Muscle strength prediction method and device based on FMG and sEMG signals

Abstract

The present invention relates to the technical field of muscle activation state evaluation, and particularly relates to a muscle strength prediction method and device based on FMG and sEMG signals. The method includes: collecting muscle pressure signals generated during muscle activities of a person to be tested based on FMG technology; collecting muscle physiological electrical signals generated during muscle activities of the person to be tested based on sEMG technology; extracting target time-domain features and target frequency-domain features of the muscle pressure signals and the muscle physiological electrical signals, and performing standardized processing according to target muscle strength levels to obtain processed features, and inputting the processed features into a target machine learning model to output a muscle strength prediction result of the person to be tested. Thus, the problem in the related art that the sEMG signal is easily interfered by noise, and high-density arrays are required for signal analysis, increasing the number of sensors and the difficulty of signal processing, and reducing the robustness of muscle strength prediction is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of muscle activation state evaluation, and in particular to a muscle force prediction method and device based on FMG (ForceMyography, muscle pressure map) and sEMG (Surface Electromyography, surface electromyography) signals. Background Art

[0002] Human muscle function and activation status is an important physiological parameter indicator that reflects human athletic ability and health status. In the fields of sports science and clinical postoperative rehabilitation, the evaluation of muscle function and activation status is of great significance. For example, unbalanced muscle force during exercise can lead to joint instability, wear and tear, and even injury. In postoperative rehabilitation, muscle damage and insufficient muscle activation may lead to limited joint mobility, decreased joint strength, joint damage and pain. Therefore, being able to evaluate muscle activation status and function is of great value to human sports health.

[0003] Currently, the commonly used methods for quantifying muscle strength and muscle function mainly include imaging methods and dynamometers. Among them, imaging methods mainly include musculoskeletal ultrasound, CT (Computed Tomography), MRI (Magnetic Resonance Imaging), etc., which characterize the muscle structure characteristics through imaging features, and further infer muscle strength. The main problem is that it relies on experienced doctors, the measurement repeatability is poor, and it is difficult to monitor the muscles in real time during exercise; dynamometers include isokinetic dynamometers and dynamometers. Isokinetic dynamometers are large and expensive, and are not widely used. The measurement scenarios applicable to dynamometers are relatively limited, and they are also more dependent on doctors' measurement experience and standardized measurement methods, making them difficult to promote.

[0004] In the postoperative rehabilitation of musculoskeletal injuries, patients usually need long-term rehabilitation training. Real-time monitoring of muscle activation status through wearable devices in daily training can quantify the rehabilitation training effect of specific muscles. For example, sEMG is a common muscle physiological information. At present, there are some methods based on sEMG signals and machine learning models to predict muscle strength, thereby providing effective information for the formulation of rehabilitation training programs.

[0005] However, in the related technology, since sEMG signals are easily interfered by noise and sEMG is usually mixed with signals of deep muscles and superficial muscles, high-density arrays are required for signal analysis, which increases the number of sensors and the difficulty of signal processing, reducing the robustness of muscle strength prediction, which urgently needs to be solved. Summary of the invention

[0006] The present invention provides a muscle strength prediction method and device based on FMG and sEMG signals, so as to solve the problems in the related art that since the sEMG signal is easily interfered by noise, and the sEMG usually mixes the signals of deep muscles and superficial muscles, a high-density array is required for signal analysis, which increases the number of sensors, and increases the difficulty of signal processing, and reduces the robustness of muscle strength prediction.

[0007] The first aspect embodiment of the present invention provides a muscle strength prediction method based on FMG and sEMG signals, including the following steps: collecting muscle pressure signals generated during muscle activities of at least one person to be tested based on the muscle pressure map FMG technology; collecting muscle physiological electrical signals generated during muscle activities of the at least one person to be tested based on the surface electromyogram sEMG technology; extracting target time-domain features and target frequency-domain features of the muscle pressure signals and the muscle physiological electrical signals, and performing normalization processing on the target time-domain features and the target frequency-domain features according to the target muscle strength level to obtain processed features, and inputting the processed features into a target machine learning model to output the muscle strength prediction results of the at least one person to be tested.

