Signal processing method and device based on PPG signal of smart wristwatch

Abstract

The invention discloses a signal processing method and device based on a PPG signal of an intelligent wristwatch. The method comprises the steps that the PPG signal and acceleration information are acquired; the acceleration information comprises first acceleration information, second acceleration information and third acceleration information; filtering the PPG signal and the acceleration information to obtain first target signal information; and filtering the first target signal information to obtain second target signal information.

Description

Technical Field

[0001] The present invention relates to the field of signal processing technology, and in particular to a signal processing method and device based on a smart watch PPG signal. Background Art

[0002] At present, as a smart wearable device that is easy to wear and flexible to use, the new wristwatch is gradually gaining popularity among the public. It can not only track sports, navigate, pay, and provide notification reminders, but also can perform non-invasive real-time daily monitoring of human functions. Smart watches mainly use photoplethysmography (PPG) to collect human pulse waves for a long time and continuously, and then obtain physiological parameters such as human heart rate, blood pressure, blood oxygen saturation, and heart rhythm changes through analysis algorithms. In the acquisition stage of the PPG signal, the light will propagate through the skin tissue, part of it will be absorbed, and the other part will be reflected back to the photosensitive sensor. When the light emitted by the light-emitting diode enters the human body, due to the existence of a "pulse" in the human body - the hemodynamic mechanism, this will affect the absorption efficiency of the blood and cause the intensity of the light reflected by the skin to change accordingly. If these changes are captured by the photosensitive sensor, they will be converted into electrical signals consistent with heart activity. However, the PPG signal collected on the watch is easily affected by various noises. Therefore, a signal processing method and device are provided to improve the noise filtering efficiency and accuracy of the PPG signal, thereby improving the efficiency and accuracy of the PPG signal extraction. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a signal processing method and device based on the PPG signal of a smart watch, which is conducive to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0004] In order to solve the above technical problem, a first aspect of an embodiment of the present invention discloses a signal processing method, the method comprising:

[0005] Acquire a PPG signal and acceleration information; the acceleration information includes first acceleration information, second acceleration information and third acceleration information;

[0006] Filtering the PPG signal and the acceleration information to obtain first target signal information;

[0007] The first target signal information is filtered to obtain second target signal information.

[0008] A second aspect of an embodiment of the present invention discloses a signal processing device, the device comprising:

[0009] An acquisition module, used to acquire a PPG signal and acceleration information; the acceleration information includes first acceleration information, second acceleration information and third acceleration information;

[0010] A first processing module, configured to filter the PPG signal and the acceleration information to obtain first target signal information;

[0011] The second processing module is used to filter the first target signal information to obtain second target signal information.

[0012] A third aspect of the present invention discloses another signal processing device, the device comprising:

[0013] A memory storing executable program code;

[0014] a processor coupled to the memory;

[0015] The processor calls the executable program code stored in the memory to execute part or all of the steps in the signal processing method disclosed in the first aspect of the embodiment of the present invention.

[0016] The fourth aspect of the present invention discloses a computer-readable storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute some or all of the steps in the signal processing method disclosed in the first aspect of an embodiment of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 is a schematic diagram of a scenario of a signal processing system provided by an embodiment of the present invention;

[0019] Figure 2 It is a flowchart of a signal processing method disclosed in an embodiment of the present invention;

[0020] Figure 3 is a structural schematic diagram of a signal processing device disclosed in an embodiment of the present invention;

[0021] Figure 4 is a structural schematic diagram of another signal processing device disclosed in an embodiment of the present invention;

[0022] Figure 5 is a structural schematic diagram of a first filtering model disclosed in an embodiment of the present invention;

[0023] Figure 6 It is a schematic diagram of a filtering effect disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or equipment.

[0026] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0027] In this application, the word "exemplary" is used to mean "used as an example, illustration, or description." Any embodiment described in this application as "exemplary" is not necessarily to be construed as being preferred or advantageous over other embodiments. The following description is given to enable any technician in the field to implement and use the present application. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present application can be implemented without using these specific details. In other instances, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present application with unnecessary details. Therefore, the present application is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0028] It should be noted that since the method of the embodiment of the present application is executed in a computer device, the processing objects of each computer device exist in the form of data or information. For example, time is actually time information. It can be understood that if size, quantity, position, etc. are mentioned in subsequent embodiments, they are all corresponding data for processing by the computer device. The details will not be repeated here.

[0029] It should be noted that the artificial intelligence related technologies that may be involved in this application are briefly described. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that machines have the functions of perception, reasoning and decision-making.

[0030] Artificial intelligence technology is a comprehensive discipline that covers a wide range of fields, including both hardware-level and software-level technologies. The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0031] Computer Vision (CV) is a science that studies how to make machines "see". To put it more specifically, it refers to machine vision such as using cameras and computers to replace human eyes to identify and measure targets, and further perform graphic processing so that computer processing becomes an image that is more suitable for human eye observation or transmission to instrument detection. As a scientific discipline, computer vision studies related theories and technologies, and attempts to establish an artificial intelligence system that can obtain information from images or multi-dimensional data. Computer vision technology usually includes image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content / behavior recognition, three-dimensional object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous positioning and map construction, and other technologies, as well as common biometric recognition technologies such as face recognition and fingerprint recognition.

