A method for quantitatively evaluating lip muscle function based on myoelectric and pressure signals

Abstract

The application discloses a kind of lip muscle function quantitative evaluation methods based on myoelectricity and pressure signal, belong to rehabilitation evaluation technical field, including: obtaining the lip surface myoelectricity signal and interlabial pressure signal of subject in lip muscle action time synchronization, data pre-processing is carried out in time domain, frequency domain, nonlinear feature extraction, feature selection is carried out by ReliefF algorithm, and the age and BMI index of subject together constitute feature set, input the quantitative evaluation model of lip muscle function of classifier training.The myoelectricity and pressure signal of the person to be tested are collected, and the age and BMI index are combined to form a feature set and input into the trained evaluation model to obtain the quantitative evaluation result of the current lip muscle function of the person to be tested.The application can accurately and scientifically quantify the lip muscle function state of the patient by combining the surface myoelectricity signal, pressure signal, age and BMI index evaluation method.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rehabilitation medicine and pattern recognition, and particularly relates to a lip muscle function quantitative evaluation method based on electromyography and pressure signals. BACKGROUND

[0002] Evolution of congenital genetic or environmental factors can cause malocclusion, which directly affects the development of craniofacial and oral health in children during growth, and leads to abnormal oral function, among which the reduction of lip muscle function is a common one, and lip muscle function rehabilitation training is an important part in the process of orthodontic treatment.

[0003] At present, the methods for evaluating lip muscle function mainly include observation, X-ray examination, electromyography, and pressure method, etc. The observation method is a qualitative evaluation method with strong subjectivity, and is generally used for preliminary diagnosis in clinical practice. Domestic Ji Xiaolei et al. verified the effect of muscle function training through X-ray head lateral comparison. Foreign Saori et al. collected the surface electromyography signals of orbicularis oris muscle when closing lips before and after training in patients with abnormal lip function, and evaluated the effectiveness of training through the reduction of integral electromyography value after training. Fujiwara et al. measured the pressure around the lips with multi-directional pressure sensor after one week of training, and evaluated the effect of training through the increase of closing lip pressure. At present, there is no lip muscle function quantitative evaluation method at home and abroad that combines multi-modal signals (surface electromyography signals and pressure signals) with human development. SUMMARY

[0004] The present application is directed to the fact that the research on quantitative evaluation of lip muscle function in the prior art is mainly based on a single signal, but the fiber structure of the lip muscle is relatively complex, and a single analysis of a signal cannot accurately quantitatively evaluate the lip muscle function. The present application designs a lip muscle function quantitative evaluation method based on electromyography and pressure signals by taking surface electromyography signals and pressure signals as original data. First, the synchronous collected surface electromyography signals of orbicularis oris muscle and the pressure signals between lips are obtained, the characteristics of the surface electromyography signals and the pressure signals between lips are analyzed and selected, and then the age and BMI index of the subject are used to form a feature set, so as to train a quantitative evaluation model, so as to realize the grade classification of the lip muscle function and quantitatively evaluate the lip muscle function state.

[0005] To achieve the above purpose, the present application adopts the following technical solutions:

[0006] A lip muscle function quantitative evaluation method based on electromyography and pressure signals, comprising the following steps:

[0007] 1) obtaining the synchronous lip muscle surface electromyography signals and the pressure signals between lips of the subject when performing lip muscle action, age and height body mass index BMI;

[0008] 2) pre-process the myoelectric signal and the pressure signal, and perform activity segment division based on the pre-processed myoelectric signal and the pressure signal;

[0009] 3) perform feature extraction on the myoelectric signal and the pressure signal of the activity segment, the extracted features of the myoelectric signal include time domain features, frequency domain features and nonlinear features, and the extracted features of the pressure signal include time domain features;

[0010] 4) perform feature selection among the extracted time domain features, frequency domain features and nonlinear features of the myoelectric signal and the time domain features of the pressure signal;

[0011] 5) combine the selected features, the age and the BMI index of the subject to form a feature set, and input the feature set into a classifier to train a quantitative evaluation model of lip muscle function;

[0012] 6) obtain the surface myoelectric signal and the interlabial pressure signal of the subject to be tested, extract features after data pre-processing, and input the extracted features, the age and the BMI index into the trained quantitative evaluation model to obtain a quantitative evaluation result of the current lip muscle function grade of the subject to be tested.

