Biological signal processing method and device, computer equipment and storage medium

Abstract

The invention relates to the field of signal processing, and discloses a biological signal processing method and device, computer equipment and a storage medium, and the method comprises the steps: collecting a biological signal, and carrying out the preliminary filtering of the biological signal, and obtaining a preliminary filtering signal; performing noise characteristic analysis on the preliminary filtering signal, selecting a proper filtering mode according to an analysis result, and designing filter parameters through fuzzy logic processing; inputting the filter parameters into an adaptive filter to process the signal so as to obtain a filtered biological signal; in combination with a de-noising convolutional neural network model, noise identification and fuzzy logic processing are performed on the filtered biological signals, and filter parameters are dynamically adjusted; and performing multi-round iterative optimization until a target filtering effect is obtained, and outputting an optimized biological signal. According to the biological signal processing method, complex noise in the biological signal is effectively removed, the filtering precision is improved, meanwhile, the signal integrity is kept, and a reliable basis is provided for subsequent biological signal analysis and processing.

Description

Technical Field

[0001] The present application relates to the field of signal processing technology, and in particular to a biological signal processing method, device, computer equipment and storage medium. Background Art

[0002] Biosignals (e.g., electrocardiogram, electroencephalogram, and electromyography) are widely used in medical diagnosis, health monitoring, and scientific research. However, during the acquisition process, biosignals are often interfered by power frequency interference, motion artifacts, and other noises, which reduces the signal-to-noise ratio and affects the accuracy of subsequent analysis and processing.

[0003] For biological signal denoising, existing technologies mainly include analog filters, digital filters and deep learning-based methods. Analog filters filter out noise of specific frequencies, such as power frequency interference, through hardware circuits such as band-stop, low-pass, and high-pass. Although it has strong real-time performance, its performance is easily affected by temperature changes or aging of electronic components, and its design is fixed, making it difficult to adapt to dynamically changing noise. Digital filters rely on methods such as wavelet transform and adaptive filters to process the collected signals. Among them, wavelet transform can perform multi-resolution analysis, but the selection of appropriate wavelet functions is complex and may introduce pseudo-harmonics. Although adaptive filters can dynamically adjust filter coefficients, they have problems with slow convergence speed and large steady-state error. In recent years, neural network methods based on deep learning have shown potential in noise modeling and removal, but their engineering implementation is still immature and there are certain difficulties in promoting their application.

[0004] Despite the breakthroughs in the above technologies, there are still problems in practical applications, such as insufficient dynamic adaptability, the contradiction between filtering accuracy and signal fidelity, and weak comprehensive processing capabilities for multiple noise sources. For example, traditional filters have difficulty following power frequency noise with frequency and phase changes, low-pass or high-pass filters may weaken the effective frequency components of the target signal, and a single filter performs poorly in a multi-noise mixed environment. In addition, both wavelet transform and deep learning methods have the problems of high implementation complexity and high requirements for hardware resources.

[0005] Therefore, how to efficiently remove power frequency interference and motion artifacts in the process of biological signal acquisition while maintaining signal integrity and significantly improving filtering flexibility, stability and accuracy has become an important issue that needs to be solved urgently. Summary of the invention

[0006] In view of this, the embodiments of the present application provide a biological signal processing method, device, computer equipment and storage medium, which can effectively solve the problems in the prior art such as insufficient dynamic adaptability of the filter, low noise removal accuracy, inability to fully guarantee signal integrity and weak comprehensive processing capability of multiple noises.

[0007] In a first aspect, an embodiment of the present application provides a biological signal processing method, comprising: Collect biological signals; Performing preliminary filtering processing on the collected biological signal to obtain a preliminary filtered signal; Performing noise characteristic analysis on the preliminary filtered signal, selecting a filtering method based on the analysis result, and designing parameters of the adaptive filter through fuzzy logic processing; Inputting the parameters of the adaptive filter into the adaptive filter for processing to obtain a filtered biological signal; The denoising convolutional neural network model is used to perform repeated noise identification and fuzzy logic processing on the filtered biological signal, and the parameters of the adaptive filter are iteratively adjusted until the target filtering effect is obtained, and the optimized biological signal is output.

[0008] In some embodiments, collecting biological signals includes: At least one surface-mounted electrode is attached to different parts of the biological body surface as a detection electrode to capture the biological signal, and the amplitude of the biological signal is amplified by a preamplifier.

[0009] In some embodiments, performing preliminary filtering on the collected biological signal to obtain a preliminary filtered signal includes: The frequency components below the first frequency threshold and above the second frequency threshold are filtered out from the collected biological signal to obtain the preliminary filtered signal.