[0032] Unimodal information is data of only one type, such as text, image, audio, video, electromagnetic signal, etc. Multimodal information is data that includes at least two types of unimodal information. Furthermore, multimodal information is suitable for complex tasks that require the integration of multiple information sources, such as sentiment analysis, robot interaction, autonomous driving, etc. By integrating information from multiple modalities, higher performance and accuracy can usually be achieved on the task.

[0033] A large model refers to an artificial neural network model with a very large number of parameters. In the field of artificial intelligence, a large model generally refers to a model with hundreds of millions to trillions of parameters. Models usually need to be trained on large-scale data sets and require a large amount of computing resources to be optimized and adjusted. Large models are often used to solve complex tasks such as natural language processing, computer vision, and speech recognition. Generative AI is an AI that can create new content and ideas, including conversations, stories, images, videos, and music. In an embodiment of the present application, the large model may be ChatGPT, BERT, XLNet, Zhipu model, Claude, Moonshot AI model, ChatGLM model, Tongwen Qianyi model, MiniMax model, Spark model, Llama model, 360GPT model, Qwen model, Baichuan model, Skylark model, vivoLM model, and Wenxin Yiyan scale language models, which are not limited in the embodiments of the present application.

[0034] The embodiments of the present application provide a signal processing method, apparatus, computer device, and computer-readable storage medium, which are described in detail below.

[0035] See also Figure 1 , Figure 1 Schematic diagram of a signal processing system provided in an embodiment of the present application. The signal processing system may include a computer device 100, in which a signal processing device is integrated, such as Figure 1 Computer equipment in.

[0036] In the embodiment of the present application, the computer device 100 is mainly used to obtain PPG signals and acceleration information; the acceleration information includes first acceleration information, second acceleration information and third acceleration information;

[0037] Filtering the PPG signal and the acceleration information to obtain first target signal information;

[0038] The first target signal information is filtered to obtain second target signal information.

[0039] It can improve the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0040] In the embodiment of the present application, the computer device 100 may be an independent server, or a server network or server cluster composed of servers. For example, the computer device 100 described in the embodiment of the present application includes but is not limited to a computer, a network host, a single network server, a plurality of network server sets or a cloud server composed of a plurality of servers. The cloud server is composed of a large number of computers or network servers based on cloud computing.

[0041] It is understandable that the computer device 100 used in the embodiments of the present application may be a device including both receiving and transmitting hardware, that is, a device having receiving and transmitting hardware capable of performing two-way communication on a two-way communication link. Such a device may include: a cellular or other communication device having a single-line display or a multi-line display or a cellular or other communication device without a multi-line display. The specific computer device 100 may be a desktop terminal or a mobile terminal, and the computer device 100 may also be one of a mobile phone, a tablet computer, a laptop computer, etc.

[0042] Those skilled in the art will understand that Figure 1 The application environment shown in the figure is only one application scenario of the present application solution and does not constitute a limitation on the application scenario of the present application solution. Other application environments may also include Figure 1 More or less computer equipment as shown in Figure 1 Only one computer device is shown in the figure. It can be understood that the signal processing system can also include one or more other services, which are not specifically limited here.

[0043] In addition, if Figure 1 As shown, the signal processing system may further include a memory 200 for storing data, such as image data, position information, etc.

[0044] It should be noted that Figure 1 The scenario diagram of the signal processing system shown is merely an example. The signal processing system and scenario described in the embodiment of the present application are intended to more clearly illustrate the technical solution of the embodiment of the present application, and do not constitute a limitation on the technical solution provided in the embodiment of the present application. A person of ordinary skill in the art can appreciate that with the evolution of the signal processing system and the emergence of new business scenarios, the technical solution provided in the embodiment of the present application is equally applicable to similar technical problems.

[0045] The present invention discloses a signal processing method and device based on a smart watch PPG signal, which is beneficial to improving the noise filtering efficiency and accuracy of the PPG signal, thereby improving the PPG signal extraction efficiency and accuracy. The following are detailed descriptions.

[0046] Embodiment 1

[0047] See also Figure 2 , Figure 2 is a flow chart of a signal processing method disclosed in an embodiment of the present invention. Figure 2 The signal processing method described is applied to a management system, such as a local server or a cloud server for management, and the embodiments of the present invention are not limited thereto. Figure 2 As shown, the signal processing method may include the following operations:

[0048] 101. Obtain PPG signal and acceleration information.

[0049] In the embodiment of the present invention, the acceleration information includes first acceleration information, second acceleration information and third acceleration information.

[0050] 102. Filter the PPG signal and the acceleration information to obtain first target signal information.

[0051] 103. Filter the first target signal information to obtain second target signal information.

[0052] It should be noted that the above-mentioned PPG signal and acceleration information can be obtained through smart collection devices such as smart watches, such as Huawei bracelets, Xiaomi bracelets, etc., which are not limited in the embodiments of the present invention.

[0053] It should be noted that there are still many "burrs" in the first target signal information obtained by filtering the PPG signal and acceleration information. These "burrs" will affect the pulse wave feature extraction based on the PPG waveform, thereby reducing the accuracy of human body function parameter detection. The present application further filters the first target signal information to smooth and shape the signal, thereby facilitating the extraction of an ideal pulse wave, which is not limited in the embodiments of the present invention.

