A perineal massage performance evaluation feedback method and system

Abstract

The present application provides a kind of perineal massage efficiency evaluation feedback method and system, including by obtaining perineal area multi-channel surface myoelectric signal, adjacent channel signal correlation is monitored in real time, when the normalized cross-correlation coefficient is lower than threshold value, utilize the myoelectric signal before signal correlation reduces to construct projection operator, project to subsequent signal to suppress displacement artifact;To the corrected or original signal, do time-frequency analysis, combine background noise adjustment kurtosis threshold to purify myoelectric time-frequency spectrum;Again based on pure spectrum, extract muscle tension relaxation rate and muscle recruitment synchronism index, the former is obtained by the first derivative of relaxation band energy integral time, and the latter is obtained by the kurtosis of recruitment band energy integral, and the band center frequency is adjusted with muscle tension relaxation rate history value;According to two indexes, calculate comprehensive efficiency score, and the muscle recruitment synchronism value is positively adjusted with muscle tension relaxation rate weight.

Description

Technical Field

[0001] This application belongs to the field of evaluation, and in particular relates to a method and system for evaluating and providing feedback on the effectiveness of perineal massage. Background Technology

[0002] Perineal massage is an important non-pharmacological intervention in the field of physical therapy and rehabilitation, particularly in postpartum recovery and the treatment of pelvic floor dysfunction. It enhances elasticity and coordination by relieving excessive tension in the perineal muscles. Surface electromyography (sEMG) is an important monitoring tool for objectively evaluating the immediate and long-term effects of perineal massage. sEMG can detect the electrophysiological activity of muscles during massage. However, the massage procedure and the unavoidable slight body movements of the subject can cause relative displacement between the electrodes and the skin. This displacement introduces motion artifacts, severely contaminating or even drowning out the weak, true electromyographic signals, distorting the signal analysis results and failing to reflect the true state of the muscles. Furthermore, sEMG signals themselves have a low signal-to-noise ratio and are easily affected by background noise, power frequency interference, and other factors. Existing filtering methods struggle to adapt to the changing noise environment during massage, resulting in poor signal purification and affecting the accuracy of subsequent analysis.

[0003] Furthermore, traditional assessment methods often rely on therapists' subjective judgment or patients' verbal feedback, lacking objectivity and reproducibility. Some studies attempting to use sEMG for assessment are limited to macroscopic indicators such as the root mean square or integrated electromyography (EMG) values ​​to reflect the overall muscle activation level. While these methods can represent changes in muscle tension to some extent, they fail to identify the deeper mechanisms of massage efficacy. For example, they cannot represent the rate of muscle tone release under massage, nor can they reveal the improvement process of motor unit recruitment patterns from disorder to coordination within the muscle. Even if multiple physiological indicators can be extracted, how to scientifically integrate these indicators into a scoring system that reflects the overall efficacy of massage remains a challenge that current technology has not yet solved. Summary of the Invention

[0004] This invention proposes a perineal massage efficacy evaluation feedback method to address the problems of existing technologies failing to reflect the true state of muscles, adapting poorly to changing noise environments during massage, and failing to identify the deep physiological mechanisms of massage efficacy. The method includes the following steps:

[0005] Acquire multi-channel surface electromyography (SEMG) signals from a multi-channel SEMG electrode array arranged in the perineal region; monitor the signal correlation between adjacent channels in the multi-channel SEMG signals in real time; when the normalized cross-correlation coefficient is lower than a preset stability threshold, perform displacement correction: extract the multi-channel SEMG signals within the previous preset time window, construct a signal subspace and generate a projection operator, and then use the projection operator to project the subsequently acquired multi-channel SEMG signals to suppress displacement artifacts, thereby obtaining displacement-corrected SEMG signals; The displacement-corrected electromyographic signal, or the original multichannel surface electromyographic signal without displacement correction, is subjected to time-frequency analysis to obtain the electromyographic time spectrum; and the local time-frequency blocks of the electromyographic time spectrum are purified based on a kurtosis threshold adjusted according to the background noise level to obtain a pure electromyographic time spectrum. Based on the pure electromyography time spectrum, two indicators are extracted: muscle tone release rate and muscle recruitment synchronicity. The muscle tone release rate is obtained by calculating the first derivative of the energy integral over time within the preset relaxation band; the muscle recruitment synchronicity is obtained by calculating the kurtosis of the energy integral within the recruitment band. A comprehensive performance score is calculated based on the muscle tone release rate and muscle recruitment synchronicity.

[0006] Optionally, the real-time monitoring of signal correlation between adjacent channels in the multi-channel surface electromyography signal, and the execution of displacement correction when the normalized cross-correlation coefficient is lower than a preset stability threshold, includes: Using a first preset duration as the sliding time window, calculate the Pearson cross-correlation coefficient for all adjacent channel pairs within each window, and use the average of all coefficients as the normalized cross-correlation coefficient.

