from users with different hand shapes to output an optimized personalized standard hand shape.

[0042] In generating correction instructions describing posture deviations, specifically by comparing hand movement data after compensating for consistency deviations with a personalized standard hand shape, this step clarifies the method of generating correction instructions. By comparing the user's actual hand movement data, which has already undergone systematic error compensation, with a personalized standard hand shape tailored to the user, specific deviations in the user's posture can be identified more accurately and effectively, generating more instructive correction instructions. This can be achieved using a three-dimensional posture matching algorithm, comparing the user's real-time hand movement data, such as joint angles and positions, with the corresponding joint data of the personalized standard hand shape, calculating the difference vector between the two. This vector can represent the positional deviation, rotational angle deviation, etc., of each joint in three-dimensional space. Subsequently, this numerical deviation information is converted into correction instructions that are easy for the user to understand, such as the index finger should bend a few more degrees, the hand... The system provides information such as the wrist's inward rotation and the wrist's height, and presents these parameters to the user in real-time through visual or auditory feedback. This allows the system to acquire the user's individual hand physiological parameters and adaptively adjust the preset standard playing hand shape based on these parameters. This generates a unique, personalized standard hand shape for each user. When the system generates posture correction instructions for a second type of user, it compares the instruction to that user's personalized standard hand shape, rather than a generic standard hand shape. This customized comparison method makes the generated correction instructions more accurate and better suited to the user's actual hand physiological conditions and motor abilities. This significantly improves the effectiveness and operability of posture correction, avoiding ineffective or uncomfortable practice due to hand shape differences. Furthermore, it helps users master the correct guqin playing posture more efficiently, thereby improving learning efficiency and user experience.

[0043] In some of the embodiments described above in this application, the feedback output module generates and presents corresponding virtual images and sound feedback based on the correction instructions output by the calibration execution unit. However, in practical applications, if only abstract correction instructions are provided, users may find it difficult to intuitively understand the specific deviation position and movement direction of their hand posture, thereby affecting the correction efficiency and learning experience.

[0044] To address this, this application proposes an improved feedback output method. The feedback output module includes a graphics rendering unit, which uses visual markers superimposed on the user's virtual hand model in a virtual scene to indicate the position and direction of the posture deviation corresponding to the correction command in real time. The graphics rendering unit is the core component of the feedback output module, responsible for converting abstract correction commands into visual information perceptible to the user. Its function is to construct and present three-dimensional images in the virtual environment based on the data processed by the system, including a virtual guqin, a user's virtual hand model, and various visual cues. This unit typically includes a graphics processor, a rendering engine, and related graphics libraries, which can efficiently handle complex graphics calculations and ensure the smoothness and realism of the virtual scene. The virtual scene refers to the interactive three-dimensional environment constructed and presented by the graphics rendering unit. This virtual scene typically simulates a real guqin performance environment and includes elements such as a virtual guqin, strings, and a user's virtual hand model. Users immerse themselves in it by wearing VR virtual reality headsets or AR augmented reality devices and interact with the virtual environment. This immersive experience helps users perceive and understand the feedback information provided by the system more intuitively, enhancing the sense of immersion and effectiveness of practice.

[0045] Visual markers superimposed on the user's virtual hand model refer to the visual elements used by the graphics rendering unit to directly present on the virtual representation of the user's hand in the virtual scene, indicating posture deviations. The user's virtual hand model is constructed or driven in real time based on hand motion sensing data collected by the data acquisition module, and can accurately reflect the user's actual hand posture. Visual markers can take the form of color changes, arrows, highlighted areas, transparent overlays, or dynamic lighting effects, etc. They are precisely mapped onto the virtual hand model to highlight specific fingers, joints, or palm areas that need adjustment.