[0054] In this optional embodiment, as an optional implementation manner, the filtering process is performed on the first target signal information to obtain the second target signal information, including:

[0055] Performing multi-scale wavelet decomposition on the first target signal information to obtain signal frequency coefficient information; the signal frequency coefficient information includes first frequency coefficient information and second frequency coefficient information; the second frequency coefficient information includes a plurality of second sub-frequency coefficient information;

[0056] For any second sub-frequency coefficient information in the second frequency coefficient information, the frequency coefficient information in the second sub-frequency coefficient information is divided and classified to obtain a plurality of frequency coefficient information sets arranged in sequence; the frequency coefficient information includes a basic frequency coefficient and a coefficient time; each frequency coefficient information set includes a plurality of frequency coefficient information;

[0057] For any frequency coefficient information set, the maximum value of all basic frequency coefficients in the frequency coefficient information set is used as the frequency coefficient reference value corresponding to the frequency coefficient information set;

[0058] According to the order of the frequency coefficient information sets, the frequency coefficient reference values ​​corresponding to the first 6 frequency coefficient information sets in the order are used as reference reference values;

[0059] Using a first numerical update model, the frequency coefficient reference value corresponding to the frequency coefficient information set is updated;

[0060] Among them, the first numerical update model is:

[0061]

[0062] Where x1 represents the updated frequency coefficient reference value; x0 xh and x0 i Respectively represent the xhth and ith frequency coefficient reference values; a1 and a2 represent the first calculation coefficient and the second calculation coefficient respectively;

[0063] Performing square calculation processing on the frequency coefficient corresponding to the frequency coefficient information in the second sub-frequency coefficient information to obtain a coefficient square value corresponding to the frequency coefficient;

[0064] For any updated frequency coefficient reference value, taking the coefficient time corresponding to the updated frequency coefficient reference value as the reference time point, determining the update time interval corresponding to the updated frequency coefficient reference value;

[0065] Sum the coefficient square values ​​corresponding to all frequency coefficients within one timing time adjacent to the upper and lower limit times of the update time interval to obtain a first calculated square value corresponding to the updated frequency coefficient reference value;

[0066] Calculating the first calculated square value using the second numerical update model to obtain a second calculated square value;

[0067] Among them, the second numerical update model is:

[0068] x2=a3·pl·x3;

[0069] In the formula, x2 represents the second calculated square value; x3 represents the first calculated square value; pl represents the decomposition frequency corresponding to the second sub-frequency coefficient information; a3 represents the third calculated coefficient;

[0070] Using a third numerical update model, the coefficient square value corresponding to the update time interval corresponding to the updated frequency coefficient reference value is updated;

[0071] Among them, the third numerical update model is:

[0072]

[0073] In the formula, x4 represents the square value of the coefficient after the update; x5 represents the square value of the coefficient before the update;

[0074] Performing a square root calculation on the updated coefficient square value, and assigning the original positive and negative signs of the frequency coefficient of the frequency coefficient information corresponding to the coefficient square value to the square root value, to obtain the coefficient square root value;

[0075] The square root of the coefficient is used as a new frequency coefficient in the frequency coefficient information corresponding to the square value of the coefficient;

[0076] Signal reconstruction is performed on the first frequency coefficient information and all updated second sub-frequency coefficient information to obtain second target signal information.

[0077] It should be noted that the above-mentioned use of the first numerical update model to update the frequency coefficient reference value corresponding to the frequency coefficient information set can effectively analyze the changes in the PPG signal, so as to adjust and update the threshold in a timely manner, so that PPG signals of different amplitudes can be detected, thereby improving signal processing efficiency, which is not limited in the embodiments of the present invention.

[0078] It should be noted that the above-mentioned use of the third numerical update model to update the coefficient square value corresponding to the update time interval corresponding to the updated frequency coefficient reference value uses the square value corresponding to the average frequency coefficient of the signal interference neighborhood as the benchmark of the interference interval threshold, and replaces the value greater than the neighborhood average frequency coefficient, and does not process the value less than the neighborhood average frequency coefficient. At the same time, the second numerical update model is used to dynamically update the benchmark coefficient of each interference area, thereby achieving effective filtering of interference components to make the signal in the interference interval smoother, which is not limited to the embodiments of the present invention.

[0079] It should be noted that the multi-scale wavelet decomposition of the first target signal information can be performed by using the wavedec function to perform wavelet decomposition on the input normalized first target signal information. The function decomposes the PPG signal into components of different frequency bands. The decomposition level is controlled by the level parameter (set to 3 in the function), and the wavelet basis is specified by the waveletName parameter, which is not limited in the embodiment of the present invention.

[0080] It should be noted that the first frequency coefficient information and the second frequency coefficient information respectively represent the high-frequency signal component and the low-frequency signal component, which is not limited in the embodiment of the present invention.

[0081] It should be noted that the above-mentioned signal reconstruction of the first frequency coefficient information and all updated second sub-frequency coefficient information can be to reconstruct the denoised coefficients through the waverec function to generate a smoothed and shaped PPG signal waveform, which is not limited in the embodiment of the present invention.

[0082] It should be noted that the above timing time includes M coefficient times. Furthermore, M is not less than 6, which is not limited in the embodiment of the present invention.

[0083] It should be noted that the first calculation coefficient and the second calculation coefficient are values ​​between 0 and 1, and the sum of the two is 1, which is not limited in the embodiment of the present invention. Further, the third calculation coefficient is a positive number not greater than 0.1, which is not limited in the embodiment of the present invention.