[0013] Further, the pre-processing in the step 2) includes baseline correction and filtering operation.

[0014] Further, the activity segment division in the step 2) includes: performing empirical mode decomposition on the surface myoelectric signal, performing Hilbert transform, and using the average value of the baseline plus twice the standard deviation as the activity segment starting point and endpoint threshold value.

[0015] Further, the extracted features of the surface myoelectric signal in the step 3) include: time domain features: root mean square RMS, mean absolute value MAV, integral myoelectric value iEMG, zero-crossing rate ZC, wavelength WL, and variance VAR; frequency domain features: mean power frequency MPF and median frequency MDF; and nonlinear features: sample entropy SampEn and fuzzy entropy FuzzyEn.

[0016] Further, the extracted features of the pressure signal in the step 3) are time domain features, including: mean root mean square RMS, variance VAR, and standard deviation σ.

[0017] Further, the feature selection among the extracted frequency domain features and time domain features in the step 4) selects a preset number of features, and the ReliefF algorithm is used for feature selection.

[0018] Further, the input classifier in step 5) trains the quantitative evaluation model of lip muscle function, and divides the feature set of the subjects into a training set and a test set according to leave-one-out method, and takes the feature set of k-1 persons as the training set, and takes the feature set of the remaining 1 person as the test set.

[0019] The present application collects signals from two aspects, and through synchronous analysis and processing, the obtained data is more scientific and accurate, and more scientific and effective evaluation results can be obtained. The positive effects of the technical scheme of the present application are as follows:

[0020] 1. The present application can realize the familyization and communityization of rehabilitation evaluation, realize the self-assessment of patients by digitizing the guidance of doctors, and improve the training effect.

[0021] 2. The present application processes the information collected from two aspects of electromyography and pressure, combines the age and BMI index of the patient, and quantitatively obtains the lip muscle function state of the patient, so that the rehabilitation effect of the patient is more scientifically and comprehensively evaluated. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 is a flow chart of the training quantitative evaluation model according to the embodiment of the present application.

[0023] Figure 2 is a flow chart of the evaluation of the lip muscle function of the testee according to the embodiment of the present application.

[0024] Figure 3 is a schematic diagram of the measurement angle of the lateral cephalogram according to the embodiment of the present application. DETAILED DESCRIPTION

[0025] The present application will be described in detail below with reference to the drawings, and the present application will be clearly and completely described with reference to the preferred embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0026] Combined with Figure 1 and Figure 2 shown, the present embodiment is a lip muscle function quantitative evaluation method based on electromyography and pressure signals, which comprises the following steps:

[0027] Step 1: Obtain the synchronous lip muscle surface electromyography signal and interlabial pressure signal of the subject when performing lip muscle action, age and height body mass index BMI;

[0028] In one embodiment of the present application, k subjects are recruited, all of which are children aged 8-15 years old, and each 2 years old is a group, a total of 4 groups, each group contains at least 10 children with lip muscle function rehabilitation training and an equal number of healthy control children. In order to improve the classification accuracy of the model, more subjects are recruited, and it is recommended that k be greater than 100.

[0029] In one embodiment of the present application, the lip muscle function grade of the subject is evaluated by X-ray examination of the lateral head film, and three angles, SNA (sella-nasal root point- upper alveolar point angle), SNB (sella-nasal root point-lower alveolar point angle), and ANB (upper alveolar point angle-nasal root point-lower alveolar point angle), are measured. See Figure 3 The recommended values of the three angles are 82°, 79° and 4°, respectively, and the sum of the absolute values of the deviations of the three measured values of the subject from the recommended values is calculated, denoted as Sum, and the lip muscle function grade is divided into 0, 1, 2, 3 according to the Sum value, and the healthy subject grade is 0, and the Sum value range is shown in Table 1.

[0030] Table 1 Lip muscle function grade classification

[0031]

[0032] In one embodiment of the present application, the lip muscle action refers to the close closing of the upper and lower lips of the subject, and the closing degree is controlled by the subject. The lip muscle action needs to be maintained for 10s, and then rested for 10s, repeated 5 times.

[0033] Step 2: Preprocess the electromyography signal and the pressure signal, and divide the activity segment based on the preprocessed electromyography signal and the pressure signal.

[0034] In one embodiment of the present application, the preprocessing includes baseline correction and filtering operation.