[0007] Optionally, when the normalized cross-correlation coefficient is lower than a preset stability threshold, the multi-channel surface electromyography signal within the previous preset time window is extracted, a signal subspace is constructed, and a projection operator is generated, including: Extract multi-channel surface electromyography signal data within a second preset time window before the normalized cross-correlation coefficient falls below a stable threshold; Principal component analysis is performed on the data matrix, eigenvectors corresponding to a preset number of principal components are selected, a signal subspace is constructed, and an orthogonal projection operator is generated to project onto the subspace.

[0008] Optionally, the purification of local time-frequency blocks of the electromyographic time spectrum based on a kurtosis threshold adjusted according to the background noise level includes: Before the massage begins, collect resting electromyographic signals for a third preset duration as background noise. The background noise time spectrum is divided into multiple local time-frequency blocks. The kurtosis of the energy value sequence in each time-frequency block is calculated to obtain the kurtosis distribution of all blocks, and its mean μ and standard deviation σ are calculated. Set the kurtosis threshold to , where k is a preset coefficient; For the electromyographic time spectrum, it is also divided into local time-frequency blocks and the kurtosis value of each block is calculated. The energy of all time-frequency units in the time-frequency block with a kurtosis value lower than the threshold is set to zero.

[0009] Optionally, obtaining the muscle tone release rate by calculating the first derivative of the energy integral over time within a preset relaxation frequency band includes: The preset relaxation frequency band is set as the lower frequency limit. Up to frequency limit Scope; The energy integral value within the frequency band is calculated every fourth preset time interval; The least squares method is used to perform linear fitting on the most recent preset number of energy integral data points M, and the slope of the fitted line is used as the muscle tension release rate at the current moment.

[0010] Optionally, the center frequency of the recruitment band is adjusted based on historical values ​​of the muscle tone release rate, including: The bandwidth of the recruitment frequency band is fixed at a preset bandwidth B; Center frequency The calculation formula is: ,in Based on the center frequency, This is the frequency adjustment coefficient. This represents the muscle tone release rate in the previous calculation cycle, and The value range is limited to the lower limit of the center frequency. Up to the center frequency limit between.

[0011] Optionally, the calculation of the comprehensive efficacy score based on the muscle tone release rate and muscle recruitment synchronicity includes: The formula for calculating the overall performance score S is as follows: ,in To normalize the muscle tone release rate to the fractional lower limit Up to the maximum score Values ​​within the range To normalize muscle recruitment synchronicity to the lower limit of the fraction. Up to the maximum score The values ​​are within the range, where W is the weight of the muscle tension release rate.

[0012] Optionally, the weight of the muscle tone release rate in the calculation is adjusted based on the current value of the muscle recruitment synchronicity, including: The formula for calculating the weight W is as follows: ,in Based on weights, This is the weighting adjustment coefficient. A normalized score for current muscle recruitment synchronicity.

[0013] Furthermore, the present invention also relates to a perineal massage efficacy evaluation and feedback system, comprising the following modules: The execution module is used to acquire multi-channel surface electromyography (SEMG) signals collected by a multi-channel SEMG electrode array arranged in the perineal region; monitor the signal correlation between adjacent channels in the multi-channel SEMG signals in real time; and perform displacement correction when the normalized cross-correlation coefficient is lower than a preset stability threshold: extract the multi-channel SEMG signals within the previous preset time window, construct a signal subspace and generate a projection operator, and then use the projection operator to project the subsequently acquired multi-channel SEMG signals to suppress displacement artifacts and obtain displacement-corrected SEMG signals. An adjustment module is used to perform time-frequency analysis on the displacement-corrected electromyographic signal, or the original multi-channel surface electromyographic signal without displacement correction, to obtain the electromyographic time spectrum; and to purify the local time-frequency blocks of the electromyographic time spectrum based on a kurtosis threshold adjusted according to the background noise level to obtain a pure electromyographic time spectrum. The calculation module is used to extract two indicators, muscle tone release rate and muscle recruitment synchronicity, based on the pure electromyography time spectrum: the muscle tone release rate is obtained by calculating the first derivative of the energy integral over time within a preset relaxation frequency band; the muscle recruitment synchronicity is obtained by calculating the kurtosis of the energy integral within the recruitment frequency band. The adjustment module is used to calculate a comprehensive performance score based on the muscle tone release rate and muscle recruitment synchronicity.

[0014] Preferably, the real-time monitoring of signal correlation between adjacent channels in the multi-channel surface electromyography signal, and the execution of displacement correction when the normalized cross-correlation coefficient is lower than a preset stability threshold, includes: Using a first preset duration as the sliding time window, calculate the Pearson cross-correlation coefficient for all adjacent channel pairs within each window, and use the average of all coefficients as the normalized cross-correlation coefficient.