[0008] Optionally, in an embodiment of the present invention, the extracting the target time-domain features and target frequency-domain features of the muscle pressure signals and the muscle physiological electrical signals includes: extracting a first average value feature of the muscle pressure signal based on a target activity segment and a second average value feature based on a non-activity segment; extracting the root mean square feature, average absolute value feature, slope sign change rate feature, zero crossing rate feature, median frequency feature and average power frequency feature of the muscle physiological electrical signal.

[0009] Optionally, in an embodiment of the present invention, the performing normalization processing on the target time-domain features and the target frequency-domain features according to the target muscle strength level to obtain processed features includes: extracting signal features of each muscle of the at least one person to be tested under a target action type; performing signal normalization processing on the signal features of the at least one person to be tested based on a training label weight range to obtain the processed features.

[0010] Optionally, in an embodiment of the present invention, before inputting the processed features into the target machine learning model, it further includes: determining a feature data set by using the processed features, and dividing the feature data set into a training set and a test set; training a pre-constructed machine learning model by using the training set to obtain a trained machine learning model; inputting the test set into the trained machine learning model to output a model prediction result, and generating the target machine learning model when the model prediction result meets a preset test condition.

[0011] In a second aspect embodiment of the present invention, a muscle strength prediction device based on FMG and sEMG signals is provided, including: a first acquisition module, configured to acquire muscle pressure signals generated during muscle activities of at least one person to be tested based on the muscle pressure graph (FMG) technology; a second acquisition module, configured to acquire muscle physiological electrical signals generated during muscle activities of the at least one person to be tested based on the surface electromyogram (sEMG) technology; a prediction module, configured to extract target time-domain features and target frequency-domain features of the muscle pressure signals and the muscle physiological electrical signals, perform normalization processing on the target time-domain features and the target frequency-domain features according to target muscle strength levels to obtain processed features, and input the processed features into a target machine learning model to output muscle strength prediction results of the at least one person to be tested.

[0012] Optionally, in an embodiment of the present invention, the prediction module includes: a first extraction unit, configured to extract a first average value feature of the muscle pressure signal based on a target activity segment and a second average value feature based on a non-activity segment; a second extraction unit, configured to extract a root mean square feature, an average absolute value feature, a slope sign change rate feature, a zero crossing rate feature, a median frequency feature, and an average power frequency feature of the muscle physiological electrical signal.

[0013] Optionally, in an embodiment of the present invention, the prediction module includes: a third extraction unit, configured to extract signal features of each muscle of the at least one person to be tested under a target action type; a processing unit, configured to perform signal normalization processing on the signal features of the at least one person to be tested based on a training label weight range to obtain the processed features.

[0014] Optionally, in an embodiment of the present invention, the device of the present invention embodiment further includes: a determination module, configured to determine a feature data set using the processed features and divide the feature data set into a training set and a test set before inputting the processed features into a target machine learning model; an acquisition module, configured to train a pre-constructed machine learning model using the training set to obtain a trained machine learning model before inputting the processed features into a target machine learning model; a generation module, configured to input the test set into the trained machine learning model to output a model prediction result and generate the target machine learning model when the model prediction result meets a preset test condition before inputting the processed features into a target machine learning model.

[0015] In a third aspect embodiment of the present invention, an electronic device is provided, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement the muscle strength prediction method based on FMG and sEMG signals as described in the above embodiments.

[0016] In a fourth aspect embodiment of the present invention, a computer-readable storage medium is provided, where the computer-readable storage medium stores a computer program, and when the program is executed by a processor, the muscle strength prediction method based on FMG and sEMG signals as described above is implemented.

[0017] In a fifth aspect embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed, it is used to implement the muscle strength prediction method based on FMG and sEMG signals as described above.

[0018] Embodiments of the present invention can collect muscle pressure signals generated during the muscle activities of a person to be tested based on FMG technology, and collect muscle physiological electrical signals generated during the muscle activities of the person to be tested based on sEMG technology. Then, extract the target time-domain features and target frequency-domain features of the muscle pressure signals and muscle physiological electrical signals, and perform standardization processing according to the target muscle strength level to obtain the processed features, and input the processed features into the target machine learning model to output the muscle strength prediction result of the person to be tested, effectively reducing the cost requirements and improving the robustness of muscle strength prediction. Thus, the problem in the related art that due to the sEMG signal being easily affected by noise interference and requiring the use of a high-density array for signal parsing, increasing the number of sensors and the difficulty of signal processing, and reducing the robustness of muscle strength prediction is solved.