[0084] It should be noted that if Figure 6 As shown, in order to verify the performance of the method of the present application, the frequency spectra under low-speed and high-speed motion states (i.e., weak motion artifacts and strong motion artifacts) were selected for simulation analysis, and in order to highlight the effectiveness of the method of the present application, the present application also compared the denoising results of the PPG signal in the same time window using two single denoising methods, RLS and NLMS, and a joint denoising method. In the state of weak motion artifacts, i.e., low-speed running, an original PPG signal of a time window was selected, and the time domain and frequency domain diagrams of the PPG signal were obtained after three denoising processes, i.e., the RLS filtering algorithm, the NLMS filtering algorithm, and the method of the present application. Figure 6 In a, the red vertical line is the position of the true heart rate obtained from the ECG data, and the blue circle is the position of the maximum spectral peak of the signal. It can be clearly seen that although the maximum spectral peak in the original PPG signal spectrum is not the spectral peak corresponding to the true heart rate due to the interference of motion artifacts, after denoising by the above three algorithms, the positions of the two become overlapping. This shows that in the state of weak motion artifacts, the PPG signal is less affected by motion artifacts, and the three denoising algorithms can achieve good filtering effects. Next, consider the spectrum of the original PPG signal, after RLS filtering, after NLMS filtering, and after filtering by the method of this application under strong motion artifacts. Three representative time windows are selected for analysis. The results are as follows: Figure 6 b. Figure 6 c. Figure 6 d. Figure 6It can be seen from b that the spectral peak corresponding to the true heart rate is not the maximum spectral peak of the original PPG signal. After RLS filtering, the positions of the two spectral peaks coincide, which means that the RLS filter can better remove motion artifacts and retain the spectral information of the heart rate. After NLMS filtering, the spectral peak corresponding to the true heart rate is still not the maximum spectral peak, which means that the true heart rate is still submerged by motion artifacts, and the denoising effect of the NLMS filter is not obvious at this time. When the method of the present application is used to process the PPG signal, the highest peak position in the spectrum is the spectral peak position corresponding to the true heart rate, which means that the method of the present application performs well in removing motion artifacts. From Figure 6 It can be seen from c that the spectrum peak corresponding to the true heart rate point after NLMS filtering coincides with the highest peak in the PPG spectrum, indicating that the NLMS filter has a good denoising effect in this time window. At this time, the denoising effect of the RLS filter is not obvious, and the spectrum peak corresponding to the true heart rate cannot be accurately identified. After filtering the PPG signal using the method of the present application, the true spectrum peak is found again, proving that the method of the present application performs well in removing motion artifacts. Figure 6 d, it can be seen that neither RLS nor NLMS filtering alone found the spectral peak corresponding to the true heart rate, and the location of the true heart rate point is not the highest peak in the PPG signal spectrum, but the second highest peak or even the third highest peak. This shows that the motion artifact has too much interference on the heart rate, and the spectral peak corresponding to the true heart rate is still submerged. However, the method of the present application denoises the PPG signal, and the true heart rate spectral peak will become more obvious, thereby finding the true spectral peak. From the experimental results, the denoising effect of the three methods is good in the state of weak motion artifacts, and the spectral peak corresponding to the true heart rate can be found. However, under the interference of strong motion artifacts, the first two denoising algorithms have large errors in the effect of removing motion artifacts, and the method of the present application can reduce this error. This is because under the interference of strong motion artifacts, after denoising by the first two denoising algorithms, the true spectral peak corresponding to the heart rate may not be the highest spectral peak in the PPG signal spectrum, but a secondary or auxiliary peak. The highest spectral peak after denoising by the method of the present application may just be staggered, such as Figure 6 b and Figure 6 c, it is also possible that it is not the highest peak at all, such as Figure 6 As shown in d, the highest spectrum peak of the method of the present application most likely corresponds to the location of the real heart rate.

[0085] It can be seen that implementing the signal processing method described in the embodiment of the present invention is conducive to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0086] In an optional embodiment, the filtering process of the PPG signal and the acceleration information to obtain the first target signal information includes:

[0087] Filtering the PPG signal using a filter to obtain first filtered signal information;

[0088] The first filtering model is used to filter the first filtering signal information and the acceleration information to obtain the first target signal information.

[0089] It should be noted that the filtering frequency of the above-mentioned filter is 0.5-5Hz to remove power frequency interference, electromyographic noise, and baseline drift, which is not limited in the embodiment of the present invention. Furthermore, baseline drift is mainly caused by human breathing, usually changes slowly, and the frequency does not exceed 0.5Hz. Power frequency interference usually comes from large equipment such as power grids. When the human body forms an electromagnetic field loop with these devices, power frequency interference will occur, and the frequency is usually 50Hz or 60Hz. Electromyographic noise is interference caused by human muscle activity, which will also be mixed into the PPG signal. Its frequency is mostly in the range of 10Hz to 300Hz. The normal heartbeat is 40 to 240 times per minute, which is equivalent to a frequency range of 0.7 to 4Hz. Therefore, this application uses a 0.5-5Hz filter to effectively filter the above three types of noise, which is not limited in the embodiment of the present invention.