[0046] The real-time indication of the posture deviation position and movement direction corresponding to the correction command means that the graphics rendering unit can instantly transform the correction command generated by the calibration execution unit into intuitive visual cues and immediately display them on the virtual hand model. The posture deviation position refers to the specific part of the user's hand that differs from the standard playing hand shape, and the movement direction refers to the direction that the deviation part needs to be adjusted, such as upward, downward, inward, or outward rotation. This real-time nature ensures that the user can immediately perceive and correct the error, avoiding the long-term solidification of incorrect posture. The graphics rendering unit can transform the abstract correction command generated by the calibration execution unit into intuitive and specific visual markers and overlay them on the user's virtual hand model in real time. This feedback method allows the user to clearly identify the specific deviation position of their hand posture and the movement direction that needs to be adjusted, thereby avoiding the understanding difficulties brought about by traditional abstract commands. Users do not need to do complex thinking or guessing; they can adjust their hand posture instantly based on visual cues, significantly improving the efficiency and accuracy of posture correction. In addition, providing this immersive and visual feedback in the virtual scene greatly enhances the user's learning experience and sense of immersion, enabling them to more effectively self-correct and improve their skills.

[0047] In some of the embodiments described above in this application, the user type determination unit classifies users into a first type of user or a second type of user based on a preset determination criterion. However, if the determination criterion is fixed, the determination criterion in this embodiment, i.e., the preset threshold, may not accurately adapt to the actual performance of users at different learning stages or in different playing states. For example, when a user's skills improve, a fixed threshold may cause the system to become overly concerned with minor posture deviations, while when a user is in poor condition, a fixed threshold may not be able to identify situations where posture correction is needed in a timely manner, thereby affecting the accuracy of feedback and the effectiveness of teaching.

[0048] In response, this application further proposes a preset threshold for determining user type in the user type determination unit, which is dynamically adjusted based on the user's historical playing data. The logic of this dynamic adjustment is as follows: when the deviation characteristics of the user's consecutive playing meet the determination conditions of the first type of user, the current preset threshold is maintained; when the user's deviation characteristics show a trend of changing to meet the determination conditions of the second type of user, the preset threshold is adjusted according to the preset adjustment step size until the user's deviation characteristics stabilize again to meet the determination conditions of the first type of user or the second type of user.

[0049] To enable dynamic adjustments based on the user's historical playing data, the system continuously records various data for each performance, including hand motion sensor data, consistency deviations and posture deviations extracted by the deviation extraction unit, as well as the matching degree between note features and standard note features and the final performance effect. This historical data is stored and used to analyze the user's long-term performance trends.

[0050] In the dynamic adjustment logic, when the deviation characteristics of a user's consecutive playing meet the judgment criteria of the first type of user, the system will track the deviation characteristics of the user's most recent N playing sessions. If the impact of posture deviation on the performance effect in these N playing sessions is lower than the current preset threshold, that is, the user is continuously judged as the first type of user, the preset threshold remains unchanged. This indicates that the user has stably mastered the basic skills and there is no need to lower the judgment standard, thereby maintaining the consistency of feedback.

[0051] When a user's deviation characteristics show a trend towards meeting the criteria for a second-type user, the system detects that during continuous playing, the impact of the user's posture deviation on the performance begins to approach or exceed the current preset threshold, or that the number of times the user is identified as a second-type user increases across multiple playing sessions. The system then identifies this trend and gradually adjusts the preset threshold with a preset adjustment step size. This adjustment can either raise the threshold to make it easier to be identified as a first-type user, encouraging the user, or lower the threshold to make it easier to be identified as a second-type user, prompting the system to provide more posture correction. The adjustment continues until the user's deviation characteristics stably meet the criteria for a certain type of user again, meaning the user's performance stabilizes above or below the new threshold. This logic allows the system to adjust the judgment criteria in a timely manner when the user's performance declines or encounters a bottleneck, in order to more accurately identify the user's current state and provide more targeted guidance.

