阅读说明：本技术 一种肌肉老化程度检测方法、装置及系统 (Muscle aging degree detection method, device and system ) 是由 何嘉杰 于 2020-12-29 设计创作，主要内容包括：本发明涉及一种肌肉老化程度检测方法、装置、系统及计算机存储介质,该方法包括以下步骤：获取待测者在指定运动状态下的多通道原始肌肉电信号；分别提取每一通道原始肌肉信号中用于表征肌肉老化程度的信号特征,得到信号特征矩阵；建立肌肉电信号样本集,采用分类器对所述肌肉电信号样本集进行训练,得到分类模型；将所述信号特征矩阵输入所述分类模型,得到所述原始肌肉电信号的老化类别,以实现肌肉老化程度检测。本发明可以实现肌肉老化程度的无创式检测,且检测效率高。(The invention relates to a muscle aging degree detection method, a device, a system and a computer storage medium, wherein the method comprises the following steps: acquiring a multichannel original muscle electric signal of a person to be detected in a designated motion state; respectively extracting signal characteristics used for representing muscle aging degree in original muscle signals of each channel to obtain a signal characteristic matrix; establishing a muscle electric signal sample set, and training the muscle electric signal sample set by adopting a classifier to obtain a classification model; and inputting the signal characteristic matrix into the classification model to obtain the aging category of the original muscle electric signal so as to realize muscle aging degree detection. The invention can realize the non-invasive detection of the muscle aging degree and has high detection efficiency.)

1. A muscle aging degree detection method is characterized by comprising the following steps:

acquiring a multichannel original muscle electric signal of a person to be detected in a designated motion state;

respectively extracting signal characteristics used for representing muscle aging degree in original muscle signals of each channel to obtain a signal characteristic matrix;

establishing a muscle electric signal sample set, and training the muscle electric signal sample set by adopting a classifier to obtain a classification model;

and inputting the signal characteristic matrix into the classification model to obtain the aging category of the original muscle electric signal so as to realize muscle aging degree detection.

2. The method of detecting muscle aging degree according to claim 1, wherein the signal characteristic is selected based on a hypothesis characteristic of a muscle aging mechanism.

3. The muscle aging degree detection method according to claim 1, wherein the signal characteristic is a time-domain signal characteristic parameter.

4. The muscle aging degree detection method according to claim 1, wherein a muscle electrical signal sample set is established, and a classifier is adopted to train the muscle electrical signal sample set to obtain a classification model, specifically:

collecting a sample signal of a muscle electric signal, calculating the signal characteristics of the sample signal to obtain a signal characteristic vector, and marking the age interval of the sample signal to obtain a label category;

and training a classifier by taking the signal characteristic vector of the sample signal as input and taking the label category of the sample signal as output to obtain a classification model.

5. The muscle aging degree detection method according to claim 4, wherein a classifier is trained to obtain a classification model by taking the signal feature vector of the sample signal as an input and taking the label category of the sample signal as an output, specifically:

initializing various parameters of a classifier;

dividing the sample signal into a training sample and a test sample;

inputting the signal feature vector of the training sample into the classifier to obtain a prediction category corresponding to the training sample;

comparing the prediction category with the corresponding label category, and correcting the parameters of the classifier according to the comparison result to obtain a classification model;

testing the current classification model by adopting the test sample;

and judging whether the test result meets a termination condition, if so, outputting the current classification model, and otherwise, adopting the next training sample to continue training the classifier.

6. The muscle aging degree detection method according to claim 5, wherein the classifier is a support vector machine, and the parameters of the support vector machine include a boundary function parameter, a decision function parameter, and a loss function parameter.

7. The method for detecting muscle aging degree according to claim 4, wherein a muscle electrical signal sample set is established, and a classifier is used to train the muscle electrical signal sample set to obtain a classification model, further comprising:

marking the gender and the muscle contraction degree of the sample signals in the muscle electric signal sample set to obtain a gender label and a muscle contraction degree label, and establishing a multi-label muscle electric signal sample set;

and training the classifier by adopting a multi-label muscle electric signal sample set to obtain a multi-label classification model.