[0090] It can be seen that implementing the signal processing method described in the embodiment of the present invention is conducive to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0091] In another optional embodiment, Figure 5 As shown, the first filtering model includes a first filtering module, a second filtering module, a third filtering module, a fourth filtering module, a fifth filtering module, a sixth filtering module and a fusion activation module; wherein,

[0092] The first input end of the first filter module is configured to receive the first model input information of the first filter model, the second input end of the first filter module is configured to receive the second model input information of the first filter model, and the output end of the first filter module is connected to the first input end of the second filter module; the second input end of the second filter module is configured to receive the second model input information of the first filter model, and the output end of the second filter module is connected to the first input end of the third filter module; the second input end of the third filter module is configured to receive the second model input information of the first filter model, and the output end of the third filter module is connected to the first input end of the fusion activation module; the first input end of the fourth filter module is configured to receive the first model input information of the first filter model, the second input end of the fourth filter module is configured to receive the second model input information of the first filter model, and the output end of the fourth filter module is connected to the first input end of the fifth filter module; the second input end of the fifth filter module is configured to receive the second model input information of the first filter model, and the output end of the fifth filter module is connected to the first model input information of the sixth filter model; the second input end of the sixth filter model is configured to receive the second model input information of the first filter model, and the output end of the sixth filter module is connected to the second input end of the fusion activation module; the output end of the fusion activation module is configured to output the model output information of the first filter model.

[0093] It should be noted that the model architectures of the above-mentioned first filtering module, second filtering module and third filtering module are consistent, and the embodiments of the present invention do not limit them. The model architectures of the fourth filtering module, the fifth filtering module and the sixth filtering module are consistent, and the embodiments of the present invention do not limit them. Furthermore, in the technical solution of the present application, the filtering modules constructed based on the first filtering function and the second filtering function are respectively subjected to three-level cascaded parallel filtering, and then the fusion activation module constructed based on the third filtering function performs convex combination filtering on the filtered signals of the two channels, thereby realizing deep filtering of motion artifacts and improving the filtering quality and efficiency of the signal, which is not limited by the embodiments of the present invention.

[0094] It should be noted that there are more or less some difficulties in the traditional methods of removing motion artifacts from PPG signals. For example, wavelet transform can provide a joint representation of time and frequency, but it is necessary to select an appropriate wavelet basis and wavelet decomposition level according to the signal, and the determination of the threshold also requires multiple experiments. Inappropriate selection may lead to the loss of important signal features, so it is not suitable for processing highly dynamic PPG signals. The empirical mode decomposition method is suitable for analyzing nonlinear and non-stationary signals without pre-setting basis functions, but the intrinsic mode functions IMFs generated during the decomposition process are not strictly orthogonal, so there may be problems with mode aliasing. In addition, it also has end effects, especially near the signal boundary, which may generate high-frequency noise or oscillation, resulting in instability of the decomposition results. These problems are particularly prominent in PPG signal processing, and its high computational complexity makes it unsuitable for real-time processing of PPG signals on embedded devices (watches).

[0095] Furthermore, the first filtering model of the present application uses adaptive filtering and three-axis acceleration signals to process motion artifacts in the PPG signal, so that the filtering calculation efficiency is high, the signal has the ability to be processed in real time, and the parameters can be dynamically adjusted according to the characteristics of the interference, thereby improving the filtering effect, which is not limited to the embodiments of the present invention.

[0096] It can be seen that implementing the signal processing method described in the embodiment of the present invention is conducive to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0097] In yet another optional embodiment, the second model input information is acceleration information; the acceleration direction of the first acceleration information, the acceleration direction of the second acceleration information, and the acceleration direction of the third acceleration information are perpendicular to each other;

[0098] The second model input information received by the second input end of the first filter module and the second input end of the fourth filter module is the first acceleration information in the acceleration information;

[0099] The second model input information received by the second input end of the second filter module and the fifth input end of the fourth filter module is the second acceleration information in the acceleration information;

[0100] The second model input information received by the second input end of the third filter module and the second input end of the sixth filter module is the third acceleration information in the acceleration information.

[0101] It should be noted that the above-mentioned first acceleration information, second acceleration information and third acceleration information respectively represent the acceleration in the X direction, the acceleration in the Y direction and the acceleration in the Z direction, which are not limited in the embodiments of the present invention. Furthermore, the acceleration information is introduced as reference information for filtering in the present application, considering that under the interference of motion artifacts, the spectrum peak corresponding to the real heart rate is submerged by the motion artifacts in the PPG signal, becoming the second largest spectrum peak in the PPG signal spectrum, and the main peak of the PPG signal corresponds to the main peak of the acceleration data of the x-axis, y-axis and z-axis, that is, the motion artifacts are closely related to the acceleration signal, and using acceleration as a reference signal to remove the motion artifacts in the PPG signal can improve the filtering quality and efficiency, which are not limited in the embodiments of the present invention.

[0102] It should be noted that in this application, the first acceleration information, the second acceleration information and the third acceleration information are used as reference signals of the first, second and third filter modules of two parallel filtering routes, respectively, to remove the artifacts caused by motion in the PPG signal according to the direction, thereby achieving full-dimensional artifact removal, which is not limited in the embodiment of the present invention. Furthermore, compared with the synthetic acceleration information obtained by square-summing the acceleration information in three directions and then taking the square root (based on the synthetic acceleration signal, only a single type of motion can be processed, but in real life, the motion state of the human body is very complex, such as running, fitness, etc., acceleration will be generated in three directions, and it is not known which specific direction of the acceleration signal will produce stronger motion artifacts), this application processes and analyzes the acceleration signals in three directions separately, which can more effectively cope with the complexity of the motion state and improve the filtering effect, which is not limited in the embodiment of the present invention.

