Respiration detection system and method, electronic equipment and storage medium

Abstract

The invention relates to the technical field of respiration detection, in particular to a respiration detection system and method, electronic equipment and a storage medium, and the system comprises a PPG sensor used for obtaining original PPG data; the wavelet transformation module is used for performing signal decomposition on the original PPG data and extracting frequency band data related to respiration; and the respiration rate prediction module is used for performing feature extraction and respiration rate prediction on the frequency band information related to respiration to obtain a target detection result. The wavelet transformation module can effectively extract frequency band data related to respiration from PPG signals, the decomposition can distinguish physiological fluctuations of different frequencies caused by heartbeat and respiration, particularly, low-frequency changes caused by respiration can be accurately captured, and then the accuracy of respiratory rate prediction is improved; and the accuracy of the target detection result is ensured.

Description

Technical Field

[0001] The present application relates to the field of respiratory detection technology, and in particular to a respiratory detection system, method, electronic device and storage medium. Background Art

[0002] Respiration detection is a key component in physiological monitoring and is widely used in critical care, sleep monitoring, and real-time health tracking. Traditional volume- or pressure-based methods used in clinical settings include handheld spirometers, masks with flow sensors, and piezoelectric pressure-based respiratory sensors. Alternatively, respiratory chest belts or straps are used to detect the movement caused by breathing. Although these methods provide high accuracy, they do not support simultaneous monitoring of other vital signs. In addition, their bulky and inconvenient design makes them unsuitable for long-term use, such as continuous monitoring during sleep or physical activity.

[0003] In order to detect breathing signals non-invasively while ensuring patient comfort, photoplethysmography (PPG) has been widely proposed as an alternative method for breathing detection. PPG is a non-invasive optical technology that measures changes in blood volume within blood vessels. The method works by illuminating the skin, usually using near-infrared or red light-emitting diodes. The sensor detects the changes in light intensity caused by the pulsation of blood through the blood vessels and extracts the breathing signal from the modulation of the pulse. PPG provides a portable, continuous breathing detection solution while also being able to monitor other physiological signals such as heart rate and blood oxygen saturation, blood sugar levels, and arterial stiffness.

[0004] PPG is used to detect breathing through several methods. The simplest method directly measures the modulation associated with breathing by analyzing the changes in the signal peaks, but this method is highly sensitive to noise and motion artifacts, which can greatly reduce its effectiveness in real-world scenarios. Frequency-based methods, such as empirical mode decomposition and holographic-Hilbert spectroscopy, are used to isolate specific breathing signals from PPG signals, but they still require a large computational load when implemented in hardware. In addition, frequency-based methods are unable to extract deeper and more complex features in the signal, limiting their ability to capture subtle changes caused by factors such as skin tone, age, and gender, which can lead to inaccuracies in PPG vital sign detection. Summary of the invention

[0005] In order to solve the problem of inaccuracy in the above-mentioned PPG vital signal detection, the present application provides a breathing detection system, method, electronic device and storage medium, which can be applied to complex application scenarios and have accurate detection.

[0006] In a first aspect, the present application provides a breathing detection system, the system comprising:

[0007] PPG sensor, used to obtain raw PPG data;

[0008] A wavelet transform module, used to perform signal decomposition on the original PPG data and extract frequency band data related to breathing;

[0009] The breathing rate prediction module is used to perform feature extraction and breathing rate prediction on the frequency band information related to breathing to obtain a target detection result.

[0010] In some embodiments of the present application, the PPG sensor includes a light source component, a photodetector, a signal amplifier, and an analog-to-digital converter.

[0011] In some embodiments of the present application, obtaining raw PPG data includes:

[0012] The light source assembly emits light, and the photoelectric detector receives a reflection signal of the emitted light;

[0013] amplifying the reflected signal by the signal amplifier;

[0014] The amplified reflected signal is converted into digital form by the analog-to-digital converter to obtain original PPG data.