[0052] Through the above technical solutions, this system can more accurately reflect the user's actual learning progress and performance level. When the user's performance is stable and good, the preset threshold remains unchanged, avoiding unnecessary interference. When the user's performance fluctuates or regresses, the preset threshold can be adaptively adjusted, allowing the user type judgment unit to promptly identify changes in the user's state. This ensures that the calibration execution unit can provide feedback that better meets the user's current needs. For example, when the user's performance declines, the system can more sensitively identify them as a second-type user, thereby triggering posture deviation correction instructions to help the user correct bad habits in time. Conversely, when the user improves, the preset threshold may be adjusted accordingly, allowing the system to identify the user as a first-type user earlier, reducing unnecessary posture correction and focusing instead on compensating for consistency deviations, thereby improving learning efficiency and user experience. This dynamic adjustment mechanism avoids the problems of misjudgment and untimely feedback that may be caused by fixed thresholds, enabling the virtual guqin practice system to provide more personalized, intelligent, and effective teaching assistance.

[0053] In some of the embodiments described above in this application, the deviation extraction unit identifies and quantifies consistency deviations by statistically analyzing hand motion sensing data from consecutive frames, and the calibration execution unit calls the corresponding calibration parameters for compensation based on the fingering type. However, in actual guqin practice, the deviation characteristics of the user's hand movements may not be constant. Simply identifying and compensating for the current consistency deviation may not be effective in dealing with changes in deviation over time or in the playing context, thus affecting the stability and accuracy of the calibration effect.

[0054] In response, this application further proposes that the deviation extraction unit is configured to also perform deviation trend analysis, which includes: in the time dimension, counting the frequency of the same fingering in different playing periods of the user; in the spatial dimension, counting the overlap range of the direction and magnitude of consistency deviation in each playing session; wherein, the basis for determining irregular fluctuation is that the change in the direction or magnitude of the currently detected deviation does not follow the change pattern obtained from the statistical analysis of historical action data.

[0055] Deviation trend analysis aims to systematically examine the evolution and behavioral patterns of identified consistency deviations over multiple instances or time periods. It goes beyond single-point-of-time measurements or short-term statistical averages, striving to understand the dynamic characteristics of deviations and thus providing a basis for more adaptive and robust compensation strategies. In practice, the system collects and stores historical data related to consistency deviations, including consistency deviation values ​​such as direction and amplitude identified for specific finger techniques in different practice periods.

[0056] To understand the frequency of use of specific fingering techniques and the performance of their associated consistency deviations across different practice sessions, the system maintains a log or database for each user, recording the fingering type, timestamp, and consistency deviations such as direction and amplitude for each fingering execution. During trend analysis, the system queries this log to count the number of times a specific fingering technique appears within a predefined time window, such as daily, weekly, or per session. This frequency information can be used to assess the reliability and representativeness of the deviation data. For example, if a fingering technique is rarely used, its deviation data may not be as robust as the deviation data of commonly used fingering techniques.

[0057] To quantify the variability or stability of consistency deviation itself, the system stores the consistency deviation vector extracted for each fingering in each performance. The overlap range can be calculated by determining the minimum and maximum values ​​of each component of the X, Y, Z direction or amplitude in all historical instances, or by calculating statistics such as the standard deviation or variance of these components. A smaller range or standard deviation indicates higher consistency. This spatial analysis helps to define confidence intervals or expected ranges for consistency deviation.

[0058] To distinguish between predictable, slowly evolving systematic errors and truly unstable, unpredictable errors, the system uses historical data to build a model of the "pattern" of consistency deviation. This can be achieved through various statistical or machine learning techniques, such as regression analysis, time series analysis, statistical process control, charts, or training machine learning models to predict the expected range or pattern of deviation. When a new consistency deviation is detected, its direction and magnitude are compared with the predicted or expected pattern. If the new deviation exceeds the statistically defined confidence interval, or if its change significantly deviates from the learned trend of previous measurements, it is marked as an irregular fluctuation. This may trigger further investigation, prompt the user to recalibrate, or temporarily revert to a more conservative compensation method.