[0103] It should be noted that the signals of the first filtering model at each moment in the present application are represented in vector form, such as the input signals of the first input terminal and the second input terminal of the first filtering module, the output signal of the output terminal of the first filtering module, the output signal of the output terminal of the fourth filtering module, etc., which are not limited in the embodiments of the present invention.

[0104] It can be seen that implementing the signal processing method described in the embodiment of the present invention is conducive to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0105] In yet another optional embodiment, the first filtering module is constructed based on the first filtering function:

[0106]

[0107] Where SC1 i and SC1 i-1 Respectively represent the i-th and i-1-th output signals of the output end of the first filter module; SR1 i and SR2i Respectively represent the i-th input signal of the first input terminal and the second input terminal of the first filter module; WC1 i Signal difference information representing the i-th input signal of the second input terminal and the first input terminal of the first filtering module; ZY1 i Characterizes the i-th gain parameter in the first filtering module; QZ1 i represents the i-th weight parameter in the first filtering module; λ represents the first filtering coefficient.

[0108] It should be noted that when the signal of the first filter module constructed based on the first filter function is input at each moment, the weight parameters and gain parameters of the first filter module will be continuously updated according to the observation value and the error signal, thereby gradually improving the performance of the filtering model, making it robust to noise interference, able to quickly converge to the optimal solution, and having a good filtering effect, which is not limited to the embodiments of the present invention.

[0109] It should be noted that the first filter coefficient may be set by the user or a default value given by the system, which is not limited in the embodiment of the present invention. Further, the first filter coefficient may be between 0.5 and 2, which is not limited in the embodiment of the present invention.

[0110] It can be seen that implementing the signal processing method described in the embodiment of the present invention is conducive to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0111] In an optional embodiment, the fourth filtering module is constructed based on the second filtering function:

[0112]

[0113] Where SC2 j Represents the j-th output signal of the output end of the fourth filter module; SR3 j and SR4 j Respectively represent the j-th input signal of the first input terminal and the second input terminal of the fourth filter module; WC2 represents the signal difference information of the j-th input signal of the second input terminal and the first input terminal of the fourth filter module; QZ2 j Characterizes the jth weight parameter in the fourth filtering module; ZY2 j represents the jth gain parameter in the fourth filtering module; α and β represent the second filtering coefficient and the third filtering coefficient respectively.

[0114] It should be noted that the second filter coefficient and the third filter coefficient may be set by the user or may be a default value given by the system, which is not limited in the embodiment of the present invention. Further, the third filter coefficient is between 0 and 2, which is not limited in the embodiment of the present invention. The third filter coefficient is a very small positive number, such as 0.00001, which is not limited in the embodiment of the present invention.

[0115] It should be noted that the gain parameters in the fourth filtering module constructed based on the second filtering function are dynamically adjusted, that is, standardized parameter updates can be performed during signal processing at each moment to overcome the problem of poor adaptability of fixed parameters (error and convergence constitute a contradictory relationship), so that the filtering module can more effectively adapt to different input conditions, which not only improves the stability of the filtering model, but also accelerates the convergence speed, which is not limited in the embodiments of the present invention.

[0116] It can be seen that implementing the signal processing method described in the embodiment of the present invention is conducive to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0117] In another optional embodiment, the fusion activation module is constructed based on the third filter function:

[0118]

[0119] Where Y l+1 and Y l Respectively represent the l+1th and lth output signals of the output end of the fusion activation module; SC3 l and SC4 l Respectively represent the first output signal of the third filter module and the sixth filter module; SC5 l Characterizes the intermediate processed signal corresponding to the lth output signal of the third filtering module and the sixth filtering module; γ l represents the lth fourth filter coefficient; ξ represents the fifth filter coefficient.

[0120] It should be noted that the fifth filter coefficient may be set by the user (eg, between 1 and 2), may be a default parameter given by the system, or may be determined based on historical data, and the embodiment of the present invention does not limit this.

[0121] It should be noted that the fifth filter coefficient is between 0 and 1, which is not limited in the embodiment of the present invention. Further, in the present application, the fourth filter coefficient value corresponding to the output signal of the third filter module is strictly controlled between 0 and 1, and the filter coefficient (1-fourth filter coefficient) corresponding to the output signal of the corresponding sixth filter module is also between 0 and 1, and their sum is 1, and the value of the fourth filter coefficient is adaptively updated in each iteration. Further, in the present application, the latter signal is updated by the previous signal. This weight update mechanism allows the system to dynamically adjust the output signal according to two errors (the error between a filter output and the final filter output, and the error between a filter output and another filter output), which can improve the filter's ability to adapt to signal changes, so that it can better face non-stationary signals and dynamic scenes, which is not limited in the embodiment of the present invention.

[0122] It can be seen that implementing the signal processing method described in the embodiment of the present invention is conducive to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0123] Embodiment 2

[0124] See also Figure 3 , Figure 3 is a schematic diagram of the structure of a signal processing device disclosed in an embodiment of the present invention. Figure 3 The described device can be applied to a management system, such as a local server or a cloud server for management, etc., which is not limited in the embodiments of the present invention. Figure 3 As shown, the device may include:

[0125] The acquisition module 201 is used to acquire the PPG signal and acceleration information; the acceleration information includes first acceleration information, second acceleration information and third acceleration information;

[0126] A first processing module 202 is used to filter the PPG signal and the acceleration information to obtain first target signal information;

[0127] The second processing module 203 is used to filter the first target signal information to obtain second target signal information.