[0015] Preferably, when the normalized cross-correlation coefficient is lower than a preset stability threshold, the multi-channel surface electromyography signal within the previous preset time window is extracted, a signal subspace is constructed, and a projection operator is generated, including: Extract multi-channel surface electromyography signal data within a second preset time window before the normalized cross-correlation coefficient falls below a stable threshold; Principal component analysis is performed on the data matrix, eigenvectors corresponding to a preset number of principal components are selected, a signal subspace is constructed, and an orthogonal projection operator is generated to project onto the subspace.

[0016] Preferably, the purification of the local time-frequency block of the electromyographic time spectrum based on a kurtosis threshold adjusted according to the background noise level includes: Before the massage begins, collect resting electromyographic signals for a third preset duration as background noise. The background noise time spectrum is divided into multiple local time-frequency blocks. The kurtosis of the energy value sequence in each time-frequency block is calculated to obtain the kurtosis distribution of all blocks, and its mean μ and standard deviation σ are calculated. Set the kurtosis threshold to , where k is a preset coefficient; For the electromyographic time spectrum, it is also divided into local time-frequency blocks and the kurtosis value of each block is calculated. The energy of all time-frequency units in the time-frequency block with a kurtosis value lower than the threshold is set to zero.

[0017] Preferably, the step of obtaining the muscle tone release rate by calculating the first derivative of the energy integral over time within a preset relaxation frequency band includes: The preset relaxation frequency band is set as the lower frequency limit. Up to frequency limit Scope; The energy integral value within the frequency band is calculated every fourth preset time interval; The least squares method is used to perform linear fitting on the most recent preset number of energy integral data points M, and the slope of the fitted line is used as the muscle tension release rate at the current moment.

[0018] Preferably, the center frequency of the recruitment band is adjusted based on historical values ​​of the muscle tone release rate, including: The bandwidth of the recruitment frequency band is fixed at a preset bandwidth B; Center frequency The calculation formula is: ,in Based on the center frequency, Let R(t-1) be the frequency adjustment coefficient, and R(t-1) be the muscle tone release rate in the previous calculation cycle. The value range is limited to the lower limit of the center frequency. Up to the center frequency limit between.

[0019] Preferably, the calculation of the comprehensive efficacy score based on the muscle tone release rate and muscle recruitment synchronicity includes: The formula for calculating the overall performance score S is as follows: ,in To normalize the muscle tone release rate to the fractional lower limit Up to the maximum score Values ​​within the range To normalize muscle recruitment synchronicity to the lower limit of the fraction. Up to the maximum score The values ​​are within the range, where W is the weight of the muscle tension release rate.

[0020] Preferably, the weight of the muscle tone release rate in the calculation is adjusted according to the current value of the muscle recruitment synchronicity, including: The formula for calculating the weight W is as follows: ,in Based on weights, This is the weighting adjustment coefficient. A normalized score for current muscle recruitment synchronicity.

[0021] This invention provides a more reliable and physiologically significant method for evaluating the efficacy of perineal massage. By constructing and projecting a signal subspace, strong motion artifacts caused by electrode displacement during massage can be suppressed, ensuring the stability and reliability of electromyographic (EMG) signals. Simultaneously, the signal-to-noise ratio is improved by purifying the EMG time-spectrum using a kurtosis threshold set according to the background noise level. Two physiological indicators, muscle tone release rate and muscle recruitment synchronicity, are extracted, representing the improvement process of muscle state under massage. By constructing a comprehensive efficacy score and correlating and weighting the two indicators, with the weight of muscle tone release rate adjusted according to the current value of muscle recruitment synchronicity, the evaluation results better reflect the physiological mechanisms and improve the reliability of the evaluation results. Attached Figure Description

[0022] Figure 1 A flowchart of the first embodiment; Figure 2 A schematic diagram of a 4x4 surface electromyography electrode array; Figure 3 This is a schematic diagram illustrating the suppression of displacement artifacts in signal projection. Figure 4 A schematic diagram illustrating the method for calculating the sustained release rate of muscle tone. Figure 5 This is a schematic diagram of the frequency band adjustment mechanism for recruitment. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] In the first embodiment, the present invention proposes a method for evaluating and providing feedback on the effectiveness of perineal massage, such as... Figure 1 This includes the following steps: S1, acquire the multi-channel surface electromyography (SEMG) signals collected by the multi-channel SEMG electrode array arranged in the perineal region; monitor the signal correlation between adjacent channels in the multi-channel SEMG signals in real time; when the normalized cross-correlation coefficient is lower than the preset stability threshold, perform displacement correction: extract the multi-channel SEMG signals within the previous preset time window, construct a signal subspace and generate a projection operator, and then use the projection operator to project the subsequently acquired multi-channel SEMG signals to suppress displacement artifacts, thereby obtaining displacement-corrected SEMG signals; A two-dimensional surface electromyography (EMG) electrode array, for example, 4 rows and 4 columns, is attached to the patient's perineal area, ensuring good contact between the electrodes and the skin. Signals are acquired using an EMG acquisition device at a sampling frequency of 1000 Hz, and then pre-filtered using a bandpass filter from 20 to 450 Hz to obtain the raw multi-channel surface EMG signal data stream, such as... Figure 2 .