[0059] By introducing deviation trend analysis, the system no longer relies solely on identifying consistency deviations at a single moment or in a short period. Instead, it gains a deeper understanding of the evolution of these deviations over time and their stability range in space. Specifically, by statistically analyzing the frequency of the same fingering technique in different playing periods, the system can assess the reliability and representativeness of the deviation data, avoiding inaccurate compensation due to data sparsity or randomness. Simultaneously, by statistically analyzing the overlap range of the direction and amplitude of consistency deviations in past playing sessions, the system can quantify the stability of the deviations, thereby establishing a more accurate deviation model. When the system detects that the change in the direction or amplitude of the current deviation does not follow historical patterns, it can promptly identify irregular fluctuations, avoiding misjudging abnormal or unstable deviations as regular consistency deviations for compensation. This significantly improves the accuracy, stability, and adaptability of deviation compensation, ensuring that the virtual guqin practice system can provide users with more reliable and personalized feedback, especially in long-term use and different practice scenarios.

[0060] In some of the embodiments described above in this application, the deviation extraction unit identifies and quantifies consistency deviations by statistically analyzing hand motion sensing data from consecutive frames and combining it with deviation trend analysis. However, in actual guqin practice, the complexity and diversity of user hand movements may lead to confusion between certain posture deviations and consistency deviations in their manifestation. For example, some non-standard movements caused by the user's lack of fingering skills or unstable rhythm may exhibit certain regularities in a short period of time, thereby affecting the accurate identification of consistency deviations by the deviation extraction unit, and consequently affecting the subsequent calibration and feedback effects, making it difficult for the system to accurately distinguish whether it is a device error or a problem with the user's playing skills.

[0061] In response, this application further proposes that when extracting consistency deviations, the deviation extraction unit simultaneously combines the continuity features of hand movements for auxiliary judgment; the continuity features include the connection state of trajectories between adjacent movements and the stability of time intervals; when a deviation is detected accompanied by abnormal connection state or unstable time intervals, the deviation is determined to be a posture deviation.

[0062] Continuity is a key indicator for measuring the smoothness and rhythm of a user's hand movements. The connection state of the trajectory refers to the smoothness and natural transition of the hand or finger movement path between one guqin playing action and the next. For example, when changing finger techniques, whether the trajectory of lifting, moving, and falling the fingers is continuous, without abrupt changes or unnecessary shaking can be quantified by calculating the smoothness of angle changes, rate of curvature changes, or speed and acceleration between adjacent movement trajectory segments. If the connection state is abnormal, it may manifest as trajectory interruption, abrupt changes, or unnecessary pauses. The stability of the time interval refers to whether the time interval between consecutive playing actions conforms to the expected rhythm or beat. In guqin playing, the time interval between notes is crucial to the accuracy of the rhythm. This can be quantified by calculating the difference between the timestamps of consecutive actions and evaluating the variance, standard deviation, or deviation from the standard rhythmic pattern of these differences. If the time interval is unstable, it may manifest as the rhythm being fast or slow, or pauses being too long or too short. This can be achieved by performing Fourier transforms, wavelet analysis, or using machine learning models based on hidden Markov models or recurrent neural networks to identify these patterns.

[0063] When a deviation is detected accompanied by abnormal connection states or unstable time intervals, it is determined to be a posture deviation. This is a key priority rule. When the deviation extraction unit identifies consistency deviations, if it simultaneously observes abnormal user hand movements such as disjointed or uneven trajectory connections or unstable time intervals between movements such as rhythmic instability, even if the deviation may show a certain regularity statistically, the system will prioritize classifying it as a posture deviation. This means that the system considers the user's performance technique problem to be the main cause of this deviation, rather than the systematic error of the data glove. This judgment mechanism ensures accurate identification of user performance problems, thereby providing more targeted feedback and guidance. In specific implementation, a conditional judgment branch can be introduced into the deviation classification algorithm. When the continuity abnormality condition is met, the classification result of the posture deviation is directly output without further complex analysis of consistency deviations.