[0128] It can be seen that implementation Figure 3 The described signal processing device is beneficial to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0129] In another optional embodiment, Figure 3 As shown, the PPG signal and the acceleration information are filtered to obtain the first target signal information, including:

[0130] Filtering the PPG signal using a filter to obtain first filtered signal information;

[0131] The first filtering model is used to filter the first filtering signal information and the acceleration information to obtain the first target signal information.

[0132] It can be seen that implementation Figure 3 The described signal processing device is beneficial to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0133] In yet another optional embodiment, Figure 3 As shown, the first filtering model includes a first filtering module, a second filtering module, a third filtering module, a fourth filtering module, a fifth filtering module, a sixth filtering module and a fusion activation module; wherein,

[0134] The first input end of the first filter module is configured to receive the first model input information of the first filter model, the second input end of the first filter module is configured to receive the second model input information of the first filter model, and the output end of the first filter module is connected to the first input end of the second filter module; the second input end of the second filter module is configured to receive the second model input information of the first filter model, and the output end of the second filter module is connected to the first input end of the third filter module; the second input end of the third filter module is configured to receive the second model input information of the first filter model, and the output end of the third filter module is connected to the first input end of the fusion activation module; the first input end of the fourth filter module is configured to receive the first model input information of the first filter model, the second input end of the fourth filter module is configured to receive the second model input information of the first filter model, and the output end of the fourth filter module is connected to the first input end of the fifth filter module; the second input end of the fifth filter module is configured to receive the second model input information of the first filter model, and the output end of the fifth filter module is connected to the first model input information of the sixth filter model; the second input end of the sixth filter model is configured to receive the second model input information of the first filter model, and the output end of the sixth filter module is connected to the second input end of the fusion activation module; the output end of the fusion activation module is configured to output the model output information of the first filter model.

[0135] It can be seen that implementation Figure 3 The described signal processing device is beneficial to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0136] In yet another optional embodiment, Figure 3 As shown, the second model input information is acceleration information; the acceleration direction of the first acceleration information, the acceleration direction of the second acceleration information and the acceleration direction of the third acceleration information are perpendicular to each other;

[0137] The second model input information received by the second input end of the first filter module and the second input end of the fourth filter module is the first acceleration information in the acceleration information;

[0138] The second model input information received by the second input end of the second filter module and the fifth input end of the fourth filter module is the second acceleration information in the acceleration information;

[0139] The second model input information received by the second input end of the third filter module and the second input end of the sixth filter module is the third acceleration information in the acceleration information.

[0140] It can be seen that implementation Figure 3 The described signal processing device is beneficial to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0141] In yet another optional embodiment, Figure 3 As shown, the first filtering module is constructed based on the first filtering function:

[0142]

[0143] Where SC1 i and SC1 i-1 Respectively represent the i-th and i-1-th output signals of the output end of the first filter module; SR1 i and SR2 i Respectively represent the i-th input signal of the first input terminal and the second input terminal of the first filter module; WC1 i Signal difference information representing the i-th input signal of the second input terminal and the first input terminal of the first filtering module; ZY1 i Characterizes the i-th gain parameter in the first filtering module; QZ1 i represents the i-th weight parameter in the first filtering module; λ represents the first filtering coefficient.

[0144] It can be seen that implementation Figure 3 The described signal processing device is beneficial to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0145] In yet another optional embodiment, Figure 3 As shown, the fourth filtering module is constructed based on the second filtering function:

[0146]

[0147] Where SC2 j Represents the j-th output signal of the output end of the fourth filter module; SR3j and SR4 j Respectively represent the j-th input signal of the first input terminal and the second input terminal of the fourth filter module; WC2 represents the signal difference information of the j-th input signal of the second input terminal and the first input terminal of the fourth filter module; QZ2 j Characterizes the jth weight parameter in the fourth filtering module; ZY2 j represents the jth gain parameter in the fourth filtering module; α and β represent the second filtering coefficient and the third filtering coefficient respectively.

[0148] It can be seen that implementation Figure 3 The described signal processing device is beneficial to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0149] In yet another optional embodiment, Figure 3 As shown, the fusion activation module is built based on the third filter function:

[0150]

[0151] Where Y l+1 and Y l Respectively represent the l+1th and lth output signals of the output end of the fusion activation module; SC3 l and SC4 l Respectively represent the first output signal of the third filter module and the sixth filter module; SC5 l Characterizes the intermediate processed signal corresponding to the lth output signal of the third filtering module and the sixth filtering module; γ l represents the lth fourth filter coefficient; ξ represents the fifth filter coefficient.

[0152] It can be seen that implementation Figure 3 The described signal processing device is beneficial to improving the efficiency and accuracy of PPG signal noise filtering, thereby improving the efficiency and accuracy of PPG signal extraction.

[0153] Embodiment 3

[0154] See also Figure 4 , Figure 4 is a structural diagram of another signal processing device disclosed in an embodiment of the present invention. Figure 4 The described device can be applied to a management system, such as a local server or a cloud server for management, etc., which is not limited in the embodiments of the present invention. Figure 4 As shown, the device may include:

[0155] A memory 301 storing executable program codes;

[0156] a processor 302 coupled to the memory 301;

[0157] The processor 302 calls the executable program code stored in the memory 301 to execute the steps in the signal processing method described in the first embodiment.