8. A muscle degradation degree detection apparatus comprising a processor and a memory, the memory storing thereon a computer program, the computer program, when executed by the processor, implementing the muscle degradation degree detection method according to any one of claims 1 to 7.

9. A computer storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the muscle aging degree detection method according to any one of claims 1 to 7.

10. A muscle aging degree detecting system comprising the muscle aging degree detecting apparatus according to claim 8, and further comprising a muscle electric signal collecting means;

the muscle electric signal acquisition device is used for acquiring an original muscle electric signal of a person to be detected in a designated motion state and sending the original muscle electric signal to the muscle aging degree detection device;

the muscle aging degree detection device is used for detecting the original muscle electric signal by adopting the muscle aging degree detection method to obtain the aging category of the person to be detected so as to realize the detection of the muscle aging degree.

Technical Field

The invention relates to the technical field of muscle aging diagnosis, in particular to a muscle aging degree detection method, a muscle aging degree detection device, a muscle aging degree detection system and a computer storage medium.

Background

The reports of the world health organization about aging and health of the population in 2018 show that the proportion of the population in the world over 60 years will be expected to double, rising from 12% to 22% by 2050. The market demand for health care for the elderly may be met in the future with a huge explosion. The healthy skeletal muscle system is one of the important prerequisites for the quality of life of the elderly, and over 600 muscles in the skeletal muscle system are the support of the body structure and the guarantee of daily activities. At present, the main muscle aging mechanism research is based on clinical experiments, a series of exercise tests require long experimental time, and an invasive test mode is not suitable for long-term daily monitoring.

In the prior art, the following detection methods are mainly used for diagnosing the muscle aging degree:

1. muscle aging was assessed based on the neuron number measurements. The reduction of motor neurons in aging muscles results in the remaining motor neurons of the muscles expanding the range of connecting muscles. The number of muscle neurons is estimated by stimulating muscles with externally input electrical pulses while measuring the distribution of electrical signals of the muscle surface.

2. Muscle aging was assessed based on the ability of muscle fibers to conduct electricity. The electrical signal transmission to the aging muscle fibers is reduced. The aging degree of muscle fiber is also evaluated by measuring the potential difference across the same muscle fiber through the input of external electric signals.

3. Muscle aging was assessed based on muscle endurance. The endurance of aged muscles to low-intensity exercise is stronger, and the degree of muscle aging is evaluated by measuring the length of time that a subject maintains a given motion.

The above detection method has the following problems: the intrusion detection mode is adopted, so that a person to be detected generates sense of fear on detection, signal acquisition is not facilitated, and user experience is poor; the acquisition time is long, and the detection efficiency is low; the detection mode is not combined with clinic, so that the detection method is difficult to be accepted by doctors and clinical scientific research personnel.

Disclosure of Invention

In view of the above, it is necessary to provide a method, an apparatus and a system for detecting muscle aging degree, so as to solve the technical problems of poor user experience and low detection efficiency in the conventional muscle aging detection method.

The invention provides a muscle aging degree detection method, which comprises the following steps:

acquiring a multichannel original muscle electric signal of a person to be detected in a designated motion state;

respectively extracting signal characteristics used for representing muscle aging degree in original muscle signals of each channel to obtain a signal characteristic matrix;

establishing a muscle electric signal sample set, and training the muscle electric signal sample set by adopting a classifier to obtain a classification model;

and inputting the signal characteristic matrix into the classification model to obtain the aging category of the original muscle electric signal so as to realize muscle aging degree detection.

Further, the signal characteristics are selected based on the characteristics of the muscle aging mechanism hypothesis.

Further, the signal characteristic is a time-domain signal characteristic parameter.

Further, a muscle electrical signal sample set is established, and a classifier is adopted to train the muscle electrical signal sample set to obtain a classification model, which specifically comprises the following steps:

collecting a sample signal of a muscle electric signal, calculating the signal characteristics of the sample signal to obtain a signal characteristic vector, and marking the age interval of the sample signal to obtain a label category;

and training a classifier by taking the signal characteristic vector of the sample signal as input and taking the label category of the sample signal as output to obtain a classification model.