[0019] Additional aspects and advantages of the present invention will be given in part in the following description, become apparent in part from the following description, or be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The above and / or additional aspects and advantages of the present invention will become apparent and be readily understood from the following description of the embodiments in conjunction with the drawings, where:

[0021] Figure 1 FIG. is a schematic structural diagram of a muscle strength prediction system based on FMG and sEMG signals according to an embodiment of the present invention;

[0022] Figure 2 FIG. is a schematic layout diagram of FMG sensors and sEMG sensors in a specific embodiment of the present invention;

[0023] Figure 3 FIG. is a flowchart of a muscle strength prediction method based on FMG and sEMG signals according to an embodiment of the present invention;

[0024] Figure 4 Schematic diagram of a 90-degree isometric contraction action of biceps curl and extension for a specific embodiment of the present invention;

[0025] Figure 5 Flowchart of a muscle strength prediction method based on FMG and sEMG signals for a specific embodiment of the present invention;

[0026] Figure 6 Schematic diagram of the normalization of the original data set and the model training process for a specific embodiment of the present invention;

[0027] Figure 7 Schematic diagram of the comparison of model results of three different signal strategies for a specific embodiment of the present invention;

[0028] Figure 8 Schematic diagram of the structure of a muscle strength prediction device based on FMG and sEMG signals provided according to an embodiment of the present invention;

[0029] Figure 9 Schematic diagram of the structure of an electronic device provided according to an embodiment of the present invention. Detailed description of the specific implementation

[0030] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present invention, but should not be construed as limiting the present invention.

[0031] The following describes the muscle strength prediction method and device based on FMG and sEMG signals of the embodiment of the present invention with reference to the accompanying drawings. In view of the problem that the sEMG signal is easily interfered by noise and needs to use a high-density array for signal analysis in the related technology mentioned in the background technology center, the number of sensors is increased, the difficulty of signal processing is increased, and the robustness of muscle strength prediction is reduced, the present invention provides a muscle strength prediction method based on FMG and sEMG signals, in which the muscle pressure signal generated by the muscle activity of the person to be tested can be collected based on the FMG technology, and the muscle physiological electrical signal generated by the muscle activity of the person to be tested can be collected based on the sEMG technology, then, the target time domain features and target frequency domain features of the muscle pressure signal and the muscle physiological electrical signal are extracted, and standardized according to the target muscle strength level to obtain the processed features, and the processed features are input into the target machine learning model to output the muscle strength prediction result of the person to be tested, effectively reducing the cost requirements and improving the robustness of muscle strength prediction. Thereby, the problem in the related art that sEMG signals are easily affected by noise and require the use of high-density arrays for signal analysis, which increases the number of sensors, increases the difficulty of signal processing, and reduces the robustness of muscle strength prediction is solved.

[0032] like Figure 1 As shown, an embodiment of the present invention designs a muscle strength prediction system based on FMG and sEMG signals, including a sensor module and a signal acquisition circuit.

[0033] Among them, Figure 2 As shown, based on the sensor module and the signal acquisition circuit, the embodiment of the present invention designs a wearable device for collecting FMG signals and sEMG signals, wherein the FMG sensor and the sEMG sensor synchronously and co-locately collect muscle pressure signals and muscle physiological electrical signals. The sEMG sensor is placed at the belly of the muscle to be measured. Taking the measurement of elbow joint muscle strength as an example, the sEMG sensor is placed at three positions: the biceps, the long head of the triceps, and the lateral head of the triceps. The sEMG electrode is attached to the skin surface through gel to collect sEMG signals, i.e., muscle physiological electrical signals. The FMG sensor is set on a circle of straps on the outside to collect FMG signals, i.e., muscle pressure signals.

[0034] Among them, the FMG signal is the pressure signal of the body surface-strap interface caused by muscle activity under the external strap. Its signal acquisition is highly robust and not easily interfered by noise. In addition, the FMG signal is highly correlated with muscle strength. The fusion of FMG and sEMG signals can reduce the number of sensors, improve system robustness and recognition accuracy, and is simple to operate and easy to promote, and can continuously monitor specific muscles for a long time.