[0015] In some embodiments of the present application, performing signal decomposition on the original PPG data to extract frequency band data related to breathing includes:

[0016] Dividing the original PPG data into even-indexed samples and odd-indexed samples;

[0017] Signal decomposition is performed based on the even-numbered index samples and the odd-numbered index samples to extract frequency band data related to breathing.

[0018] In some embodiments of the present application, the performing signal decomposition based on the even-indexed samples and the odd-indexed samples to extract frequency band data related to breathing includes:

[0019] The first step is to calculate the difference between the odd-indexed sample and the even-indexed sample to obtain the detail coefficient;

[0020] The second step is to calculate the average value of the odd index samples and the even index samples to obtain the approximate coefficient;

[0021] In the third step, the approximation coefficients are divided into even-indexed samples and odd-indexed samples, and the first step and the second step are sequentially repeated until the frequency band data related to breathing is extracted.

[0022] In some embodiments of the present application, the breathing rate prediction module includes a convolutional layer, a pooling layer and a fully connected layer.

[0023] In some embodiments of the present application, the step of extracting features and predicting the breathing rate of the frequency band information related to breathing to obtain a target detection result includes:

[0024] Extracting features of the frequency band information related to breathing through the convolution layer to obtain a feature map;

[0025] Performing a pooling operation on the feature map through a pooling layer;

[0026] The results after the pooling operation are integrated and analyzed through the fully connected layer to obtain a respiratory rate prediction value, which is used as the target detection result.

[0027] A second aspect of the present application provides a breathing detection method, the method comprising:

[0028] Get raw PPG data;

[0029] Performing signal decomposition on the original PPG data to extract frequency band data related to breathing;

[0030] Feature extraction and respiratory rate prediction are performed on the frequency band information related to breathing to obtain a target detection result.

[0031] A third aspect of the present application provides an electronic device, including a memory and a processor, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the breathing detection method described in the embodiment.

[0032] A fourth aspect of the present application provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the breathing detection method as described in the embodiment is implemented.

[0033] The technical solution provided in the embodiments of the present application has at least the following technical effects or advantages:

[0034] The breathing detection system in each embodiment of the present application includes: a PPG sensor for acquiring raw PPG data; a wavelet transform module for performing signal decomposition on the raw PPG data and extracting frequency band data related to breathing; a breathing rate prediction module for performing feature extraction and breathing rate prediction on the frequency band information related to breathing to obtain a target detection result. The present application implements an efficient frequency-based method by applying a wavelet transform module, filters out key information related to the breathing signal, significantly reduces the amount of data and improves the accuracy of neural network training. The PPG sensor can efficiently collect raw PPG data, and the wavelet transform module is a module with high time domain localization and high computational efficiency, which is used to efficiently process raw PPG data, and can obtain data closely related to breathing, and its multi-resolution characteristics enable it to simultaneously capture frequency bands for calculating blood oxygen content and heart rate. Compared with other computationally intensive frequency-based technologies, the wavelet transform module is more efficient in real-time signal processing on resource-constrained devices because it can be implemented in hardware without multipliers or dividers. Unlike traditional discrete wavelet transforms, the wavelet transform module does not rely on convolution, but uses a lifting scheme to decompose the signal into different frequency bands.

[0035] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0037] Figure 1 is a structural schematic diagram of a breathing detection system in an exemplary embodiment of the present application;

[0038] Figure 2 is a structural schematic diagram of another breathing detection system in an exemplary embodiment of the present application;

[0039] Figure 3 It is a schematic diagram comparing signals collected by a breathing sensor and a PPG sensor in an exemplary embodiment of the present application;

[0040] Figure 4 is a schematic diagram of an experimental test environment setting in an exemplary embodiment of the present application;

[0041] Figure 5 is a schematic diagram of an experimental result in an exemplary embodiment of the present application;

[0042] Figure 6is a schematic diagram of the steps of a breathing detection method in an exemplary embodiment of the present application;

[0043] Figure 7 It is a schematic diagram of the structure of an electronic device provided by an exemplary embodiment of the present application.