Further, training a classifier by taking the signal feature vector of the sample signal as input and taking the label category of the sample signal as output to obtain a classification model, specifically:

initializing various parameters of a classifier;

dividing the sample signal into a training sample and a test sample;

inputting the signal feature vector of the training sample into the classifier to obtain a prediction category corresponding to the training sample;

comparing the prediction category with the corresponding label category, and correcting the parameters of the classifier according to the comparison result to obtain a classification model;

testing the current classification model by adopting the test sample;

and judging whether the test result meets a termination condition, if so, outputting the current classification model, and otherwise, adopting the next training sample to continue training the classifier.

Further, the classifier is a support vector machine, and the parameters of the support vector machine include a boundary function parameter, a decision function parameter and a loss function parameter.

Further, establishing a muscle electrical signal sample set, and training the muscle electrical signal sample set by using a classifier to obtain a classification model, further comprising:

marking the gender and the muscle contraction degree of the sample signals in the muscle electric signal sample set to obtain a gender label and a muscle contraction degree label, and establishing a multi-label muscle electric signal sample set;

and training the classifier by adopting a multi-label muscle electric signal sample set to obtain a multi-label classification model.

The invention also provides a muscle aging degree detection device, which comprises a processor and a memory, wherein the memory is stored with a computer program, and the computer program is executed by the processor to realize the muscle aging degree detection method.

The present invention also provides a computer storage medium having stored thereon a computer program which, when executed by a processor, implements the muscle aging degree detection method.

The invention also provides a muscle aging degree detection system, which comprises the muscle aging degree detection device and a muscle electric signal acquisition device;

the muscle electric signal acquisition device is used for acquiring an original muscle electric signal of a person to be detected in a designated motion state and sending the original muscle electric signal to the muscle aging degree detection device;

the muscle aging degree detection device is used for detecting the original muscle electric signal by adopting the muscle aging degree detection method to obtain the aging category of the person to be detected so as to realize the detection of the muscle aging degree.

Has the advantages that: the method comprises the steps of firstly collecting original muscle electric signals of a person to be tested in an appointed movement, then extracting signal characteristics of the original muscle electric signals, establishing a muscle electric signal sample set, inputting signal characteristic vectors of the person to be tested with different physiological characteristics and different ages into a classifier, and training by adopting a machine learning classifier to obtain a classification model. And finally, the corresponding age interval classification of the person to be detected can be obtained only by inputting the signal characteristic matrix of the person to be detected into the classification model, so that the muscle aging degree is detected. The method provided by the invention can realize the detection of the muscle aging degree only by acquiring the muscle electrical signals of the person to be detected in the appointed movement, has the advantages of simple detection mode, no wound, good user experience, high detection efficiency and more reliable diagnosis of the muscle aging degree.

Drawings

FIG. 1 is a flowchart of a method for detecting muscle aging according to a first embodiment of the present invention;

FIG. 2 is a schematic diagram of signal acquisition of a first embodiment of a muscle aging degree detection method according to the present invention;

FIG. 3 is a schematic diagram of the original muscle electrical signal of the first embodiment of the muscle aging degree detection method provided by the present invention;

fig. 4 is a diagram of a signal feature extraction result of the muscle aging degree detection method according to the first embodiment of the present invention;

FIG. 5 is a schematic diagram of a classifier according to a first embodiment of the method for detecting muscle aging degree of the present invention;

fig. 6 is a system architecture diagram of a muscle aging degree detection system according to a first embodiment of the present invention.

Detailed Description

The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.

Example 1

As shown in fig. 1, embodiment 1 of the present invention provides a muscle aging degree detection method, including the steps of:

s1, acquiring multichannel original muscle electric signals of the person to be detected in the designated motion state;

s2, respectively extracting signal characteristics used for representing muscle aging degree in each channel original muscle signal to obtain a signal characteristic matrix;

s3, establishing a muscle electric signal sample set, and training the muscle electric signal sample set by adopting a classifier to obtain a classification model;

and S4, inputting the signal characteristic matrix into the classification model to obtain the aging category of the original muscle electric signal so as to realize muscle aging degree detection.