[0035] Specifically, Figure 3Schematic flow chart of a muscle strength prediction method based on FMG and sEMG signals provided by an embodiment of the present invention.

[0036] As Figure 3 shown, the muscle strength prediction method based on FMG and sEMG signals includes the following steps:

[0037] In step S301, based on the muscle pressure map FMG technology, muscle pressure signals generated during muscle activities of at least one person to be tested are collected.

[0038] It can be understood that in the embodiment of the present invention, the FMG technology in the above-mentioned wearable device can be used to collect muscle pressure signals generated during muscle activities of at least one person to be tested. For example, as Figure 4 shown, it is a muscle strength test experimental method. The test subject performs a cycle of relaxation - 90-degree isometric contraction under the voice command of computer software. The signal activity segments are divided according to the voice command time stamps. The signals in the activity segments and relaxation segments will be used for subsequent feature extraction and normalization, effectively improving the feasibility of muscle strength prediction.

[0039] Among them, the signal acquisition circuit in the above-mentioned device can collect and transmit sensing signals, generate a 1kHz sine wave to provide a power supply signal for the FMG sensor; perform voltage conversion on the impedance of the FMG sensor, and calculate the effective value as the FMG signal, that is, the muscle pressure signal.

[0040] In step S302, based on the surface electromyogram sEMG technology, muscle physiological electrical signals generated during muscle activities of at least one person to be tested are collected.

[0041] It can be understood that in the embodiment of the present invention, the sEMG technology in the above-mentioned wearable device can be used to collect muscle physiological electrical signals generated during muscle activities of at least one person to be tested. For example, in the above muscle strength test experiment, the signal acquisition circuit in the above-mentioned device can filter (10 - 500Hz band-pass filtering, human common-mode noise suppression), amplify and collect the sEMG signal, that is, the muscle physiological electrical signal, and transmit the sEMG signal to the upper computer software, effectively improving the reliability of muscle strength prediction.

[0042] Among them, since the FMG sensing power supply signal causes crosstalk to the sEMG signal, based on the above signal acquisition circuit, the sEMG acquisition frequency is set to 1600Hz. At this time, the 1kHz FMG AC power supply signal causes 600Hz noise to the sEMG, which can be filtered by the band-pass filter.

[0043] In step S303, the target time-domain features and target frequency-domain features of the muscle pressure signal and the muscle physiological electrical signal are extracted, and the target time-domain features and target frequency-domain features are standardized according to the target muscle strength level to obtain the processed features, and the processed features are input into the target machine learning model to output the muscle strength prediction results of at least one person to be tested.

[0044] In the embodiment of the present invention, the target machine learning model is a pre-trained optimal machine learning model, which will be introduced in the following steps and will not be specifically elaborated here.

[0045] It can be understood that the embodiment of the present invention can extract the target time-domain features and target frequency-domain features of the muscle pressure signal and the muscle physiological electrical signal of each person to be tested, and standardize the target time-domain features and target frequency-domain features of each person to be tested according to the muscle strength level, for example, perform normalization processing to obtain the processed features, and can also perform unified normalization processing on the personal information (height, weight, age, gender, etc.) of all persons to be tested, and input the processed features and the personal information of the person to be tested into the target machine learning model, for example, the trained SVR regression model, so as to output the muscle strength prediction results of each person to be tested, effectively reducing the cost requirements and improving the accuracy of muscle strength prediction.

[0046] Among them, in an embodiment of the present invention, extracting the target time-domain features and target frequency-domain features of the muscle pressure signal and the muscle physiological electrical signal includes: extracting the first average value feature of the muscle pressure signal based on the target activity segment and the second average value feature based on the non-activity segment; extracting the root mean square feature, average absolute value feature, slope sign change rate feature, zero crossing rate feature, median frequency feature, and average power frequency feature of the muscle physiological electrical signal.