[0010] In some embodiments, the performing noise characteristic analysis on the preliminary filtered signal, selecting a filtering method based on the analysis result, and designing the parameters of the adaptive filter through fuzzy logic processing includes: Performing noise characteristic analysis on the preliminary filtered signal to determine a corresponding filtering method; According to the filtering method, filtering is performed on the preliminary filtered signal; Performing sampling processing on the filtered preliminary filtered signal to remove noise in the preliminary filtered signal; Using a triangular fuzzy model to perform fuzzy processing on the preliminary filtered signal after noise removal, and converting the amplitude value of the signal into a membership value; According to the membership value, using a preset fuzzy rule to determine whether to filter the preliminary filtered signal again to output a fuzzy value; The fuzzy value is defuzzified, and the parameters of the adaptive filter are designed and output.

[0011] In some embodiments, judging whether to filter the preliminary filtered signal again by using a preset fuzzy rule according to the membership value includes: When the membership value is greater than a preset threshold, the preliminary filtered signal is determined to be noise and filtering is performed.

[0012] In some embodiments, inputting the parameters of the adaptive filter into the adaptive filter for processing to obtain a filtered biological signal includes: Inputting the parameters of the adaptive filter into the adaptive filter for signal processing; The adaptive filter adjusts the biological signal in real time according to the input parameters of the adaptive filter to filter out the noise components in the biological signal to obtain the filtered biological signal.

[0013] In some embodiments, the denoising convolutional neural network model is used to repeatedly perform noise identification and fuzzy logic processing on the filtered biological signal, and the parameters of the adaptive filter are iteratively adjusted until the target filtering effect is obtained, and the optimized biological signal is output, including: Inputting the filtered biological signal into the denoising convolutional neural network model to perform noise identification and fuzzy processing; Based on the processing results of the noise identification and the fuzzy processing, the filter parameters are dynamically adjusted; the optimization is iterated until the target filtering effect is obtained, and the optimized biological signal is output.

[0014] In a second aspect, an embodiment of the present application provides a biological signal processing device, comprising: A signal acquisition module, used for collecting biological signals; A filtering processing module, used for performing preliminary filtering processing on the collected biological signal to obtain a preliminary filtered signal; A parameter design module, used to analyze the noise characteristics of the preliminary filtered signal, select a filtering method based on the analysis results, and design the parameters of the adaptive filter through fuzzy logic processing; A signal acquisition module, used for inputting the parameters of the adaptive filter into the adaptive filter for processing to obtain a filtered biological signal; The target signal acquisition module is used to use a denoising convolutional neural network model to repeatedly perform noise identification and fuzzy logic processing on the filtered biological signal, iteratively adjust the parameters of the adaptive filter until the target filtering effect is obtained, and output the optimized biological signal.

[0015] In a third aspect, an embodiment of the present application provides a computer device, comprising a processor and a memory, wherein the memory stores a computer program, and the processor is used to execute the computer program to implement the biological signal processing method of the first aspect.

[0016] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, and when the computer program is executed on a processor, the biological signal processing method of the first aspect described above is implemented.

[0017] The embodiments of the present application have the following beneficial effects: The biological signal processing method of the present application can effectively remove low-frequency or high-frequency noise and retain the main components of the target signal by performing preliminary filtering on the collected biological signals, thereby obtaining a preliminary filtered signal. Subsequently, the noise characteristics of the preliminary filtered signal are analyzed, the most suitable filtering method is selected based on the analysis results, and the filter parameters are designed using fuzzy logic processing to ensure that the filter can dynamically adapt to the characteristics of different noises. The signal quality is further optimized through adaptive filter processing, and the filtered signal is repeatedly subjected to noise identification and fuzzy logic processing in combination with the DNCNN model. The filter parameters are dynamically adjusted by combining the powerful modeling ability of deep learning with the rapid decision-making ability of fuzzy logic. After multiple rounds of iterations, the filtering effect is continuously optimized, and finally complex noise (e.g., power frequency interference, motion artifacts) is effectively removed, and an optimized biological signal with a high signal-to-noise ratio is output. The biological signal processing method of the present application not only significantly improves the filtering accuracy, but also has the characteristics of flexibility, adaptability and high efficiency, can adapt to a variety of complex noise environments, and provides high-quality basic data for subsequent biological signal analysis and processing, and has important engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A flow chart of a biological signal processing method according to an embodiment of the present application is shown; Figure 2 A schematic diagram showing a design of filter parameters in a biological signal processing method according to an embodiment of the present application is shown; Figure 3 A schematic diagram showing the structure of a biological signal processing method according to an embodiment of the present application is shown; Figure 4A schematic diagram showing the working principle of a DNCNN model of a biological signal processing method according to an embodiment of the present application is shown; Figure 5 A schematic structural diagram of a biological signal processing device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0021] The components of the embodiments of the present application generally described and shown in the drawings herein may be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0022] Hereinafter, the terms "including", "having" and their cognates that can be used in various embodiments of the present application are intended only to indicate specific features, numbers, steps, operations, elements, components or a combination of the foregoing items, and should not be understood as first excluding the existence of one or more other features, numbers, steps, operations, elements, components or a combination of the foregoing items or increasing the possibility of one or more features, numbers, steps, operations, elements, components or a combination of the foregoing items. In addition, the terms "first", "second", "third" and the like are only used to distinguish descriptions and cannot be understood as indicating or implying relative importance.