In order to improve the detection efficiency of the muscle aging degree and the user experience, the classification model algorithm provided in this embodiment realizes the classification of the muscle aging degree of the person to be detected.

Firstly, signal acquisition is carried out, and muscle electric signals of a person to be measured in a designated movement are acquired. Since biceps brachii, which has a dense type 2 muscle distribution, is a suitable subject, this embodiment requires the subject to contract the muscle in a tight-grip position while keeping the length of the muscle fiber constant, and then the original muscle electrical signal is acquired by using a 64-channel high-resolution electrode matrix a 2. The specific collection position is shown in fig. 2, and the position marked by the thick solid line rectangular frame in fig. 2 is the collection position in this embodiment: biceps brachii muscle a 1. The original muscle electrical signals collected are shown in fig. 3, and only 10 channels of muscle electrical signals are shown in fig. 3.

Extracting the signal characteristics of the original muscle electric signals, representing each channel by the signal characteristic value, and finally synthesizing a matrix of 8x8 pixels. The signal feature matrix extracted in this embodiment is shown in fig. 4, and fig. 4 shows a signal feature matrix of 8 × 8, and the signal feature value is characterized by using a gray value, so that the signal feature matrix is converted into a pixel matrix of an image.

Secondly, establishing a muscle electric signal sample set, inputting signal characteristic vectors of subjects with different physiological characteristics and different ages into a classifier, and training by adopting a machine learning classifier to obtain a classification model for classifying different samples.

And finally, the corresponding age interval classification of the person to be detected can be obtained only by inputting the signal characteristic matrix of the person to be detected into the classification model, so that the muscle aging degree is detected.

The muscle aging degree detection method provided by the invention can realize the detection of the muscle aging degree only by acquiring the muscle electrical signals of the person to be detected in the designated movement, has the advantages of simple detection mode, no wound, good user experience, high detection efficiency and more reliable diagnosis of the muscle aging degree.

From the perspective of a user, the muscle aging degree detection method provided by the invention realizes an efficient noninvasive evaluation method. The non-invasive signal acquisition scheme greatly reduces the fear of patients to hospitals, and the short signal acquisition time enables users to accept the detection mode more easily, so that the detection mode can become a part of routine health care.

Preferably, the signature is selected based on a muscle aging mechanism hypothesis.

The hypothesis is to say that a phenomenon is explained according to preset, namely, the natural phenomenon and the regularity of the phenomenon are presumed and explained according to known scientific facts and principles, and the data are classified, generalized and analyzed in detail to obtain a temporary but acceptable explanation. The hypothesis of the mechanism of muscle aging is a hypothesis presumed to be based on the mechanism of muscle aging. Muscle aging has multiple mechanisms, so based on different aging mechanism hypotheses, different signal characteristics can be extracted and used, and the signal characteristics which are sensitive to the muscle aging degree can be selected.

With the rapid development of AI in recent years, AI is gradually enabling various industries. In the medical field, the use of medical AI is slow. This is not only because of the complex clinical scene of human body, but also because of the rigor of pathological research in medical industry. In other words, the accuracy on the mathematical level must be combined with the corresponding interpretation in the clinic to be more compelling. The data feature extraction mode of the invention is based on clinical muscle aging features, so that the model is more easily accepted and used by doctors and clinical scientific researchers. Therefore, from the perspective of doctors and clinical researchers, the present embodiment provides a more interpretable mechanical learning model to interpret muscle aging characteristics.

As a preferred embodiment, the signal characteristic is a time-domain signal characteristic parameter.

Since the original muscle signal is a time-domain signal, the time-domain signal characteristic parameter is used for describing the original muscle signal. For example, signal characteristics that may be selected are: root mean square, skewness, kurtosis.