[0047] In the actual execution process, the embodiment of the present invention can extract the average value of the active segment of the FMG signal, that is, the muscle pressure signal, as the signal feature when the muscle is activated, and extract the average value of the non-active segment as the initial pressure value; extract features such as root mean square, average absolute value, slope sign change rate, zero crossing rate, median frequency, and average power frequency from the sEMG signal, that is, the muscle physiological electrical signal. Among them, the root mean square can represent the overall intensity of muscle activity; the average absolute value can also represent the overall intensity of muscle activity, but is suitable for situations where noise influence needs to be reduced; the slope sign change rate can be used to identify muscle fatigue or different exercise types; the zero crossing rate can evaluate the dynamic nature of muscle activity; the median frequency can be used to monitor muscle fatigue status; the average power frequency can also be used to evaluate the degree of muscle fatigue, so as to deeply understand the essence of muscle activity, improve sports performance, and promote rehabilitation.

[0048] Optionally, in an embodiment of the present invention, the target time-domain features and target frequency-domain features are normalized according to the target muscle strength level to obtain processed features, including: extracting the signal features of each muscle of at least one person to be tested under the target action type; performing signal normalization on the signal features of at least one person to be tested based on the training label weight range to obtain the processed features.

[0049] As a possible implementation manner, in the embodiment of the present invention, under the target action type, for example, when the person to be tested performs actions such as biceps curl and 90-degree isometric contraction of arm extension, the embodiment of the present invention can extract the signal features of each muscle of each person to be tested, and based on the training label weight range, for example, perform signal normalization on the signal features of each person to be tested according to the muscle strength level ranges of 0 kg, 1 kg, 2 kg, and 3 kg to obtain the processed features, thereby improving the generalization ability and prediction accuracy of the subsequent model.

[0050] Optionally, in an embodiment of the present invention, before inputting the processed features into the target machine learning model, it further includes: determining a feature data set using the processed features, and dividing the feature data set into a training set and a test set; training a pre-constructed machine learning model using the training set to obtain a trained machine learning model; inputting the test set into the trained machine learning model to output a model prediction result, and generating a target machine learning model when the model prediction result meets a preset test condition.

[0051] In the embodiment of the present invention, in combination with Figure 5 and Figure 6 As shown, first, a signal feature normalization method can be designed. Under a specific action type, each data set contains data of multiple subjects. For each subject, their personal information (height, weight, age, gender, etc.) is recorded, features are extracted from the signals of each muscle, and the muscle strength level corresponding to each row of data is stored in the data label column to obtain the original data set. Data normalization is a conventional operation in machine learning. However, different from the traditional method of normalizing the same features of all people together, in the embodiment of the present invention, the signal features of each person can be signal-normalized according to the training label weight range. For example, the data within the 0-3 kg label range of each subject is normalized, so as to eliminate the signal baseline difference caused by individual differences and greatly improve the model output result. For the personal information of the subjects, since each person has only one piece of information, this part of the data is uniformly normalized in the traditional normalization manner.

[0052] Furthermore, in combination with Figure 5 and Figure 6As shown, the embodiments of the present invention can obtain a feature dataset after the above standardization, and divide the feature dataset into a training set and a test set. Among them, the data in the training set is used to train the model, and the test set is used for model effect evaluation. The dataset division method may include random division; the first 80% is used for the training set and the last 20% is used for the test set; n - 1 subjects are used for the training set and the remaining 1 subject is used for the test set, etc. After training the model with the training set data, the best parameters can be obtained. After saving these parameters, the model is used for prediction. Taking a pre - constructed machine learning model, such as an SVR model, as an example, its model training and parameter optimization process is as follows: Search for the best kernel function type, gamma, and C through grid parameters. The kernel function type may include linear, rbf, poly, etc. The search range of gamma and C is from 0.001 to 1000. The training features input into the model include the above - mentioned standardized signal features, as well as features such as the height, weight, age, and gender of the subject, so as to obtain the trained SVR model. Then, in the model evaluation stage, by inputting the data of the test set, the model prediction result can be obtained. The model prediction result can be compared with the label of the test set to evaluate the quality of the model. When the model evaluation result is optimal, a target machine learning model is generated. Secondly, when the model is deployed and applied, after inputting the standardized data (personal information and signal feature data of each subject to be tested) into the trained model, the prediction result can be obtained.

[0053] It should be noted that the preset test condition is the condition when the model evaluation result is optimal, which is specifically set by those skilled in the art and will not be specifically limited here.