[0023] Unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meanings as those generally understood by those skilled in the art to which the various embodiments of the present application belong. The terms (such as those defined in generally used dictionaries) will be interpreted as having the same meanings as the contextual meanings in the relevant technical field and will not be interpreted as having idealized meanings or overly formal meanings unless clearly defined in the various embodiments of the present application.

[0024] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0025] Considering the problems of insufficient dynamic adaptability of filters, low noise removal accuracy, insufficient signal integrity and weak comprehensive processing capability of multiple noises in the existing technology, a biological signal processing method is proposed. The biological signal is collected and preliminarily filtered to obtain a preliminary filtered signal. Then, the noise characteristics of the preliminary filtered signal are analyzed, and a suitable filtering method is selected based on the analysis results. The filter parameters are dynamically designed using the fuzzy logic method to adapt to different noise characteristics. Then, the designed filter parameters are input into the adaptive filter for signal processing to further remove complex noise. Finally, the denoising convolutional neural network model (DNCNN) is combined to perform noise identification and fuzzy logic processing on the filtered signal, and the filter parameters are dynamically adjusted by iterative optimization until the target filtering effect is obtained, and a high-quality optimized biological signal is output. This method not only significantly improves the accuracy of noise removal, but also effectively guarantees the integrity of the signal and the comprehensive processing capability of multiple noises. It has strong flexibility and adaptability and can meet the needs of biological signal processing in complex noise environments.

[0026] Figure 1 A flow chart of a biological signal processing method according to an embodiment of the present application is shown. Exemplarily, the method includes steps S100 to S500.

[0027] Step S100, collecting biological signals.

[0028] Exemplarily, the collection of biological signals is achieved through detection electrodes. Biosignals generally refer to signals generated by electrophysiological activities inside and outside the human body, such as electrocardiogram (ECG), electromyogram (EMG), etc. In order to obtain these signals, electrodes are required to be attached to the surface of the human body to collect the electrical signals generated by the human body. Usually, the amplitude of biological signals is small and needs to be amplified for subsequent processing. The collection of biological signals can be carried out in a variety of ways, and the appropriate collection technology can be selected according to the specific application scenario.

[0029] In this embodiment, the collected biological signals are electrical signals, representing the electrophysiological activities in the organism. The collection process relies on suitable detection electrodes, which can sense the slight current changes on the surface of the organism and convert them into electronic signals. The collected signals will serve as the basis for subsequent signal processing and can provide important data for further filtering, analysis, processing, etc.

[0030] In an optional embodiment, in step S100, collecting biological signals includes: In order to ensure the quality of the signal captured from the surface of the organism, at least one surface-mounted electrode is used as a detection electrode in this embodiment. The surface-mounted electrode is attached to the surface of the organism, such as the skin, in a non-invasive manner to collect the electrical signal of the organism. The surface-mounted electrode is usually flexible and has good conductivity, which can ensure good contact with the skin surface and will not cause any discomfort to the subject.

[0031] According to the needs of biological signal collection, surface-mount electrodes are usually attached to multiple different parts of the biological body surface. For example, taking the forearm electromyographic signal collection as an example, the detection electrodes with preamplification function are placed in five different positions of the forearm muscles, namely the radial flexor carpi, ulnar flexor carpi, long flexor pollicis, deep flexor digitorum and extensor digitorum. The location of these electrodes is selected based on the distribution characteristics of muscle electrophysiological activities, which can capture signals from different muscle parts to obtain more comprehensive electromyographic signal data. The spacing between each two collection electrodes is 20mm, and the collection electrodes are output through shielded wires to reduce the impact of external interference on the signal.

[0032] Since the amplitude of biological signals is usually small, directly collected signals may be difficult to analyze effectively. In order to solve this problem, the present embodiment uses a preamplifier on each electrode to amplify the amplitude of the signal. The role of the preamplifier is to amplify the amplitude of the signal to a range suitable for subsequent processing to avoid information loss due to the signal being too weak. The preamplifier usually has a high input impedance and a low output impedance, which can effectively reduce the noise in the signal acquisition process and improve the clarity and stability of the signal.

[0033] In the above manner, the surface-mounted electrodes and the preamplifier work together to ensure that biological signals of sufficient strength can be stably and clearly collected from the surface of the organism, laying the foundation for subsequent signal processing and analysis. Specifically, in this embodiment, by arranging the detection electrodes with preamplification at multiple locations of the forearm muscles, it is not only possible to capture muscle electrical signals at different locations, but also to ensure the quality and strength of signal collection, providing a reliable data source for subsequent signal analysis.

[0034] Step S200, performing preliminary filtering processing on the collected biological signal to obtain a preliminary filtered signal.