Extracting the root mean square of the original muscle electrical signal, specifically:

wherein, t₁Is root mean square, S_iThe signal value of the ith acquisition moment of the original muscle electric signal is obtained, and N is the total acquisition quantity of the original muscle electric signal;

extracting skewness of an original muscle electric signal, specifically:

wherein, t₂For skewness, E () represents the mean operation, μ is the mean of the original muscle electrical signals, and σ is the standard deviation of the original muscle electrical signals;

extracting kurtosis of an original muscle electric signal, specifically:

wherein, t₃Is kurtosis.

In this embodiment, three signal features are selected for training a classification model and identifying a muscle aging degree according to a muscle aging mechanism hypothesis, where the three signal features are respectively: root mean square, skewness, and kurtosis. Of course, in other embodiments, other signal characteristics may also be adopted, and are not described herein again.

Specifically, the root mean square of the original muscle electrical signal represents an average amplitude of the muscle electrical signal during the action performed by the user, and the strength of the electrical signal generated by the motor neural unit of the muscle decreases with age, i.e., the root mean square becomes smaller, so that the degree of muscle aging can be represented.

Skewness of the original muscle electrical signal represents the symmetry of the distribution of the muscle electrical signal during the performance of the action. Clinical tests show that the reduction of the fiber content of the second class of muscle of aged muscle leads to the reduction of anaerobic respiration intensity, which is reflected in the right deviation of time domain signals characterized by peaks, so that the muscle aging degree can be represented by the fiber content.

The kurtosis of the original muscle electrical signal represents the peak height of the probability density distribution curve of the muscle electrical signal at the average value in the action process. Clinical tests show that the reduction of the fiber content of the two types of muscle of aged muscle leads to the reduction of anaerobic respiration intensity, and the acuteness degree of a peak value is reflected in a time domain signal, so that the muscle aging degree can be represented by the acuteness degree.

Preferably, a muscle electrical signal sample set is established, and a classifier is adopted to train the muscle electrical signal sample set to obtain a classification model, specifically:

collecting a sample signal of a muscle electric signal, calculating the signal characteristics of the sample signal to obtain a signal characteristic vector, and marking the age interval of the sample signal to obtain a label category;

and training a classifier by taking the signal characteristic vector of the sample signal as input and taking the label category of the sample signal as output to obtain a classification model.

Before training the classifier, a muscle electrical signal sample set is established, and sample signals in the muscle electrical signal sample set cover various age stages as much as possible, so that a better classification effect is realized. The input of the classifier, i.e. the signal characteristics of the sample signal, is in accordance with the signal characteristics of the original muscle electrical signal selected above, e.g. the root mean square is chosen each, in order to achieve the classification purpose. The output of the classifier, i.e., the age interval of the sample signal, is divided according to the classification accuracy, for example, if the requirement for the classification accuracy is not high, the age interval may be divided into two intervals of 0 to 50 years and more than 50 years, the sample signal of the age interval of 0 to 5 years is labeled with "1" to represent the young person category, and the sample of the age interval of more than 50 years is labeled with "0" to represent the old person category. And after the establishment of the muscle electric signal sample set is completed, training the classifier to obtain a classification model. The classifier can be realized by adopting a naive Bayes classifier, an AdaBoost classifier, a support vector machine and the like.

Preferably, the signal feature vector of the sample signal is used as an input, the label category of the sample signal is used as an output, and the classifier is trained to obtain a classification model, specifically:

initializing various parameters of a classifier;

dividing the sample signal into a training sample and a test sample;

inputting the signal feature vector of the training sample into the classifier to obtain a prediction category corresponding to the training sample;

comparing the prediction category with the corresponding label category, and correcting the parameters of the classifier according to the comparison result to obtain a classification model;

testing the current classification model by adopting the test sample;

and judging whether the test result meets a termination condition, if so, outputting the current classification model, and otherwise, adopting the next training sample to continue training the classifier.

Preferably, the classifier is a support vector machine, and the parameters of the support vector machine include a boundary function parameter, a decision function parameter and a loss function parameter.