[0158] Embodiment 4

[0159] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the signal processing method described in the first embodiment.

[0160] Embodiment 5

[0161] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps in the signal processing method described in the first embodiment.

[0162] The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0163] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution can be essentially or partly contributed to the prior art in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0164] Finally, it should be noted that the signal processing method and device based on the PPG signal of a smart watch disclosed in the embodiment of the present invention discloses only the preferred embodiments of the present invention, which are only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A signal processing method, characterized in that: The method comprises: Acquire a PPG signal and acceleration information; the acceleration information includes first acceleration information, second acceleration information and third acceleration information; Filtering the PPG signal and the acceleration information to obtain first target signal information; The first target signal information is filtered to obtain second target signal information.

2. The signal processing method according to claim 1, characterized in that: The filtering the PPG signal and the acceleration information to obtain first target signal information includes: Filtering the PPG signal using a filter to obtain first filtered signal information; The first filtering signal information and the acceleration information are filtered using a first filtering model to obtain first target signal information.

3. The signal processing method according to claim 2, characterized in that: The first filtering model includes a first filtering module, a second filtering module, a third filtering module, a fourth filtering module, a fifth filtering module, a sixth filtering module and a fusion activation module; wherein, The first input end of the first filtering module is configured to receive the first model input information of the first filtering model, the second input end of the first filtering module is configured to receive the second model input information of the first filtering model, and the output end of the first filtering module is connected to the first input end of the second filtering module; the second input end of the second filtering module is configured to receive the second model input information of the first filtering model, and the output end of the second filtering module is connected to the first input end of the third filtering module; the second input end of the third filtering module is configured to receive the second model input information of the first filtering model, and the output end of the third filtering module is connected to the first input end of the fusion activation module; the first input end of the fourth filtering module is configured to Receive the first model input information of the first filtering model, the second input end of the fourth filtering module is configured to receive the second model input information of the first filtering model, and the output end of the fourth filtering module is connected to the first input end of the fifth filtering module; the second input end of the fifth filtering module is configured to receive the second model input information of the first filtering model, and the output end of the fifth filtering module is connected to the first model input information of the sixth filtering model; the second input end of the sixth filtering model is configured to receive the second model input information of the first filtering model, and the output end of the sixth filtering module is connected to the second input end of the fusion activation module; the output end of the fusion activation module is configured to output the model output information of the first filtering model.

4. The signal processing method according to claim 3, characterized in that: The second model input information is the acceleration information; the acceleration direction of the first acceleration information, the acceleration direction of the second acceleration information and the acceleration direction of the third acceleration information are perpendicular to each other; The second model input information received by the second input end of the first filtering module and the second input end of the fourth filtering module is the first acceleration information in the acceleration information; The second model input information received by the second input end of the second filtering module and the fifth input end of the fourth filtering module is the second acceleration information in the acceleration information; The second model input information received by the second input end of the third filtering module and the second input end of the sixth filtering module is the third acceleration information in the acceleration information.

5. The signal processing method according to claim 3, characterized in that: The first filtering module is constructed based on the first filtering function: Where SC1 i and SC1 i-1 Respectively represent the i-th and i-1-th output signals of the output end of the first filtering module; SR1 i and SR2 i Respectively represent the i-th input signal of the first input terminal and the second input terminal of the first filtering module; WC1 i Signal difference information representing the i-th input signal of the second input terminal and the first input terminal of the first filtering module; ZY1 i Characterizes the i-th gain parameter in the first filtering module; QZ1 i Characterizing the i-th weight parameter in the first filtering module; λ represents the first filter coefficient.

6. The signal processing method according to claim 3, characterized in that: The fourth filtering module is constructed based on the second filtering function: Where SC2 j Represents the j-th output signal of the output end of the fourth filtering module; SR3 j and SR4 j Respectively represent the j-th input signal of the first input terminal and the second input terminal of the fourth filter module; WC2 represents the signal difference information of the j-th input signal of the second input terminal and the first input terminal of the fourth filter module; QZ2 j Characterizes the jth weight parameter in the fourth filtering module; ZY2 j represents the jth gain parameter in the fourth filtering module; α and β represent the second filtering coefficient and the third filtering coefficient respectively.

7. The signal processing method according to claim 3, characterized in that: The fusion activation module is constructed based on the third filter function: Where Y l+1 and Y l Respectively represent the l+1th and lth output signals of the output end of the fusion activation module; SC3 l and SC4 l Respectively represent the first output signal of the third filter module and the sixth filter module; SC5 l Characterizes the intermediate processed signal corresponding to the lth output signal of the third filtering module and the sixth filtering module; γ l represents the lth fourth filter coefficient; ξ represents the fifth filter coefficient.

8. A signal processing device, characterized in that: The device comprises: An acquisition module, used to acquire a PPG signal and acceleration information; the acceleration information includes first acceleration information, second acceleration information and third acceleration information; A first processing module, configured to filter the PPG signal and the acceleration information to obtain first target signal information; The second processing module is used to filter the first target signal information to obtain second target signal information.

9. A signal processing device, characterized in that: The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the signal processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are called, they are used to execute the signal processing method according to any one of claims 1 to 7.

Images