[0035] Further, in one embodiment of the present application, for the electromyography signal: the baseline correction operation is to subtract the baseline offset of the surface electromyography signal in the resting state; the filtering operation after the baseline correction is to eliminate motion artifacts and high-frequency interference through a 10-450HZ Butterworth band-pass filter; and the power frequency interference is eliminated through a 50HZ power frequency notch filter. For the pressure signal: the baseline correction operation is to subtract the baseline offset of the pressure signal in the resting state; and the filtering operation after the baseline correction is to eliminate interference through a 20HZ Butterworth low-pass filter.

[0036] Further, in one of the embodiments of the present application, the surface myoelectric signal activity segment division method is: after the lip muscle surface myoelectric signal is subjected to empirical mode decomposition (EMD) and Hilbert transform (HT), the average value of the baseline is added by two times the standard deviation to determine the activity segment starting point and end point threshold. Since it is synchronous acquisition, the activity segment of the pressure signal is taken as the same time period as the myoelectric signal.

[0037] Step 3: The myoelectric signal and the pressure signal of the activity segment are subjected to window processing and feature extraction. The extracted features of the myoelectric signal include time domain features, frequency domain features and nonlinear features, and the features of the pressure signal include time domain features. The subsequent feature extraction process is also based on the analysis window.

[0038] In one of the embodiments of the present application, the window processing is to divide the surface myoelectric signal of each channel by using an overlapping analysis window with a window width of 200 ms and an overlapping rate of 50%.

[0039] Further, in one of the embodiments of the present application, the features of the activity segment myoelectric signal are extracted, and 6 commonly used time domain features, 2 commonly used frequency domain features and 2 commonly used nonlinear features are extracted. The 6 commonly used time domain features are root mean square RMS, mean absolute value MAV, integral myoelectric value iEMG, zero crossing rate ZC, wavelength WL and variance VAR; the 2 commonly used frequency domain features are mean power frequency MPF and median frequency MDF; and the 2 nonlinear features are sample entropy SampEn and fuzzy entropy FuzzyEn.

[0040] Further, in one of the embodiments of the present application, the features of the activity segment pressure signal are extracted, and 4 commonly used time domain features are extracted, namely mean root mean square RMS, variance VAR and standard deviation σ.

[0041] Step 4: The ReliefF algorithm is used to select all the features above, and a preset number of myoelectric features and pressure features are selected.

[0042] Step 5: The features selected in step 4, the age and BMI index of the subject are used to form a feature set, which is input into a classifier to train a quantitative evaluation model of lip muscle function, and the process can be seen from Figure 1 .

[0043] In one of the embodiments of the present application, the BMI index of the subject is the body mass index, and the calculation formula is

[0044]

[0045] Wherein, W refers to weight (unit: kg), and H refers to height (unit: m).

[0046] In one of the embodiments of the present application, the classifier used is a support vector machine (SVM).

[0047] The feature set of the k subjects is divided into a training set and a test set according to the leave-one-out method, the feature set of k-1 persons is taken as the training set, k is the total number of subjects, and the feature set of the remaining one person is taken as the test set. The classifier is trained and tested in the manner of k-fold cross-validation.

[0048] Step 6: as Figure 2 The lip muscle surface electromyography signal and the interlabial pressure signal of the to-be-tested person are acquired, the features are extracted after data preprocessing, and the age and BMI index are taken together to form a feature set, which is input into the quantitative evaluation model trained to obtain the quantitative evaluation result of the current lip muscle function grade of the to-be-tested person.

[0049] In one of the embodiments of the present application, the to-be-tested person is a child with lip muscle function training, and the quantitative evaluation of the lip muscle function can be performed before or after the training according to different needs, the lip muscle surface electromyography signal and the interlabial pressure signal are acquired, the data are preprocessed as in step 2, and then the preset number of electromyography features and pressure features selected in step 4 are extracted according to the method of step 3, the extracted features and the age and BMI index of the to-be-tested person are taken together to form a feature set, which is input into the quantitative evaluation model trained in step 5, so that the quantitative evaluation result of the current interlabial function grade of the to-be-tested person can be obtained, which can be used to compare and evaluate the effectiveness of different training methods, so as to improve the training efficiency and achieve better rehabilitation effect.