[0035] Exemplarily, the purpose of performing preliminary filtering on the collected biological signals is to remove some unnecessary components in the signals, such as power frequency noise, motion artifacts, etc. Biological signals usually contain useful physiological information and various noise interferences. Especially during the collection process, environmental noise and motion artifacts may affect the signal quality. Therefore, in order to obtain clearer and more accurate signals, it is necessary to perform preliminary filtering on the signals to eliminate these unnecessary noise components.

[0036] The purpose of preliminary filtering is to pre-process the signal through a certain filtering method to make it suitable for subsequent signal analysis and processing. The role of the filter here is to selectively attenuate or remove certain frequency components in the signal, thereby achieving the effect of reducing noise. The preliminary filtered signal is the signal processed at this stage. It will be clearer and purer than the original signal, and can provide a better foundation for subsequent signal processing.

[0037] In an optional embodiment, in step S200, the collected biological signal is subjected to preliminary filtering processing to obtain a preliminary filtered signal, including: In this embodiment, the specific filtering method is to use a bandpass filter. A bandpass filter is a frequency selective filter that allows signals within a certain frequency range to pass through while suppressing signals outside the range. In biological signal processing, especially in signal processing such as electromyography (EMG) and electroencephalography (EEG), noise in certain frequency ranges (for example, power frequency noise) will affect the analysis of the signal.

[0038] The working principle of the bandpass filter is to define the passband range by setting two thresholds: the first frequency threshold and the second frequency threshold. These two thresholds represent the lowest and highest frequency ranges that the signal is allowed to pass. Frequency components below the first frequency threshold will be filtered out, and frequency components above the second frequency threshold will also be suppressed. In this way, the bandpass filter can effectively remove low-frequency and high-frequency noise in the signal, retaining only frequency components that are useful for biological signal analysis, so that the filtered signal becomes a preliminary filtered signal.

[0039] In this embodiment, the passband frequency range of the active bandpass filter is set to 20Hz-500Hz, which is implemented by a hardware circuit. In the hardware design, the resistance values ​​of the filter resistors R1 and R2 are 40kΩ and 2.5kΩ respectively, the values ​​of the capacitors C1 and C2 are both 0.1μF, and the operational amplifier uses the OPA227 chip. This design can ensure that the filter has good frequency selectivity and stability.

[0040] After the filtering operation is completed, the signal is converted into digital signal by the STM32 chip, and the analog signal is quantized into digital signal for use by the subsequent signal processing unit. This hardware implementation method ensures the accuracy and efficiency of signal filtering.

[0041] For example, in electromyography (EMG) signal processing, it is usually necessary to retain the frequency range of 20Hz to 500Hz and remove low-frequency noise below 20Hz and high-frequency noise above 500Hz. The active bandpass filter can effectively remove these unwanted noises according to the set passband range, obtain a preliminary filtered signal, make the signal clearer, and provide more reliable basic data for further analysis.

[0042] Through bandpass filtering, combined with precise hardware design, it is possible to ensure that only the frequency components of useful signals are retained and irrelevant noise is eliminated, thereby providing cleaner and more accurate signal data for subsequent processing and analysis.

[0043] Step S300 , analyzing the noise characteristics of the preliminary filtered signal, selecting a filtering method based on the analysis result, and designing the parameters of the adaptive filter through fuzzy logic processing.

[0044] In this embodiment, in order to optimize the processing of the preliminary filtered signal, the noise characteristics of the signal are first analyzed. By analyzing the statistical characteristics of the noise (for example, Gaussian distribution, uniform distribution or random distribution), the main characteristic parameters of the noise (for example, mean, variance or frequency distribution) are identified, so as to select a suitable filtering method, for example, single-value filtering, multi-value filtering or adaptive filtering. Subsequently, the adaptive filter parameters are dynamically designed in combination with the fuzzy logic processing method. Fuzzy logic processing can convert the signal amplitude value into a membership value, use preset fuzzy rules to effectively judge the noise components in the signal, and calculate the target parameters of the filter (for example, cutoff frequency, gain or filtering range) through the membership value. Finally, the designed filter parameters can significantly improve the noise removal effect in the subsequent filtering process, while retaining the integrity and accuracy of the signal to the maximum extent.

[0045] In an optional embodiment, in step S300, noise characteristics analysis is performed on the preliminary filtered signal, a filtering method is selected based on the analysis result, and parameters of the adaptive filter are designed through fuzzy logic processing, including: In this embodiment, if Figure 2 As shown, in order to achieve efficient denoising of the preliminary filtered signal and optimal design of the adaptive filter, the following steps are specifically performed: Perform noise characteristic analysis on the preliminary filtered signal, and determine the type of noise in the signal and its main characteristic parameters (e.g., variance and mean) by analyzing the statistical characteristics of the noise (e.g., Gaussian distribution or uniform distribution). Based on the analysis results of the noise characteristics, select a suitable filtering method, such as multi-value filtering for Gaussian noise and single-value filtering for uniform noise.