[0025] A 1-second sliding time window is set for the acquired multi-channel signals. The normalized cross-correlation coefficients between all physically adjacent channel pairs are continuously calculated and averaged. A stabilization threshold is set, for example, to 0.8. Once the average correlation coefficient falls below this value, electrode displacement is considered to have occurred. At this time, the multi-channel electromyographic signal data within the last stable 1-second time window before the displacement occurs are taken. The extracted 1-second stable data matrix is ​​zero-centered, and then principal component analysis is performed to extract the first two principal component vectors. These two vectors constitute the signal subspace. A projection operator is calculated based on this set of basis vectors. For each newly acquired data point after the displacement, the same zero-centering process is first performed, and then each data point is left-multiplied by the projection operator to project the data point onto the stable signal subspace, thereby eliminating the non-physiological signal components input by the displacement and obtaining the displacement-corrected electromyographic signal, such as... Figure 3 To accurately determine whether a displacement artifact has occurred, in one embodiment, a more stringent judgment is performed. For example, when the normalized cross-correlation coefficient is lower than a preset stability threshold and the overall energy change rate of the channel is lower than a second threshold, it is determined to be a displacement artifact and displacement correction is performed.

[0026] In an optional embodiment, the real-time monitoring of signal correlation between adjacent channels in the multi-channel surface electromyography signal, and the execution of displacement correction when the normalized cross-correlation coefficient is lower than a preset stability threshold, includes: Using a first preset duration as the sliding time window, calculate the Pearson cross-correlation coefficient for all adjacent channel pairs within each window, and use the average of all coefficients as the normalized cross-correlation coefficient.

[0027] The stability of the electromyography (EMG) signal is evaluated in real time by calculating the average Pearson cross-correlation coefficient of the multi-channel EMG signal. An 8-channel EMG sensor array is used, with a sampling frequency of 1000 Hz, and the first preset duration (sliding time window) is set to 1 second. This means that at any given time, the calculation is based on data collected within the past second, containing 1000 sampling points. Adjacent sensor channels are paired, for example, channel 1 with channel 2, channel 2 with channel 3, up to channel 7 with channel 8, forming a total of 7 channel pairs.

[0028] Within each 1-second sliding window, the Pearson cross-correlation coefficient is calculated for the data sequence of each channel pair. This coefficient represents the degree of linear correlation between the two channel signals. For example, the calculated seven coefficient values ​​might be 0.92, 0.95, 0.91, 0.89, 0.93, 0.96, and 0.90. The arithmetic mean of these seven coefficient values, approximately 0.923, is the normalized cross-correlation coefficient for the current window. This coefficient value is compared to a preset stability threshold, such as 0.85. If the coefficient consistently exceeds this threshold, it indicates that the muscle activity pattern is stable and the signal quality is high within this timeframe.

[0029] To extract core patterns from stable electromyography (EMG) signals for subsequent signal separation, in an optional embodiment, when the normalized cross-correlation coefficient is lower than a preset stability threshold, the multi-channel surface EMG signals within the previous preset time window are extracted to construct a signal subspace and generate a projection operator, including: Extract multi-channel surface electromyography signal data within a second preset time window before the normalized cross-correlation coefficient falls below a stable threshold; Principal component analysis is performed on the data matrix, eigenvectors corresponding to a preset number of principal components are selected, a signal subspace is constructed, and an orthogonal projection operator is generated to project onto the subspace.

[0030] The normalized cross-correlation coefficient is continuously monitored, assuming a stability threshold of 0.85. When this coefficient drops below 0.85 at time T, the preceding signal segment is truncated. If the second preset duration is set to 5 seconds, then 8 channels of electromyographic signal data between time T minus 5 seconds and time T are selected. If the sampling frequency is 1000Hz, an 8-row, 5000-column data matrix will be formed, where each row represents the signal of one channel.

[0031] Principal component analysis was performed on an 8×5000 data matrix. The data was zero-centered, and an 8×8 covariance matrix was calculated. The eigenvalues ​​and eigenvectors of this matrix were then determined. Assuming a preset number of eigenvalues ​​(2), the eigenvectors corresponding to the two largest eigenvalues ​​were selected. These two eigenvectors form a basis for a two-dimensional signal subspace. An 8×8 orthogonal projection operator matrix was constructed using these two basis vectors. This operator can project any 8-channel electromyography (EMG) signal onto a subspace defined by stable muscle activity patterns.

[0032] In an optional embodiment, the projection operator is used to project the subsequently acquired multi-channel surface electromyography (EMG) signals to suppress displacement artifacts and obtain displacement-corrected EMG signals. Specifically, this involves: first, using 8-channel EMG signal data within the most recent 5-second time window before the displacement occurs, zero-centering the matrix, performing principal component analysis, selecting the eigenvectors corresponding to the first two principal components to construct matrix U, and calculating the orthogonal projection operator. Simultaneously record the mean vector of this stable data segment. For each new data point acquired in real time after the displacement occurs, that is, the 8-dimensional signal vector obtained at each sampling time. First, perform the same zero-centralization process: Then the vector is projected onto the stable signal subspace: This projection operation decomposes the signal into two orthogonal components: one falls within a subspace spanned by stable muscle activity patterns, and the other is orthogonal to that subspace, thereby effectively filtering out non-physiological artifacts introduced by electrode displacement and outputting a pure displacement-corrected electromyographic signal.