[0054] For example, as Figure 7 shown, it is a comparison of model evaluation results for the dataset division method of "n - 1 subjects are used for the training set and the remaining 1 subject is used for the test set". Among them, each point represents the score value output by the model when the data of a certain subject is used as the test set. It can be seen that compared with a single FMG or sEMG signal, the iFEMG signal fusion strategy can improve the situation where the model fitting fails for some subjects, making the model have a higher score value for all subjects, and further indicating that the iFEMG strategy can improve the accuracy and versatility of the model.

[0055] The muscle strength prediction method based on FMG and sEMG signals proposed according to an embodiment of the present invention can collect muscle pressure signals generated during the muscle activities of a person to be tested based on FMG technology, and collect muscle physiological electrical signals generated during the muscle activities of the person to be tested based on sEMG technology. Then, the target time-domain features and target frequency-domain features of the muscle pressure signals and muscle physiological electrical signals are extracted, and standardized processing is performed according to the target muscle strength level to obtain the processed features, and the processed features are input into the target machine learning model to output the muscle strength prediction result of the person to be tested, effectively reducing the cost requirement and improving the robustness of muscle strength prediction. Thus, the problem in the related art that the sEMG signal is easily interfered by noise, and a high-density array needs to be used for signal analysis, increasing the number of sensors and the difficulty of signal processing, and reducing the robustness of muscle strength prediction is solved.

[0056] Next, a muscle strength prediction device based on FMG and sEMG signals proposed according to an embodiment of the present invention will be described with reference to the accompanying drawings.

[0057] Figure 8 FIG. is a block diagram of a muscle strength prediction device based on FMG and sEMG signals according to an embodiment of the present invention.

[0058] As Figure 8 shown, the muscle strength prediction device 10 based on FMG and sEMG signals includes: a first acquisition module 100, a second acquisition module 200, and a prediction module 300.

[0059] Specifically, the first acquisition module 100 is configured to collect muscle pressure signals generated during the muscle activities of at least one person to be tested based on the muscle pressure map FMG technology.

[0060] The second acquisition module 200 is configured to collect muscle physiological electrical signals generated during the muscle activities of at least one person to be tested based on the surface electromyogram sEMG technology.

[0061] The prediction module 300 is configured to extract the target time-domain features and target frequency-domain features of the muscle pressure signals and muscle physiological electrical signals, and perform standardized processing on the target time-domain features and target frequency-domain features according to the target muscle strength level to obtain the processed features, and input the processed features into the target machine learning model to output the muscle strength prediction results of at least one person to be tested.

[0062] Optionally, in an embodiment of the present invention, the prediction module 300 includes: a first extraction unit and a second extraction unit.

[0063] Wherein, the first extraction unit is configured to extract the first average value feature of the muscle pressure signal based on the target activity segment and the second average value feature based on the non-activity segment.

[0064] A second extraction unit, configured to extract the root mean square feature, average absolute value feature, slope sign change rate feature, zero crossing rate feature, median frequency feature, and average power frequency feature of the muscle physiological electrical signal.

[0065] Optionally, in an embodiment of the present invention, the prediction module 300 includes: a third extraction unit and a processing unit.

[0066] Wherein, the third extraction unit is configured to extract the signal features of each muscle of at least one person to be tested under a target action type.

[0067] The processing unit is configured to perform signal normalization processing on the signal features of at least one person to be tested based on the training label weight range to obtain the processed features.

[0068] Optionally, in an embodiment of the present invention, the muscle strength prediction device 10 based on FMG and sEMG signals further includes: a determination module, an acquisition module, and a generation module.

[0069] Wherein, the determination module is configured to determine a feature data set by using the processed features and divide the feature data set into a training set and a test set before inputting the processed features into a target machine learning model.

[0070] The acquisition module is configured to train a pre-constructed machine learning model by using the training set to obtain a trained machine learning model before inputting the processed features into the target machine learning model.

[0071] The generation module is configured to input the test set into the trained machine learning model before inputting the processed features into the target machine learning model to output a model prediction result, and generate a target machine learning model when the model prediction result meets a preset test condition.

[0072] It should be noted that the foregoing explanation of the embodiment of the muscle strength prediction method based on FMG and sEMG signals is also applicable to the muscle strength prediction device 10 based on FMG and sEMG signals in this embodiment, and will not be elaborated herein.