[0046] The preliminary filtered signal is filtered according to the selected filtering method, for example, a multi-value filter is used to smooth the signal, thereby reducing the interference of random noise.

[0047] The preliminary filtered signal after filtering is sampled and the remaining noise is further removed by extracting the characteristic value of the signal. For example, the signal is smoothed by the method of average sampling and periodic sampling, and the influence of nonlinear noise on the signal is reduced.

[0048] The triangular fuzzy model is used to perform fuzzy processing on the sampled signal. Specifically, the amplitude value of the signal is converted into a membership value, where the membership value indicates the possibility that the signal corresponds to a specific noise feature.

[0049] According to the membership value, the preset fuzzy rules are used for judgment. For example, if the membership value is higher than a certain threshold, it is judged that there may still be noise in the signal, and it needs to be filtered again; otherwise, the fuzzy value is directly output. The tuning parameters of the filter are obtained through fuzzy reasoning.

[0050] The fuzzy values ​​are defuzzified and the final parameters of the adaptive filter are calculated and output. For example, the filter cutoff frequency or gain range is calculated through defuzzification and used to dynamically adjust the filter to achieve the final optimization of the signal.

[0051] Through the above specific steps, it is possible to effectively cope with different types of noise characteristics and realize dynamic adjustment of filter parameters, thereby improving filtering accuracy and signal-to-noise ratio of the signal, and providing a reliable data basis for subsequent signal analysis.

[0052] In an optional implementation, according to the membership value, a preset fuzzy rule is used for judgment, including: In this embodiment, a decision mechanism based on fuzzy logic is used for further processing of the preliminary filtered signal. Specifically, during the processing, the signal is analyzed using fuzzification and fuzzy reasoning methods to determine whether it is noise and whether further filtering is required. One of the key steps in the decision-making process is the calculation of the membership value, which reflects whether the signal meets the noise characteristics.

[0053] First, after the previous filtering step, the preliminary filtered signal will be fuzzified to convert its amplitude value into a membership value. The main function of fuzzification is to transform the data of the input signal into the corresponding membership, which is used to analyze the degree to which the signal belongs to a certain category (for example, noise or a valid signal). In this embodiment, a triangular fuzzy model is used to fuzzify the filtered signal. Specifically, the function definition of the triangular fuzzy model can map the signal amplitude value to the membership value interval [0,1]. Among them, 0 means that the signal does not belong to noise at all, and 1 means that the signal belongs to noise completely. Since the signal output by the previous level is already a digital signal, the value of the amplitude mean is 0 or 1, and these values ​​are fuzzified through the triangular fuzzy model to generate corresponding membership values.

[0054] Next, according to the membership values, fuzzy inference is performed using preset fuzzy rules. The specific method is to calculate the membership values u0 and u1 of the signal amplitude sampling value to 0 and 1, where u0 represents the membership value of the sampling value to amplitude 0, and u1 represents the membership value of the sampling value to amplitude 1. When u0 > u1, the system outputs 0 during defuzzification, indicating that the signal is relatively clean; when u0 < u1, the defuzzification output is 1, indicating that the signal contains relatively large noise and needs further filtering processing.

[0055] To achieve more efficient filtering operations, supplementary rules are designed to determine the maximum filtering noise threshold value. Specifically, when the rightmost parameter of the 0 membership function of the filtering noise is equal to the leftmost parameter of the 1 membership function, it is the maximum filtering threshold value. In this embodiment, the maximum threshold value of the triangular distribution is set to 0.204. In addition, by sampling multiple times within a period and combining single-value filtering, the signal can be further smoothed and periodic noise components can be removed.

[0056] For example, in practical applications, assuming that the preset membership threshold is 0.5, when the membership value of the signal exceeds this threshold, the system determines that the signal contains strong noise components and needs further filtering processing; while when the membership value is less than the threshold, it is considered that the signal is relatively clean and does not require further processing. This dynamic decision-making mechanism based on membership values can flexibly adjust the filtering strategy, thereby effectively removing noise from the signal.

[0057] Through the above method, this embodiment can combine the processing methods of fuzzification, fuzzy inference, and defuzzification to accurately judge and process the noise in the signal. Finally, the filtered signal output by the system has a higher signal-to-noise ratio and better quality, can adapt to the signal processing requirements in various noise environments, and has strong flexibility and adaptability.

[0058] Step S400, input the parameters of the adaptive filter into the adaptive filter for processing to obtain the filtered biological signal.

[0059] Exemplarily, the design of the filter parameters is based on the analysis of the noise characteristics of the preliminarily filtered signal and the results of fuzzy logic inference. These parameters are designed to optimize the performance of the filter to meet the requirements of complex noise environments in biological signal processing. The filter and the adaptive filter are the same filter, and the adaptive filter can dynamically adjust its behavior according to the input parameters, thereby achieving efficient signal processing.