[0033] S2, perform time-frequency analysis on the displacement-corrected electromyographic signal, or the original multi-channel surface electromyographic signal without displacement correction, to obtain the electromyographic time spectrum; and purify the local time-frequency blocks of the electromyographic time spectrum based on a kurtosis threshold adjusted according to the background noise level to obtain a pure electromyographic time spectrum; Specifically, for each channel's electromyography (EMG) signal, whether the original signal or the shift-corrected signal, a short-time Fourier transform is used to generate the EMG time-frequency spectrum, for example, using a Hanning window with a window length of 256 points and 50% overlap. Before the massage begins, a segment of the signal in a resting state is acquired as a background noise sample. The time-frequency spectrum of the background noise sample is divided into multiple regular local time-frequency blocks. For example, each block covers two adjacent time points in time, corresponding to a duration of 256 ms; and covers five consecutive frequencies in frequency, corresponding to a bandwidth of approximately 20 Hz. The kurtosis of the sequence of all energy values ​​within each time-frequency block is calculated. Based on the kurtosis value distribution of all noise time-frequency blocks, its mean μ and standard deviation σ are calculated. A kurtosis threshold is set according to this distribution, for example, the threshold is equal to the mean plus twice the standard deviation. The real-time generated EMG time-frequency spectrum is also divided into local time-frequency blocks, and the kurtosis value of each block is calculated. By iterating through each time-frequency block in the time spectrum, if the kurtosis value of the block is lower than the threshold T, it is determined that the block mainly contains noise. The energy values ​​of all time-frequency units within it are set to zero, thereby filtering out noise, preserving the real muscle activity components, and obtaining a pure electromyographic time spectrum.

[0034] To establish a filter capable of distinguishing real electromyographic bursts from background noise, in an optional embodiment, the purification of local time-frequency blocks of the electromyographic time spectrum based on a kurtosis threshold adjusted according to the background noise level includes: Before the massage begins, collect resting electromyographic signals for a third preset duration as background noise. The background noise time spectrum is divided into multiple local time-frequency blocks. The kurtosis of the energy value sequence in each time-frequency block is calculated to obtain the kurtosis distribution of all blocks, and its mean μ and standard deviation σ are calculated. Set the kurtosis threshold to , where k is a preset coefficient; For the electromyographic time spectrum, it is also divided into local time-frequency blocks and the kurtosis value of each block is calculated. The energy of all time-frequency units in the time-frequency block with a kurtosis value lower than the threshold is set to zero.

[0035] Before the massage begins, the user is asked to keep the target muscles completely relaxed, and a resting electromyography (EMG) signal is collected for approximately 30 seconds, serving as pure background noise. This noise signal is converted into a time-frequency spectrogram using a short-time Fourier transform, representing the distribution of noise energy over time and frequency. The noise time-frequency spectrogram is divided into multiple local time-frequency blocks, for example, each block covering approximately 200 ms in time and 30 Hz in frequency. The kurtosis of the energy value sequence within each time-frequency block is calculated. By analyzing the kurtosis values ​​of all time-frequency blocks, their overall distribution characteristics, i.e., the mean, can be obtained. and standard deviation For example, if the calculated mean is 3.1, the standard deviation is 0.5, and the preset coefficient k is 2, then the kurtosis threshold is 4.1. During subsequent massage, the time-frequency spectrum of the electromyographic signal is calculated in real time and divided into time-frequency blocks, and the kurtosis value of each block is calculated. Any time-frequency block with a kurtosis value lower than 4.1 is identified as primarily containing background noise or artifacts, and all corresponding energy values ​​within that block are set to zero, thereby purifying the electromyographic time-frequency spectrum.

[0036] S3. Based on the pure electromyography time spectrum, extract two indicators: muscle tone release rate and muscle recruitment synchronicity. The muscle tone release rate is obtained by calculating the first derivative of the energy integral over time within the preset relaxation band; the muscle recruitment synchronicity is obtained by calculating the kurtosis of the energy integral within the recruitment band. Specifically, the relaxation band is defined as 20 to 60 Hz. The energy of the pure EMG time-frequency spectrum within this band is integrated along the frequency axis to obtain an energy curve that varies over time. This energy curve is then linearly fitted using the least squares method to the data from the most recent 10 seconds, and the slope of the fitted line is taken as the muscle tone release rate. An initial recruitment band with a center frequency of 100 Hz and a width of 30 Hz is defined. Based on the average muscle tone release rate calculated over the past 5 seconds, the center frequency of this recruitment band is adjusted; for example, for every unit increase in the release rate, the center frequency is adjusted 2 Hz towards lower frequencies to capture potentially lower neuromuscular activation frequencies in a deep relaxation state. The energy of the pure EMG time-frequency spectrum is then integrated within the adjusted recruitment band, again yielding an energy time-history curve. Within a 5-second sliding window, the kurtosis value of this energy time-history curve is calculated; this value represents the muscle recruitment synchronicity.