[0064] By combining the continuity features of hand movements with auxiliary judgment during the extraction of consistency deviations, this application can more accurately distinguish between consistency deviations caused by the data glove itself and posture deviations caused by insufficient user playing skills. Specifically, when the deviation extraction unit detects abnormal trajectory connection or unstable time intervals of hand movements, even if the deviation may show a certain regularity statistically, the system will prioritize judging it as a posture deviation. This mechanism effectively avoids misjudging playing deviations caused by user's lack of fingering skills or unstable rhythm as equipment system errors, thereby ensuring that the calibration execution unit can specifically compensate for consistency deviations and generate more accurate correction instructions for posture deviations. This not only improves the system's recognition accuracy of user playing problems, but also enables the feedback output module to provide more instructive virtual image and sound feedback, thereby significantly improving the efficiency and effectiveness of user practice.

[0065] In some of the embodiments described above in this application, the system can identify and compensate for consistency deviations in the user's hand movements to improve the accuracy of virtual guqin practice. However, such compensation is usually based on real-time processing of data collected currently or recently. The system lacks the ability to proactively process dynamic changes or potential trends in consistency deviations, which may affect the timeliness and accuracy of the compensation.

[0066] In response, this application further proposes that the data processing module also includes a deviation prediction unit. This deviation prediction unit is a module specifically designed to predict future consistency deviations. Its core function is to learn the patterns and trends of deviations by analyzing historical data, thereby providing a preliminary estimate before actual deviations occur. Specifically, the deviation prediction unit performs trend fitting based on the direction, magnitude, and interval of consistency deviations for consecutive identical fingering movements in the user's historical playing data, generating a prediction result for subsequent consistency deviations. The user's historical playing data is usually stored in the system's database, containing hand motion sensor data collected during each playing session, consistency deviations identified and quantified by the deviation extraction unit, and the corresponding guqin fingering types and the interval time between consecutive movements. The analysis of consecutive identical fingering movements aims to capture the cumulative or periodic trends of consistency deviations that may occur when repeatedly performing specific fingerings due to factors such as slight slippage of the data glove, sensor drift, or user fatigue. The direction of the consistency deviation can be a vector in three-dimensional space, representing the direction of the deviation; the magnitude of the deviation represents the size of the deviation; and the interval between movements is the time length between two consecutive identical fingering movements.

[0067] Trend fitting is a key technical means for deviation prediction units to achieve prediction. This can be achieved through various statistical or machine learning methods. For example, time series analysis models such as the Autoregressive Moving Average (ARIMA) model and the exponential smoothing method can be used to establish a model of the relationship between deviation and time or the number of actions. Regression analysis such as linear regression and multinomial regression can also be used to fit the changing trend of deviation. For more complex nonlinear trends, even deep learning methods such as neural networks can be used for modeling. Through these methods, deviation prediction units can learn the inherent laws of consistency deviation changing with time or action sequences from historical data, and generate a predicted value of consistency deviation for a future point in time or a specific action.

[0068] When the difference between the actual consistency deviation calculated from the newly acquired hand motion sensing data and the predicted result exceeds the preset tolerance range, the deviation prediction unit automatically triggers model parameter updates. The actual consistency deviation is calculated in real time by the deviation extraction unit. The preset tolerance range is a configurable threshold, for example, it can be set to an angle of less than 5 degrees in the deviation direction or a relative error of less than 10% in the deviation magnitude. When the vector distance, angle, or magnitude difference between the actual deviation and the predicted result exceeds this preset tolerance range, the system will automatically start the model retraining or parameter adjustment process. This update can be incremental online learning, that is, incorporating the latest actual deviation data into the historical dataset for minor adjustments; or it can be periodic offline batch updates to make the model better adapt to the user's current state, changes in playing habits, and the long-term use characteristics of the data gloves.

[0069] Through the above technical solution, this application can perform trend fitting on consistency deviation based on the user's historical playing data, thereby generating a prediction result for subsequent deviations. This allows the system to predict deviations before they actually occur, thus achieving more timely and accurate compensation. When there is a significant difference between the actual detected consistency deviation and the prediction result, the system can automatically trigger model parameter updates to ensure the continuous accuracy and adaptability of the prediction model. This forward-looking deviation handling mechanism effectively solves the lag problem that may exist in traditional compensation methods, significantly improves the robustness of deviation compensation and the smoothness of user experience in the virtual guqin practice system, and enables the system to better adapt to the dynamic changes in user playing habits and the long-term use characteristics of data gloves.