[0060] The unity of the filter and the adaptive filter ensures a simplified design of the system, that is, the core algorithm and physical structure of the filter remain unchanged, but the parameters are adjusted in real time to adapt to different noise characteristics. By inputting the designed filter parameters into the adaptive filter, the filter can further process the biological signal according to these parameters, filter out the noise in the signal, and retain useful biological information.

[0061] For example, if the input filter parameters specify a bandpass filter range (e.g., 20Hz to 500Hz), the adaptive filter will perform filtering operations within this range to remove low-frequency noise below 20Hz and high-frequency noise above 500Hz while retaining meaningful signal components. Through this process, the adaptive filter can effectively improve the quality of biological signals and lay the foundation for subsequent signal analysis.

[0062] In an optional embodiment, in step S400, the parameters of the adaptive filter are input into the adaptive filter for processing to obtain a filtered biological signal, including: In the specific implementation process, the filter parameters are input into the adaptive filter in real time. These parameters include but are not limited to the type of filter (e.g., bandpass filter or low-pass filter), cutoff frequency (e.g., high-frequency and low-frequency thresholds), gain, etc. These parameters are determined by the previous fuzzy logic analysis and noise characteristic analysis, and are intended to optimize the performance of the filter so that it can adapt to the dynamic changes of the signal.

[0063] The core feature of an adaptive filter is its ability to adjust its behavior in real time based on input parameters. For example, when power frequency noise (e.g., 50 Hz) is detected in a signal, the adaptive filter can automatically adjust its filtering parameters to specifically suppress the 50 Hz frequency band without affecting other frequency components. When the noise characteristics change (e.g., more random noise), the adaptive filter can adapt to the new noise environment by dynamically adjusting the gain or frequency range.

[0064] During the signal processing process, the adaptive filter continuously updates and optimizes the filtering parameters to ensure the filtering effect. For example, the filter analyzes the frequency distribution and intensity of the input signal in each sampling cycle and adjusts the cutoff frequency range so that it can filter out noise more accurately. This dynamic adjustment mechanism ensures that the filter can always efficiently process the noise in the biological signal and output a higher quality signal.

[0065] Finally, after being processed by the adaptive filter, the filtered biological signals are output. These signals not only remove the interference components, but also retain the useful information in the original signals to the greatest extent, providing higher quality data input for subsequent signal analysis.

[0066] Step S500, using the denoising convolutional neural network model to repeatedly perform noise identification and fuzzy logic processing on the filtered biological signal, iteratively adjust the parameters of the adaptive filter until the target filtering effect is obtained, and output the optimized biological signal.

[0067] To further improve the signal quality, the filtered biological signal is processed again using the Denoising Convolutional Neural Network (DNCNN). Denoising Convolutional Neural Network is a signal optimization method based on deep learning that can accurately model and remove complex noise. It is particularly suitable for nonlinear or random noise that is difficult to handle with traditional filtering methods. The following introduces the principle and workflow of the model. Figure 4 As shown: 1. The filtered biological signal x is input into the DNCNN model. The goal is to calculate the denoised signal v by learning the noise signal n, and the relationship is: v = x − n; Among them, x represents the noisy signal after filtering, n is the residual noise signal predicted by the model, and v is the output signal after denoising.

[0068] 2. Through end-to-end training, the DNCNN model learns the residual noise in the signal and optimizes the denoising effect through residual connections. The calculation formula for each layer of the network is:

[0069] in, Represents convolution and deconvolution operations, represents the Lth neuron in the ith input layer. represents the scale of the filter, represents the bias value, and f represents the activation function; The model predicts the noise signal through the residual learning mechanism. The formula of residual learning is: x=F(x1)+x1 Where F(x1) represents the residual prediction function; 3. The length of each layer of signal is dynamically adjusted according to the step size S, and the calculation formula is:

[0070] in: Indicates the length of the Lth layer signal; S: step size.

[0071] 4. Each layer output is processed by batch normalization, the formula is:

[0072] Where: BNγ,β: is the batch normalization function; γ,β: are trainable normalization parameters.

[0073] 5. During the optimization of the model, the discriminator of the generative adversarial network (GAN) is combined to compare the denoised signal and the original clean signal. Through the feedback mechanism, the network parameters of DNCNN are adjusted to optimize the denoising effect.

[0074] 6. Based on DNCNN, the image signal denoising is performed in combination with the weighted anisotropic diffusion model. The reconstructed signal is classified by the discriminator, and the signal tuning process is gradually completed under the feedback of the discriminator, thereby achieving the best denoising effect.

[0075] 7. Through multiple rounds of iterations, the DNCNN model continuously adjusts the filter parameters to optimize the denoising effect of the signal. The model dynamically adjusts the diffusion coefficient threshold K and stops updating the threshold K until the denoising effect is optimal.