[0037] In an optional embodiment, obtaining the muscle tone release rate by calculating the first derivative of the energy integral over time within a preset relaxation frequency band includes: The preset relaxation frequency band is set as the lower frequency limit. Up to frequency limit Scope; The energy integral value within the frequency band is calculated every fourth preset time interval; The least squares method is used to perform linear fitting on the most recent preset number of energy integral data points M, and the slope of the fitted line is used as the muscle tension release rate at the current moment.

[0038] Define a frequency range related to muscle relaxation as a preset relaxation band, for example, set a lower frequency limit. 20Hz, upper frequency limit The frequency was 60Hz. The total energy of the purified electromyography spectrum within this frequency band was calculated, with the calculation period determined by a fourth preset duration, such as once every 1 second. A sequence of energy integral values ​​changing over time was obtained. To obtain the rate of energy change, a sliding linear fitting method was used. Assuming the preset number of data points M is 10, the latest 10 energy integral data points are taken at each calculation time. The trend of the above 10 data points changing over time is fitted with a straight line using the least squares method, such as... Figure 4 The slope of this fitted straight line represents the average rate of change of energy in the relaxation band over the last 10 seconds. This slope value is defined as the current rate of muscle tension release; a negative slope with a larger absolute value indicates that muscle tension is decreasing rapidly, and the relaxation effect is better.

[0039] In an optional embodiment, the center frequency of the recruitment band is adjusted based on historical values ​​of the muscle tone release rate, including: The bandwidth of the recruitment frequency band is fixed at a preset bandwidth B; Center frequency The calculation formula is: ,in Based on the center frequency, This is the frequency adjustment coefficient, in units of... This ensures dimensional consistency. The specific values ​​are determined through system calibration. This represents the muscle tone release rate in the previous calculation cycle, and The value range is limited to the lower limit of the center frequency. Up to the center frequency limit between.

[0040] Set a set of parameters, such as the fundamental center frequency. 150Hz, frequency adjustment coefficient For 20 The preset bandwidth B is 40Hz, and the center frequency range is limited to between 120Hz and 180Hz. This recruitment band is used to assess muscle activation, and the recruitment band frequency is typically higher than the relaxation band. In each calculation cycle, the muscle tone release rate calculated in the previous cycle is used. To update the center frequency of the current cycle Assuming the previous cycle's slow-release rate was -0.5, indicating muscle relaxation, the new center frequency is calculated to be 140Hz. Since 140Hz falls within the preset range of 120 to 180Hz, the new recruitment band will be set between 120Hz and 160Hz. As the relaxation effect improves, the center frequency of the recruitment band can be appropriately lowered to guide the massage towards a mode more conducive to deep relaxation, such as... Figure 5 .

[0041] S4. Calculate the overall performance score based on the muscle tone release rate and muscle recruitment synchronicity.

[0042] Specifically, the calculated muscle tone release rate and muscle recruitment synchronicity are normalized so that their values ​​fall between 0 and 1. The weight of the muscle tone release rate is set as a function of the normalized value of muscle recruitment synchronicity; for example, the weight is 0.5 plus 0.4 times the normalized value of muscle recruitment synchronicity. The weight of muscle recruitment synchronicity is 1 minus the weight of the muscle tone release rate. The normalized values ​​of the two indicators are multiplied by their respective weights and then summed to obtain the overall performance score. For example, when muscle recruitment synchronicity is high, the weight of the muscle tone release rate will increase accordingly, and vice versa. In another embodiment, the normalized muscle tone release rate and muscle recruitment synchronicity are weighted and summed using preset weights, such as 0.6 and 0.4, to obtain the overall performance score.

[0043] To integrate two key physiological indicators into a single, easily understood massage effectiveness score, in an optional embodiment, the calculation of the comprehensive efficacy score based on the muscle tone release rate and muscle recruitment synchronicity includes: The formula for calculating the overall performance score S is as follows: ,in To normalize the muscle tone release rate to the fractional lower limit Up to the maximum score Values ​​within the range To normalize muscle recruitment synchronicity to the lower limit of the fraction. Up to the maximum score The values ​​are within the range, where W is the weight of the muscle tension release rate.

[0044] The two original metrics, namely muscle tone release rate and muscle recruitment synchronicity, are normalized. The common range of the normalized scores is set to [0,1]. For example, an original release rate of -1.5 is mapped to a normalized score according to a preset mapping rule. It equals 0.8. Similarly, a raw synchronicity index of 0.7 maps to a normalized score. It equals 0.7.