The process of training the classification model by using the support vector machine is as follows:

the classifier is a support vector machine and initializes a boundary function, a decision function and a loss function of the classifier;

dividing the sample signal into a training sample and a test sample;

inputting the signal characteristic vector of the training sample into the decision function to obtain a prediction category corresponding to the training sample;

comparing the prediction type with the corresponding label type, and correcting the boundary function according to the comparison result;

testing the current classification model by adopting the test sample, and calculating the loss value of the current classification model according to the loss function;

and judging whether the loss value is larger than a set threshold value, if so, adopting the next training sample to continue training the classifier, and otherwise, outputting the current classification model.

In this embodiment, a support vector machine is used as a classifier to implement training of a classification model, and the following specifically describes the training process for the support vector machine.

Initializing a boundary function and a decision function, and constructing a loss function;

the total samples were divided into 2 age intervals based on the age label of the sample signal. The 8x8 signal feature matrix of each sample signal is expanded into 64x1 signal feature vectors as input x_jClass label y corresponding to sample signal_jTwo types of results are: 0 represents the elderly and 1 represents the young.

Equally dividing the training sample into a plurality of subsets;

training the classifier by sequentially adopting each subset, namely inputting the signal characteristic vector of the training sample in the subset into a decision function to obtain a prediction category corresponding to the training sample, comparing the prediction category with the corresponding label category, and correcting the boundary function according to the comparison result;

after the training of a subset is completed, testing the current classification model by adopting a test sample, and calculating the loss value of the current classification model according to a loss function;

and judging whether the loss value is greater than a set threshold value, if so, adopting the next subset to continue training the classifier, and otherwise, outputting the current classification model.

Preferably, the boundary function is:

w*x+b＝0；

wherein, w and b are undetermined parameters, and x is an independent variable of a boundary function, namely a signal characteristic vector;

the decision function is:

y＝sign(w*x+b)；

y is a dependent variable of the decision function, namely an aging category, x is an independent variable of the decision function, namely a signal characteristic vector, sign () represents an aging category corresponding to an age interval where the age is located;

the loss function is:

wherein γ is a loss value, y_j(w*x_j+ b) the feature vector x for the jth test sample_jThe distance to the plane of the boundary function, M is the number of test samples.

Specifically, as shown in fig. 5, the implementation in fig. 5 is a boundary function of the support vector machine, the dotted lines on both sides of the solid line are support vectors, and the solid circle and the hollow circle represent two types of training samples, i.e., a young person and an old person, respectively. The goal of the training iterative process is to minimize the distance between the training sample points and the boundary function, and the goal of the training iterative process is formulated as:

wherein Minimzie represents the minimization, and | | represents the norm of the matrix;

i.e. the iteration target is a function of the minimum loss, such that y_j(w*x_j+b)-1≥0。

After the training of the classification model is completed, the signal characteristic matrix of the original muscle electrical signal of the person to be tested is input into the classification model, and then the corresponding aging category of '0' or '1', namely the person to be tested is a young person or an old person, can be obtained.

In this embodiment, when the age groups are divided, two classifications are taken as an example to explain, and the obtained classification model is a two-classification model. It should be understood that the division of the age intervals may be more detailed, for example, four age intervals, five age intervals, etc., and the multi-classification model is trained to achieve more detailed muscle aging detection.

Specifically, the support vector machine is still taken as an example to perform the training process description of the multi-classification model. The muscle electrical signal sample set is divided into three age intervals of 0-20 years old, 20-50 years old, and over 50 years old, and assigned category labels "0", "1", "2", respectively.

And constructing a decision function, a loss function, a boundary function and a radial basis kernel function.

The radial basis kernel function employed in this embodiment is:

K(x_i,x_j)＝exp(-γ||x_i-x_j||²),γ>0；

wherein, K (x, x)_i) Representing a radial basis kernel function value, gamma being a basis kernel parameter, exp () representing an exponential function for solving a natural constant;

the training process is similar to that of the binary model and is not described in detail herein. The objective of the training iterative process of the multi-classification model is also a minimum loss function, and the iterative objective is expressed as:

where xi is a relaxation variable, c is a penalty factor, xi_jFor the feature vector x of the jth test sample_jThe relaxation variable value of phi (omega, xi) represents the loss function.