[0050] The above embodiments of the present application aim at the restriction of the current existing technology that the quantitative evaluation of the lip muscle function is mainly based on a single signal, and the problem that the patient cannot self-evaluate the rehabilitation progress, the lip muscle surface electromyography signal and the interlabial pressure signal of the subject are analyzed, the features are extracted and taken together with the age and BMI index of the subject to form a feature set, and the quantitative evaluation model is trained, so that the quantitative evaluation of the lip muscle function is realized.

[0051] In summary, the quantitative evaluation method of the lip muscle function based on electromyography and pressure signals of the present application acquires the lip muscle surface electromyography signal and the interlabial pressure signal of the subject during lip muscle movement, extracts the time domain, frequency domain and nonlinear features after data preprocessing, selects the features through the ReliefF algorithm, takes the age and BMI index of the subject together to form a feature set, and inputs the quantitative evaluation model of the lip muscle function trained by the classifier. The electromyography and pressure signals of the to-be-tested person are collected, the age and BMI index are taken together to form a feature set, and the evaluation model trained is input to obtain the quantitative evaluation result of the current lip muscle function of the to-be-tested person. The evaluation method combining the surface electromyography signal, the pressure signal, the age and the BMI index can accurately and scientifically quantitatively evaluate the lip muscle function state of the patient.

[0052] The above description is only to illustrate the technical solutions of the present application but not to limit the present application. Other modifications or equivalent replacements of the technical solutions of the present application made by those skilled in the art should be covered in the scope of claims of the present application as long as they are not deviated from the spirit and scope of the technical solutions of the present application.

Claims

1. A method for quantitatively assessing lip muscle function based on electromyography and pressure signals, characterized in that, Includes the following steps: 1) Obtain the electromyographic signals of the lip muscle surface and the interlip pressure signal, age, and body mass index (BMI) of the subjects while they are performing lip muscle movements. 2) Preprocess the electromyographic and pressure signals, and divide the active segment based on the preprocessed electromyographic and pressure signals; 3) Feature extraction of electromyographic and pressure signals from the active segment. The extracted features of electromyographic signals include time-domain features, frequency-domain features, and nonlinear features, while the extracted features of pressure signals include time-domain features. 4) Feature selection is performed from the time-domain features, frequency-domain features, and nonlinear features extracted from electromyographic signals, as well as the time-domain features of pressure signals; 5) Combine the selected features, the subject's age, and BMI to form a feature set, and input it into a classifier to train a quantitative assessment model for lip muscle function; 6) Acquire the synchronous surface electromyographic signals of the lip muscles and the interlip pressure signals of the test subject. After data preprocessing, extract features and input them along with age and BMI index into the trained quantitative assessment model to obtain the quantitative assessment results of the test subject's current lip muscle function level. The features extracted from the surface electromyography signal in step 3) include: temporal features: root mean square (RMS) RMS Average absolute value MAV Integrated electromyography value iEMG Zero crossing rate ZC ,wavelength WL ,variance VAR Frequency domain characteristics: average power frequency MPF and median frequency MDF Nonlinear characteristics: Sample entropy SampEn and fuzzy entropy FuzzyEn ; The extracted features of the pressure signal in step 3) are time-domain features, including: mean. Root mean square RMS ,variance VAR Standard deviation σ; In step 4), feature selection is performed on the extracted frequency domain features and time domain features. A preset number of features are selected, and the ReliefF algorithm is used for feature selection.

2. The method for quantitative assessment of lip muscle function based on electromyography and pressure signals according to claim 1, characterized in that, The preprocessing in step 2) includes baseline correction and filtering operations.

3. The method for quantitative assessment of lip muscle function based on electromyography and pressure signals according to claim 1, characterized in that, Step 2) involves dividing the activity segment into segments, including: performing empirical mode decomposition on the surface electromyography signal, and using the mean of the baseline plus twice the standard deviation after Hilbert transformation as the threshold for determining the start and end points of the activity segment.

4. The method for quantitative assessment of lip muscle function based on electromyography and pressure signals according to claim 1, characterized in that, In step 5), the input classifier is used to train a quantitative assessment model of lip muscle function. The feature set of the subjects is divided into a training set and a test set according to the leave-one-out method. The feature set of k-1 people is taken as the training set, where k is the total number of subjects, and the feature set of the remaining 1 person is taken as the test set.

Images