[0045] The overall effectiveness score S is calculated using a weighted summation method. The weight W varies and determines the importance of the sustained-release rate in the total score. Assuming the currently calculated weight W is 0.6, the overall effectiveness score is S = 0.76. The resulting score reflects the overall effect of the massage at the current moment; a higher score indicates a better effect.

[0046] In an optional embodiment, the weight of the muscle tone release rate in the calculation is adjusted based on the current value of the muscle recruitment synchronicity, including: The formula for calculating the weight W is as follows: ,in Based on weights, This is the weighting adjustment coefficient. A normalized score for current muscle recruitment synchronicity.

[0047] The comprehensive scoring system can intelligently adjust the evaluation focus based on muscle condition, and preset basic weights. For example, 0.5 indicates that, without a specific preference, the sustained-release rate and recruitment synchronicity are equally important. A weighting adjustment coefficient is also set. For example, 0.3.

[0048] When calculating the overall score, first obtain the normalized score of the current muscle recruitment synchronization. Assuming the current situation... A score of 0.85 indicates good synergy in muscle recruitment. At this point, the weight W for the sustained-release rate, calculated using the formula, is 0.755. This means that when muscles are well activated and recruited synchronously, more attention is paid to whether the muscles can relax quickly, thus increasing the weight of the sustained-release rate in the overall score. Conversely, if synchronicity is poor, the weight W will be close to the baseline value, giving equal weight to both indicators.

[0049] In a second embodiment, the present invention also provides a perineal massage efficacy evaluation feedback system, comprising the following modules: The execution module is used to acquire multi-channel surface electromyography (SEMG) signals collected by a multi-channel SEMG electrode array arranged in the perineal region; monitor the signal correlation between adjacent channels in the multi-channel SEMG signals in real time; and perform displacement correction when the normalized cross-correlation coefficient is lower than a preset stability threshold: extract the multi-channel SEMG signals within the previous preset time window, construct a signal subspace and generate a projection operator, and then use the projection operator to project the subsequently acquired multi-channel SEMG signals to suppress displacement artifacts and obtain displacement-corrected SEMG signals. An adjustment module is used to perform time-frequency analysis on the displacement-corrected electromyographic signal, or the original multi-channel surface electromyographic signal without displacement correction, to obtain the electromyographic time spectrum; and to purify the electromyographic time spectrum based on a kurtosis threshold adjusted according to the background noise level to obtain a pure electromyographic time spectrum. The calculation module is used to extract two indicators, muscle tone release rate and muscle recruitment synchronicity, based on the pure electromyography time spectrum: the muscle tone release rate is obtained by calculating the first derivative of the energy integral over time within a preset relaxation frequency band; the muscle recruitment synchronicity is obtained by calculating the kurtosis of the energy integral within the recruitment frequency band. The adjustment module is used to calculate a comprehensive performance score based on the muscle tone release rate and muscle recruitment synchronicity.

[0050] In this specification, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise limited, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. In this document, "a," "an," "the," "the," and "its" may also include plural forms unless the context clearly indicates otherwise. "Multiple" refers to at least two, such as 2, 3, 5, or 8, etc. "And / or" includes any and all combinations of the associated listed items.

[0051] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0052] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for evaluating and providing feedback on the effectiveness of perineal massage, characterized in that, Includes the following steps: Acquire multi-channel surface electromyography (SEMG) signals from a multi-channel SEMG electrode array arranged in the perineal region; monitor the signal correlation between adjacent channels in the multi-channel SEMG signals in real time; when the normalized cross-correlation coefficient is lower than a preset stability threshold, perform displacement correction: extract the multi-channel SEMG signals within the previous preset time window, construct a signal subspace and generate a projection operator, and then use the projection operator to project the subsequently acquired multi-channel SEMG signals to suppress displacement artifacts, thereby obtaining displacement-corrected SEMG signals; The displacement-corrected electromyographic signal, or the original multichannel surface electromyographic signal without displacement correction, is subjected to time-frequency analysis to obtain the electromyographic time spectrum; and the local time-frequency blocks of the electromyographic time spectrum are purified based on a kurtosis threshold adjusted according to the background noise level to obtain a pure electromyographic time spectrum. Based on the pure electromyography time spectrum, two indicators, muscle tension release rate and muscle recruitment synchronicity, are extracted: the muscle tension release rate is obtained by calculating the first derivative of the energy integral over time within the preset relaxation frequency band. The muscle recruitment synchronicity is obtained by calculating the kurtosis of the energy integral within the recruitment band; A comprehensive performance score is calculated based on the muscle tone release rate and muscle recruitment synchronicity.

2. The method according to claim 1, characterized in that, The real-time monitoring of signal correlation between adjacent channels in the multi-channel surface electromyography signal, when the normalized cross-correlation coefficient is lower than a preset stability threshold, performs displacement correction, including: Using a first preset duration as the sliding time window, calculate the Pearson cross-correlation coefficient for all adjacent channel pairs within each window, and use the average of all coefficients as the normalized cross-correlation coefficient.