[0070] In some of the embodiments described above in this application, a virtual image and sound feedback is generated by a feedback output module to guide users in practicing the guqin. However, in practical applications, if the acquisition sensitivity of the data acquisition module fails to fully match the user's actual playing movements or correction needs, the feedback effect may be poor. For example, if the acquisition sensitivity is insufficient, some subtle posture deviations cannot be accurately captured, which in turn affects the subsequent deviation correction effect, thus preventing the optimization potential of the entire practice system from being fully realized.

[0071] The virtual image and sound feedback output by the feedback output module are also used to reverse-optimize the acquisition sensitivity of the data acquisition module. The specific logic of reverse optimization is as follows: when the feedback indicates that the correction of posture deviation has not achieved the expected effect, the acquisition sensitivity of the data acquisition module for the corresponding key nodes of the action is increased; when the feedback indicates that the correction is effective, the current acquisition sensitivity is maintained, and the correlation between the acquisition sensitivity parameter and the corresponding effective action features is recorded.

[0072] The virtual image and sound feedback output by the feedback output module is used to reverse-optimize the acquisition sensitivity of the data acquisition module. This means that the system no longer provides feedback unidirectionally, but forms a closed-loop control system. The visual and auditory information generated by the feedback output module carries information about the user's performance and the system's correction effect. This information is reused to intelligently adjust the performance of the data acquisition module. This reverse optimization mechanism aims to improve the adaptability and accuracy of the entire system, ensuring that the quality of data acquisition can better serve subsequent deviation identification and correction. This can be achieved by integrating an evaluation unit into the feedback output module. This unit analyzes the user's response to the feedback or the correction effect indicated by the feedback. For example, by analyzing the user's action data after receiving the correction instruction, it determines whether the correction is effective and sends this evaluation result as a signal to the data acquisition module to adjust its acquisition parameters.

[0073] The specific logic of reverse optimization is that when feedback indicates that the correction of posture deviation has not achieved the expected effect, the system will increase the sensitivity of the data acquisition module to the corresponding key nodes of the action. This logic ensures that the system can specifically solve the problem of poor correction effect. When the system finds that the user's posture deviation has not been effectively improved after receiving correction feedback, it may mean that the data acquisition module is insufficient in capturing the action details related to the specific posture deviation. By increasing the acquisition sensitivity, the system can capture the action data of these key nodes more precisely, providing high-quality input for more accurate deviation identification and correction in the future. The judgment of "not achieving the expected effect" can be based on a variety of indicators. For example, after the user receives the correction instruction, the posture deviation value does not drop below the preset threshold within a certain period of time, or the matching degree with the standard hand shape is not significantly improved. Improving the acquisition sensitivity can include increasing the sensor sampling frequency, increasing the sensor resolution, or adjusting the gain parameters of specific sensors on the data gloves.

[0074] Meanwhile, when feedback indicates that the correction is effective, the system maintains the current acquisition sensitivity and records the correlation between the acquisition sensitivity parameter and the corresponding effective motion feature. This logic aims to maintain the stability of the system in an effective working state and accumulate experience. When the correction is proven effective, it means that the current data acquisition sensitivity is appropriate. Maintaining the current sensitivity can avoid unnecessary adjustments and ensure the smooth operation of the system. At the same time, recording the correlation between the acquisition sensitivity parameter and the effective motion feature is to build a knowledge base or model, so that the system can learn the optimal acquisition sensitivity configuration under different motion features, provide data support for future optimization, and achieve more intelligent adaptive adjustment. The judgment of "correction effective" can be the opposite of the judgment of "not achieving the expected effect". For example, the posture deviation value drops below the preset threshold, or the matching degree is significantly improved. Recording the correlation can be done by binding the currently effective acquisition sensitivity configuration with the specific motion feature that caused the correction to be effective and storing it in the database.