[0073] The muscle strength prediction device based on FMG and sEMG signals proposed according to an embodiment of the present invention can collect the muscle pressure signals generated during the muscle activities of the person to be tested based on FMG technology, and collect the muscle physiological electrical signals generated during the muscle activities of the person to be tested based on sEMG technology. Then, the target time-domain features and target frequency-domain features of the muscle pressure signals and the muscle physiological electrical signals are extracted, and standardized processing is performed according to the target muscle strength level to obtain the processed features, and the processed features are input into the target machine learning model to output the muscle strength prediction result of the person to be tested, effectively reducing the cost requirement and improving the robustness of muscle strength prediction. Thus, the problem in the related art that due to the sEMG signal being easily interfered by noise and requiring the use of a high-density array for signal parsing, the number of sensors is increased, and the difficulty of signal processing is increased, reducing the robustness of muscle strength prediction is solved.

[0074] Figure 9 The structural schematic diagram of the electronic device provided by the embodiment of the present invention. The electronic device may include:

[0075] A memory 901, a processor 902, and a computer program stored on the memory 901 and executable on the processor 902.

[0076] When the processor 902 executes the program, it implements the muscle strength prediction method based on FMG and sEMG signals provided in the above embodiment.

[0077] Further, the electronic device further includes:

[0078] A communication interface 903 for communication between the memory 901 and the processor 902.

[0079] The memory 901 is used to store a computer program executable on the processor 902.

[0080] The memory 901 may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory.

[0081] If the memory 901, the processor 902, and the communication interface 903 are implemented independently, the communication interface 903, the memory 901, and the processor 902 can be interconnected through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 9 only a thick line is used to represent it in Figure 9 , but it does not mean that there is only one bus or one type of bus.

[0082] Optionally, in a specific implementation, if the memory 901, the processor 902, and the communication interface 903 are integrated on a single chip, the memory 901, the processor 902, and the communication interface 903 can communicate with each other through an internal interface.

[0083] The processor 902 may be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0084] This embodiment also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned muscle strength prediction method based on FMG and sEMG signals is implemented.

[0085] This embodiment also provides a computer program product, including a computer program, which is used to implement the above-mentioned muscle strength prediction method based on FMG and sEMG signals when the computer program is executed.

[0086] In the description of this specification, the descriptions with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0087] In addition, the terms "first" and "second" are used only for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise specifically defined.

[0088] Any process or method description shown in a flowchart or described in other ways herein can be understood to represent a module, segment, or portion of code including one or N executable instructions for implementing a customized logical function or process, and the scope of the preferred embodiments of the present invention includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0089] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, which can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in conjunction with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection part (electronic device) having one or N wirings, a portable computer disk cartridge (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0090] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits having suitable combinational logic gate circuits, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), and the like.

[0091] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of the above embodiments can be completed by a program instructing relevant hardware, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0092] In addition, each functional unit in various embodiments of the present invention may be integrated into a processing module, may exist separately as individual physical units, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0093] The above-mentioned storage medium may be a read-only memory, a magnetic disk, an optical disc, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A muscle strength prediction method based on FMG and sEMG signals, characterized in that: The following steps are involved: Based on the muscle pressure graph FMG technology, the muscle pressure signal generated by the muscle activity of at least one person to be tested is collected; Based on the surface electromyography (sEMG) technology, collecting muscle physiological electrical signals generated by muscle activity of at least one person to be tested; Extract target time domain features and target frequency domain features of the muscle pressure signal and the muscle physiological electrical signal, and standardize the target time domain features and the target frequency domain features according to the target muscle strength level to obtain processed features, and input the processed features into the target machine learning model to output the muscle strength prediction result of the at least one person to be tested; wherein, extracting the target time domain features and the target frequency domain features of the muscle pressure signal and the muscle physiological electrical signal includes: extracting a first average value feature based on a target active segment and a second average value feature based on an inactive segment of the muscle pressure signal; extracting the root mean square feature, the average absolute value feature, the slope sign change rate feature, the zero crossing rate feature, the median frequency feature and the average power frequency feature of the muscle physiological electrical signal; Among them, the first average value feature is the signal feature when the muscle is activated, and the second average value feature is used as the initial pressure value; the root mean square feature represents the overall intensity of the muscle activity; the average absolute value feature also represents the overall intensity of the muscle activity; the slope sign change rate feature is used to identify muscle fatigue or different types of exercise; the zero crossing rate feature is used to evaluate the dynamic properties of muscle activity; the median frequency feature is used to monitor the muscle fatigue status; and the average power frequency feature is used to evaluate the degree of muscle fatigue.