3. The method according to claim 1, characterized in that, When the normalized cross-correlation coefficient is lower than a preset stability threshold, the multi-channel surface electromyography signals within the previous preset time window are extracted, a signal subspace is constructed, and a projection operator is generated, including: Extract multi-channel surface electromyography signal data within a second preset time window before the normalized cross-correlation coefficient falls below a stable threshold; Principal component analysis is performed on the data matrix, eigenvectors corresponding to a preset number of principal components are selected, a signal subspace is constructed, and an orthogonal projection operator is generated to project onto the subspace.

4. The method according to claim 1, characterized in that, The purification of the local time-frequency block of the electromyographic time spectrum based on a kurtosis threshold adjusted according to the background noise level includes: Before the massage begins, collect resting electromyographic signals for a third preset duration as background noise. The background noise time spectrum is divided into multiple local time-frequency blocks. The kurtosis of the energy value sequence in each time-frequency block is calculated to obtain the kurtosis distribution of all blocks, and its mean μ and standard deviation σ are calculated. Set the kurtosis threshold to , where k is a preset coefficient; For the electromyographic time spectrum, it is also divided into local time-frequency blocks and the kurtosis value of each block is calculated. The energy of all time-frequency units in the time-frequency block with a kurtosis value lower than the threshold is set to zero.

5. The method according to claim 1, characterized in that, The step of obtaining the muscle tone release rate by calculating the first derivative of the energy integral over time within a preset relaxation frequency band includes: The preset relaxation frequency band is set as the lower frequency limit. Up to frequency limit Scope; The energy integral value within the frequency band is calculated every fourth preset time interval; The least squares method is used to perform linear fitting on the most recent preset number of energy integral data points M, and the slope of the fitted line is used as the muscle tension release rate at the current moment.

6. The method according to claim 1, characterized in that, The center frequency of the recruitment band is adjusted based on historical values ​​of the muscle tone release rate, including: The bandwidth of the recruitment frequency band is fixed at a preset bandwidth B; Center frequency The calculation formula is: ,in Based on the center frequency, This is the frequency adjustment coefficient, in units of... This ensures dimensional consistency. The specific values ​​are determined through system calibration. This represents the muscle tone release rate in the previous calculation cycle, and The value range is limited to the lower limit of the center frequency. Up to the center frequency limit between.

7. The method according to claim 1, characterized in that, The calculation of the comprehensive efficacy score based on the muscle tone release rate and muscle recruitment synchronicity includes: The formula for calculating the overall performance score S is as follows: ,in To normalize the muscle tone release rate to the fractional lower limit Up to the maximum score Values ​​within the range To normalize muscle recruitment synchronicity to the lower limit of the fraction. Up to the maximum score The values ​​are within the range, where W is the weight of the muscle tension release rate.

8. The method according to claim 7, characterized in that, The weight of the muscle tone release rate in the calculation is adjusted based on the current value of the muscle recruitment synchronicity, including: The formula for calculating the weight W is as follows: ,in Based on weights, This is the weighting adjustment coefficient. A normalized score for current muscle recruitment synchronicity.

9. A perineal massage efficacy evaluation and feedback system, characterized in that, Includes the following modules: The execution module is used to acquire multi-channel surface electromyography (SEMG) signals collected by a multi-channel SEMG electrode array arranged in the perineal region; monitor the signal correlation between adjacent channels in the multi-channel SEMG signals in real time; and perform displacement correction when the normalized cross-correlation coefficient is lower than a preset stability threshold: extract the multi-channel SEMG signals within the previous preset time window, construct a signal subspace and generate a projection operator, and then use the projection operator to project the subsequently acquired multi-channel SEMG signals to suppress displacement artifacts and obtain displacement-corrected SEMG signals. An adjustment module is used to perform time-frequency analysis on the displacement-corrected electromyographic signal, or the original multi-channel surface electromyographic signal without displacement correction, to obtain the electromyographic time spectrum; and to purify the local time-frequency blocks of the electromyographic time spectrum based on a kurtosis threshold adjusted according to the background noise level to obtain a pure electromyographic time spectrum. The calculation module is used to extract two indicators, muscle tension release rate and muscle recruitment synchronicity, based on the pure electromyography time spectrum: the muscle tension release rate is obtained by calculating the first derivative of the energy integral over time within a preset relaxation frequency band; The muscle recruitment synchronicity is obtained by calculating the kurtosis of the energy integral within the recruitment band; The adjustment module is used to calculate a comprehensive performance score based on the muscle tone release rate and muscle recruitment synchronicity.

10. The system according to claim 9, characterized in that, The real-time monitoring of signal correlation between adjacent channels in the multi-channel surface electromyography signal, when the normalized cross-correlation coefficient is lower than a preset stability threshold, performs displacement correction, including: Using a first preset duration as the sliding time window, calculate the Pearson cross-correlation coefficient for all adjacent channel pairs within each window, and use the average of all coefficients as the normalized cross-correlation coefficient.

Images