[0075] By introducing a feedback output module to perform reverse optimization of the data acquisition module, this application effectively solves the problem of mismatch between data acquisition sensitivity and actual performance needs. When the system detects that the correction effect of posture deviation is not as expected, it can intelligently improve the acquisition sensitivity of the data acquisition module for the corresponding key nodes of the movement, thereby ensuring that even subtle and hard-to-capture posture deviations can be accurately acquired and identified, providing high-quality input data for the deviation extraction unit and calibration execution unit, and significantly improving the accuracy and effectiveness of posture deviation correction. Conversely, when the correction effect is good, the system can maintain the current effective acquisition sensitivity, avoid unnecessary resource consumption, and gradually build an adaptive optimization model by recording the correlation between acquisition sensitivity parameters and effective movement characteristics. This allows the system to dynamically adjust the data acquisition strategy according to the user's playing habits and fingering characteristics, thereby achieving more personalized and efficient guqin practice guidance and improving user experience and learning efficiency.

[0076] Unlike existing technologies that focus solely on motion capture and scene rendering and lack calibration schemes tailored to different user groups, this system distinguishes between consistency deviation and posture deviation and dynamically adjusts the calibration strategy based on user type. For the first type of user, the system primarily eliminates equipment errors while preserving the user's personalized playing style; for the second type of user, the system not only eliminates equipment errors but also provides personalized posture correction instructions. This layered and adaptive calibration mechanism improves motion sensing accuracy and solves the problem of limited application of existing equipment in professional practice and beginner teaching scenarios.

Claims

1. A virtual guqin practice system, characterized in that, It includes a data acquisition module, a data processing module, and a feedback output module; The data acquisition module is used to collect hand motion sensing data of the user in a virtual guqin performance scene by using a data glove worn on the user's hand. The data processing module is communicatively connected to the data acquisition module. The data processing module includes a deviation extraction unit, a user type judgment unit, a calibration execution unit, and a standard repertoire library. The standard repertoire library stores standard performance data of at least one target repertoire. The standard performance data includes at least the standard note characteristics of the repertoire. The deviation extraction unit is used to identify and extract systematic errors that exhibit regularity in continuous movements from the hand motion sensing data, which are caused by the combined effects of the wear position offset of the data glove and the inherent characteristics of the sensor. This systematic error is defined as consistency deviation. At the same time, the portion of the hand motion sensing data that is not classified as consistency deviation but is caused by the user's hand posture or movement trajectory deviating from the standard requirements is identified as posture deviation. The calibration execution unit is communicatively connected to the deviation extraction unit and the user type determination unit, and is used to first perform the consistency deviation compensation on the hand motion sensing data; The user type determination unit is communicatively connected to the calibration execution unit and accesses the standard repertoire library. It is used to calculate the matching degree in pitch and timing between the continuous note sequence generated by the user and the standard note sequence of the target repertoire after the compensation for the consistency deviation. If the matching degree is higher than the preset threshold, the user will be classified as a first type of user; If the matching degree is lower than or equal to the preset threshold, the user is judged as a second type of user; wherein, the influence of the posture deviation corresponding to the first type of user on the performance effect is lower than the preset influence level, and the influence of the posture deviation corresponding to the second type of user on the performance effect is higher than or equal to the preset influence level. The calibration execution unit is further configured to: when the user type determination unit determines that the user is the first type of user, output compensated hand movement data; when the determination result is the second type of user, compare the compensated hand movement data with a preset standard playing hand shape, and generate a correction instruction describing the posture deviation. The feedback output module is communicatively connected to the calibration execution unit and is used to generate and present corresponding virtual image and sound feedback based on the compensated hand movement data or the correction command output by the calibration execution unit.

2. The virtual qin practicing system according to claim 1, wherein, The deviation extraction unit is specifically used to: perform statistical analysis on the hand motion sensing data of continuous frames, identify and quantify error components that recur in the continuous frames and have a fixed direction or amplitude, so as to obtain the consistency deviation. The calibration execution unit has a pre-stored set of calibration parameters. The parameter values ​​in the calibration parameter set are associated with different guqin playing fingering types. The calibration execution unit is configured to: identify the current fingering type based on the hand motion sensing data, and call the calibration parameter values ​​associated with the current fingering type to perform the compensation for the consistency deviation.