[0076] Through the above process, the DNCNN model can accurately identify and remove complex noise (such as power frequency interference and motion artifacts) in biological signals, dynamically optimize filter parameters, and significantly improve the signal-to-noise ratio of the signal. The final output optimized signal not only has significantly reduced noise components, but also retains the integrity of the original signal to the greatest extent, ensuring the reliability and accuracy of subsequent signal analysis.

[0077] In an optional embodiment, in step S500, the denoising convolutional neural network model is used to perform repeated noise identification and fuzzy logic processing on the filtered biological signal, and the parameters of the adaptive filter are iteratively adjusted until the target filtering effect is obtained, and the optimized biological signal is output, including: In order to further improve the quality of the biological signal after filtering, the denoising convolutional neural network model (DNCNN) is used to perform noise identification and fuzzy processing on the signal. First, the filtered biological signal is input into the DNCNN model, and its deep learning ability is used to identify the residual noise in the signal. At the same time, the signal amplitude value is fuzzified through fuzzy logic processing, the membership value of the signal is calculated, and the signal is classified and analyzed in combination with the preset fuzzy rules. Based on the results of noise identification and fuzzy processing, the parameters of the filter, such as the cutoff frequency, filter gain or noise threshold value, are dynamically adjusted to adapt to the characteristics of the current signal.

[0078] This process uses an iterative optimization method, that is, after each noise identification and fuzzy processing, the filter parameters are updated and the biological signal is processed again to gradually reduce the residual noise and improve the signal-to-noise ratio and quality of the biological signal. The target filtering effect refers to the effect of removing noise, retaining signal integrity and improving the signal-to-noise ratio by dynamically adjusting the adaptive filter parameters in multiple rounds of iterations. Through multiple cycles of processing, the filtering effect is continuously optimized under the synergy of the DNCNN model and the fuzzy logic system until the noise is fully removed and the quality of the biological signal reaches the preset standard. Finally, the output biological signal has a higher signal-to-noise ratio and integrity, providing reliable basic data for subsequent analysis and processing.

[0079] The method of the embodiment of the present application realizes efficient collection and dynamic optimization filtering of biological signals by combining fuzzy logic and denoising convolutional neural network model (DNCNN), such as Figure 3 As shown in the figure: First, the collected biological signals are preprocessed by preliminary filtering to eliminate some low-frequency and high-frequency noises and preliminarily improve the signal quality. Secondly, the filter parameters are dynamically designed through noise characteristic analysis and fuzzy logic processing, so that the filter can be adaptively adjusted for different noise types (for example, Gaussian noise, random noise or uniformly distributed noise), thereby improving the filtering accuracy. At the same time, the adaptive filter uses the designed parameters for dynamic adjustment to further filter out complex noise components. Finally, the DNCNN model is introduced to perform repeated noise identification and fuzzy logic processing on the filtered signal, and the residual noise is accurately identified through the deep learning ability of the neural network, and the filter parameters are iteratively optimized to form a closed-loop optimization mechanism. This scheme can not only effectively denoise in complex and diverse noise environments and significantly improve the signal-to-noise ratio of the signal, but also dynamically adjust the filtering strategy to adapt to the signal processing requirements under different noise conditions. The final output biological signal has high accuracy and integrity, providing a reliable foundation for subsequent biological signal analysis and processing. This method is highly flexible and adaptable, and has important practical application value.

[0080] Figure 5 A schematic diagram of the structure of a biological signal processing device according to an embodiment of the present application is shown. Exemplarily, the device includes: A signal acquisition module 51, used for acquiring biological signals; A filtering processing module 52 is used to perform preliminary filtering processing on the collected biological signal to obtain a preliminary filtered signal; A parameter design module 53, for performing noise characteristic analysis on the preliminary filtered signal, selecting a filtering method based on the analysis result, and designing the parameters of the adaptive filter through fuzzy logic processing; A signal acquisition module 54, used for inputting the parameters of the adaptive filter into the adaptive filter for processing to obtain a filtered biological signal; The target signal acquisition module 55 is used to use a denoising convolutional neural network model to repeatedly perform noise identification and fuzzy logic processing on the filtered biological signal, iteratively adjust the parameters of the adaptive filter until the target filtering effect is obtained, and output the optimized biological signal.

[0081] It can be understood that the device of this embodiment corresponds to the method of the above embodiment, and the options in the above embodiment are also applicable to this embodiment, so they will not be described repeatedly here.

[0082] The present application also provides a computer device. Exemplarily, the computer device includes a processor and a memory, wherein the memory stores a computer program, and the processor runs the computer program, thereby enabling the computer device to execute the functions of each module in the above method or the above apparatus.

[0083] Among them, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including a central processing unit (CPU), a graphics processing unit (GPU), a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or at least one of other programmable logic devices, discrete gates or transistor logic devices, and discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc., which can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of the present application.