When the basic kernel function parameter gamma is smaller, the training error is small, but all samples can be concentrated around the support vector, so that the classification practicability of the support vector machine for different age intervals is poor, when the basic kernel function parameter gamma is larger, the number of the support vectors is reduced, the error is increased, and the practicability of the support vector machine is improved. The penalty factor c represents the weight of the interval size and the classification accuracy, overfitting can be caused when the penalty factor c tends to be infinite, and underfitting can occur when the penalty factor c tends to be 0.

In order to select an appropriate kernel function parameter γ and penalty factor c, the present embodiment optimizes these two parameters by using a genetic algorithm, and the specific optimization process is as follows:

chromosomal coding: carrying out chromosome coding to establish an initial population, wherein floating point number coding can be selected as a coding rule;

determining a fitness function: when neural network training is carried out, taking whether training is used as input and expected output of a network, prediction output generated by network training, weight threshold generated by regional network training, and calculating the error square sum of the network as a target of a genetic algorithm;

selecting: selecting chromosome individuals with strong vitality in the current population by a dark fitness proportion method, and calculating the probability of selecting the individuals according to the fitness value of the individuals;

and (3) crossing: the floating point number coding individuals are properly subjected to arithmetic crossing, and the two individuals are subjected to arithmetic crossing to form two new individuals;

mutation: utilizing random numbers uniformly distributed in a certain range to replace original genes in the individual code strings by a smaller set probability to realize variation;

and when the accuracy of the training samples after the cross validation method is moved is the highest, selecting a group of penalty factors and kernel function parameters with the minimum penalty factors as the optimal solution.

Preferably, the method includes the steps of establishing a muscle electrical signal sample set, training the muscle electrical signal sample set by using a classifier, and obtaining a classification model, and further includes:

marking the gender and the muscle contraction degree of the sample signals in the muscle electric signal sample set to obtain a gender label and a muscle contraction degree label, and establishing a multi-label muscle electric signal sample set;

and training the classifier by adopting a multi-label muscle electric signal sample set to obtain a multi-label classification model.

Under the guarantee of the diversity and the quantity of the sample signals, the classifier can be further trained so as to subdivide the specific age interval, the gender, the muscle contraction degree and the like of the subject.

Example 2

Embodiment 2 of the present invention provides a muscle-aging degree detection apparatus including a processor and a memory, the memory having stored thereon a computer program, the computer program being executed by the processor to implement the muscle-aging degree detection method provided in embodiment 1.

The muscle aging degree detection device provided by the embodiment of the invention is used for realizing the muscle aging degree detection method, so that the technical effect of the muscle aging degree detection method is also realized by the muscle aging degree detection device, and the details are not repeated herein.

Example 3

Embodiment 3 of the present invention provides a computer storage medium having stored thereon a computer program that, when executed by a processor, implements the muscle-aging degree detection method provided in embodiment 1.

The computer storage medium provided by the embodiment of the invention is used for realizing the muscle aging degree detection method, so that the technical effect of the muscle aging degree detection method is also achieved by the computer storage medium, and the details are not repeated herein.

Example 4

Embodiment 4 of the present invention provides a muscle aging degree detection system, which specifically refers to fig. 6, and the system includes the muscle aging degree detection device 1, and further includes a muscle electrical signal acquisition device 2;

the muscle electric signal acquisition device 2 is used for acquiring an original muscle electric signal of a person to be detected in a designated motion state and sending the original muscle electric signal to the muscle aging degree detection device 1;

the muscle aging degree detection device 1 is used for detecting the original muscle electrical signal by adopting the muscle aging degree detection method to obtain the aging category of the person to be detected so as to realize the detection of the muscle aging degree.

The muscle electrical signal acquisition device 2 in this embodiment is implemented by using an electrode matrix device, and the muscle aging degree detection device 1 may be integrated in the muscle electrical signal acquisition device 2 or may be a detection terminal separately provided.

The muscle aging degree detection system provided by the embodiment of the invention comprises the muscle aging degree detection device 1, so that the technical effects of the muscle aging degree detection device 1 are the same as those of the muscle aging degree detection system, and the details are not repeated herein.

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.