3. The virtual qin practicing system according to claim 1, wherein, When the user type determination unit determines that the user is the second type, the calibration execution unit is further configured to: acquire the user's individual hand physiological parameters through the data glove or by manual input; based on the individual hand physiological parameters, adaptively adjust the standard playing hand shape to generate a personalized standard hand shape; and generate a correction instruction describing the posture deviation by comparing the hand movement data after compensating for the consistency deviation with the personalized standard hand shape.

4. The virtual qin practicing system according to claim 1, wherein, The feedback output module includes a graphics rendering unit, which is used to indicate the position and direction of the posture deviation corresponding to the correction command in real time in a virtual scene by using visual markers superimposed on the user's virtual hand model.

5. A virtual guqin practice system according to claim 2, characterized in that, The preset threshold used to determine the user type in the user type determination unit is dynamically adjusted based on the user's historical playing data; the logic for this dynamic adjustment is as follows: When the difference between the user's performance and the standard performance, as well as the regularity of the posture deviation, meet the criteria for the first type of user, the current preset threshold is maintained. When the difference between the user's performance and the standard performance, as well as the regularity of the posture deviation, shows a trend towards meeting the criteria for the second type of user, the preset threshold is adjusted by a preset adjustment step size until the difference between the user's performance and the standard performance, as well as the regularity of the posture deviation, re-stabilize and meet the criteria for either the first type of user or the second type of user.

6. The virtual qin practicing system according to claim 5, wherein, The deviation extraction unit is configured to also perform deviation trend analysis; the deviation trend analysis includes: in the time dimension, counting the frequency of the same fingering in different playing periods of the user; in the spatial dimension, counting the overlap range of the direction and amplitude of the consistency deviation in each playing session; and determining whether the consistency deviation is an irregular fluctuation based on the above trend analysis results. The basis for determining it as an irregular fluctuation is that the change in the direction or amplitude of the currently detected deviation does not follow the change pattern obtained from the statistical analysis of historical action data. If it is determined to be an irregular fluctuation, then this part of the error is attributed to posture deviation.

7. The virtual qin practicing system according to claim 6, wherein, When extracting the consistency deviation, the deviation extraction unit simultaneously combines the continuity features of the hand movements for auxiliary judgment; the continuity features include the connection state of the trajectory between adjacent movements and the stability of the time interval; when a deviation is detected accompanied by an abnormal connection state or an unstable time interval, the deviation is directly determined to be a posture deviation.

8. The virtual qin practicing system according to claim 1, wherein, The data processing module further includes a deviation prediction unit. The deviation prediction unit performs trend fitting based on the consistency deviation direction, deviation magnitude, and movement interval of consecutive identical fingering movements in the user's historical playing data, generates a prediction result for subsequent consistency deviations, and sends the prediction result to the calibration execution unit for pre-compensation of consistency deviations in subsequently collected hand movement sensing data. When the actual consistency deviation calculated from the newly collected hand movement sensing data differs from the prediction result from a preset tolerance range, the deviation prediction unit automatically triggers model parameter updates.

9. The virtual qin practicing system according to claim 1, wherein, The virtual image and sound feedback output by the feedback output module are also used to reverse-optimize the acquisition sensitivity of the data acquisition module. The specific logic of the reverse optimization is as follows: when the feedback indicates that the user's posture deviation after correction is still not lower than a preset deviation threshold, the acquisition sensitivity of the data acquisition module for the key movement nodes corresponding to the posture deviation is increased; when the feedback indicates that the user's posture deviation after correction is lower than the preset deviation threshold, i.e., the correction is effective, the current acquisition sensitivity is maintained, and the correlation between the acquisition sensitivity parameter and the corresponding effective movement features is recorded.

Images