2. The muscle strength prediction method based on FMG and sEMG signals according to claim 1, characterized in that: The target time domain features and the target frequency domain features are normalized according to the target muscle strength level to obtain the processed features, including: Under the target action type, extracting the signal feature of each muscle of the at least one person to be tested; Based on the training label weight range, signal normalization processing is performed on the signal features of the at least one person to be tested to obtain the processed features.

3. The muscle strength prediction method based on FMG and sEMG signals according to claim 1, characterized in that: Before inputting the processed features into the target machine learning model, the method further includes: Determine a feature data set using the processed features, and divide the feature data set into a training set and a test set; Using the training set to train a pre-built machine learning model to obtain a trained machine learning model; The test set is input into the trained machine learning model to output the model prediction result, and when the model prediction result meets the preset test conditions, the target machine learning model is generated.

4. A muscle strength prediction device based on FMG and sEMG signals, characterized in that: include: The first acquisition module is used to acquire a muscle pressure signal generated by muscle activity of at least one person to be tested based on the muscle pressure graph FMG technology; A second acquisition module is used to acquire muscle physiological electrical signals generated by muscle activity of the at least one person to be tested based on surface electromyography (sEMG) technology; A prediction module is used to extract target time domain features and target frequency domain features of the muscle pressure signal and the muscle physiological electrical signal, and standardize the target time domain features and the target frequency domain features according to the target muscle strength level to obtain processed features, and input the processed features into a target machine learning model to output a muscle strength prediction result of at least one person to be tested; wherein, the extraction of the target time domain features and the target frequency domain features of the muscle pressure signal and the muscle physiological electrical signal includes: extracting a first average value feature based on a target active segment and a second average value feature based on an inactive segment of the muscle pressure signal; extracting a root mean square feature, an average absolute value feature, a slope sign change rate feature, a zero crossing rate feature, a median frequency feature, and an average power frequency feature of the muscle physiological electrical signal; Among them, the first average value feature is the signal feature when the muscle is activated, and the second average value feature is used as the initial pressure value; the root mean square feature represents the overall intensity of muscle activity; the average absolute value feature also represents the overall intensity of muscle activity; the slope sign change rate feature is used to identify muscle fatigue or different types of exercise; the zero crossing rate feature is used to evaluate the dynamic properties of muscle activity; the median frequency feature is used to monitor muscle fatigue status; and the average power frequency feature is also used to evaluate the degree of muscle fatigue.

5. The muscle strength prediction device based on FMG and sEMG signals according to claim 4, characterized in that: The prediction module comprises: A third extraction unit is used to extract the signal characteristics of each muscle of the at least one person to be tested under the target action type; The processing unit is used to perform signal normalization processing on the signal features of the at least one person to be tested based on the training label weight range to obtain the processed features.

6. The muscle strength prediction device based on FMG and sEMG signals according to claim 4, characterized in that: Also includes: A determination module, configured to determine a feature data set using the processed features before inputting the processed features into a target machine learning model, and divide the feature data set into a training set and a test set; An acquisition module, used for training a pre-built machine learning model using the training set before inputting the processed features into a target machine learning model to obtain a trained machine learning model; A generation module is used to input the test set into the trained machine learning model before inputting the processed features into the target machine learning model to output the model prediction results, and generate the target machine learning model when the model prediction results meet the preset test conditions.

7. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the muscle strength prediction method based on FMG and sEMG signals as described in any one of claims 1 to 3.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the muscle strength prediction method based on FMG and sEMG signals as described in any one of claims 1 to 3.

9. A computer program product, comprising a computer program, characterized in that The computer program is executed by a processor to implement the muscle strength prediction method based on FMG and sEMG signals as described in any one of claims 1 to 3.

Images