[0084] The memory may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electric erasable programmable read-only memory (EEPROM), etc. The memory is used to store a computer program, and the processor may execute the computer program accordingly after receiving an execution instruction.

[0085] The present application also provides a computer-readable storage medium for storing the computer program used in the above-mentioned computer device. For example, the computer-readable storage medium may include but is not limited to: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0086] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and structure diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in an alternative implementation, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the structure diagram and / or the flow diagram, and the combination of boxes in the structure diagram and / or the flow diagram, can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0087] In addition, the functional modules or units in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0088] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a smart phone, a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application.

[0089] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A biological signal processing method, characterized in that: The method comprises: Collect biological signals; Performing preliminary filtering processing on the collected biological signal to obtain a preliminary filtered signal; Performing noise characteristic analysis on the preliminary filtered signal, selecting a filtering method based on the analysis result, and designing parameters of the adaptive filter through fuzzy logic processing; Inputting the parameters of the adaptive filter into the adaptive filter for processing to obtain a filtered biological signal; The denoising convolutional neural network model is used to perform repeated noise identification and fuzzy logic processing on the filtered biological signal, and the parameters of the adaptive filter are iteratively adjusted until the target filtering effect is obtained, and the optimized biological signal is output.

2. The biological signal processing method according to claim 1, characterized in that: The collecting of biological signals comprises: At least one surface-mounted electrode is attached to different parts of the biological body surface as a detection electrode to capture the biological signal, and the amplitude of the biological signal is amplified by a preamplifier.

3. The biological signal processing method according to claim 1, characterized in that: The performing preliminary filtering processing on the collected biological signal to obtain a preliminary filtered signal includes: The frequency components below the first frequency threshold and above the second frequency threshold are filtered out from the collected biological signal to obtain the preliminary filtered signal.

4. The biological signal processing method according to claim 1, characterized in that: The performing of noise characteristic analysis on the preliminary filtered signal, selecting a filtering method based on the analysis result, and designing the parameters of the adaptive filter through fuzzy logic processing includes: Performing noise characteristic analysis on the preliminary filtered signal to determine a corresponding filtering method; According to the filtering method, filtering is performed on the preliminary filtered signal; Performing sampling processing on the filtered preliminary filtered signal to remove noise in the preliminary filtered signal; Using a triangular fuzzy model to perform fuzzy processing on the preliminary filtered signal after noise removal, and converting the amplitude value of the signal into a membership value; According to the membership value, using a preset fuzzy rule to determine whether to filter the preliminary filtered signal again to output a fuzzy value; The fuzzy value is defuzzified, and the parameters of the adaptive filter are designed and output.

5. The biological signal processing method according to claim 4, characterized in that: The step of using a preset fuzzy rule to determine whether to filter the preliminary filtered signal again according to the membership value includes: When the membership value is greater than a preset threshold, the preliminary filtered signal is determined to be noise and filtering is performed.

6. The biological signal processing method according to claim 1, characterized in that: The step of inputting the parameters of the adaptive filter into the adaptive filter for processing to obtain a filtered biological signal comprises: Inputting the parameters of the adaptive filter into the adaptive filter for signal processing; The adaptive filter adjusts the biological signal in real time according to the input parameters of the adaptive filter to filter out the noise components in the biological signal to obtain the filtered biological signal.

7. The biological signal processing method according to claim 1, characterized in that: The denoising convolutional neural network model is used to repeatedly perform noise identification and fuzzy logic processing on the filtered biological signal, and the parameters of the adaptive filter are iteratively adjusted until the target filtering effect is obtained, and the optimized biological signal is output, including: Inputting the filtered biological signal into the denoising convolutional neural network model to perform noise identification and fuzzy processing; Based on the processing results of the noise identification and the fuzzy processing, the filter parameters are dynamically adjusted; the optimization is iterated until the target filtering effect is obtained, and the optimized biological signal is output.

8. A biological signal processing device, characterized in that: The device comprises: A signal acquisition module, used for collecting biological signals; A filtering processing module, used for performing preliminary filtering processing on the collected biological signal to obtain a preliminary filtered signal; A parameter design module, used to analyze the noise characteristics of the preliminary filtered signal, select a filtering method based on the analysis results, and design the parameters of the adaptive filter through fuzzy logic processing; A signal acquisition module, used for inputting the parameters of the adaptive filter into the adaptive filter for processing to obtain a filtered biological signal; The target signal acquisition module is used to use a denoising convolutional neural network model to repeatedly perform noise identification and fuzzy logic processing on the filtered biological signal, iteratively adjust the parameters of the adaptive filter until the target filtering effect is obtained, and output the optimized biological signal.

9. A computer device, characterized in that: The computer device comprises a processor and a memory, wherein the memory stores a computer program, and the processor is configured to execute the computer program to implement the biological signal processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The device stores a computer program, which, when executed on a processor, implements the biological signal processing method according to any one of claims 1 to 7.

Images