[0044] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. DETAILED DESCRIPTION

[0045] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0046] Respiration detection is a very important part of physiological monitoring and is widely used in critical care, sleep monitoring, and real-time health tracking. Traditionally, respiratory monitoring is achieved through volume or pressure-based methods, including handheld spirometers, masks with flow sensors, and piezoelectric pressure respiratory sensors. In addition, there are also respiratory chest belts or belts that detect movement during breathing. Although these traditional methods have high accuracy, they cannot monitor other vital signs at the same time. At the same time, their bulky and inconvenient design makes them unsuitable for long-term use, such as continuous monitoring during sleep or physical activity. To solve these problems, photoplethysmography (PPG) has been proposed as an alternative means of respiratory detection. PPG is a non-invasive optical technology that obtains data by measuring changes in blood volume in blood vessels. The technology usually uses near-infrared light or red light-emitting diodes to illuminate the skin, and the sensor then detects the changes in light intensity caused by blood pulsation and extracts respiratory signals from it. PPG is not only portable and supports continuous respiratory detection, but also can simultaneously monitor other physiological signals such as heart rate, blood oxygen saturation, and even blood sugar levels and arterial stiffness.

[0047] When PPG detects respiratory signals, the simplest way is to directly measure the modulation related to breathing by analyzing the change in the peak value of the signal, but this method is very sensitive to noise and motion artifacts, which may lead to reduced effectiveness in practical applications. Using frequency-based methods such as empirical mode decomposition and holographic-Hilbert spectrum, specific respiratory signals can be separated from PPG signals, but these methods require a large amount of calculation when implemented in hardware. In addition, frequency-based methods have limitations in extracting deep and complex features of signals, so they cannot accurately capture subtle changes caused by skin tone, age and gender, which may affect the accuracy of PPG life signal detection.

[0048] To solve the above technical problems, the present application provides a breathing detection system in an embodiment, such as Figure 1 As shown, the system includes: a PPG sensor 101 for acquiring raw PPG data; a wavelet transform module 102 for performing signal decomposition on the raw PPG data and extracting frequency band data related to breathing; and a respiratory rate prediction module 103 for performing feature extraction and respiratory rate prediction on the frequency band information related to breathing to obtain a target detection result.

[0049] The PPG sensor plays a key role in this system, and is used to capture physiological variables related to respiratory activity. In a specific implementation, the PPG sensor includes a light source component, a photodetector, a signal amplifier, and an analog-to-digital converter. The acquisition of raw PPG data includes: emitting light through the light source component, receiving the reflected signal of the emitted light through the photodetector; amplifying the reflected signal through the signal amplifier; and performing analog-to-digital conversion on the amplified reflected signal through the analog-to-digital converter to obtain raw PPG data.

[0050] Specifically, Figure 2 As shown, LED is a light source component. For example, red light or near-infrared light generated by LED is irradiated onto the user's skin. After the light is absorbed by the skin and blood vessels, part of the light is reflected back to the photodetector in the PPG sensor, i.e. Figure 2 The photodiode (PD) in the blood vessel is used to detect the intensity of the reflected light, which changes with the pulsation of blood flow. The reflected light signal is very weak and is amplified by the signal amplifier. Figure 2 The amplified analog signal is amplified by the transimpedance amplifier (TIA) in the Figure 2 The ADC in the image is converted into a digital signal. This step converts the analog voltage signal into a digital form to facilitate subsequent digital processing. Afterwards, the digital signal is transmitted to a computing device (such as a smartphone or medical monitoring device) via Bluetooth or other wireless technology for further processing. In the embodiment of the present application, the raw PPG data in digital form is transmitted to the wavelet transform module, i.e. Figure 2 Medium LWT.

[0051] Figure 2 In the analog front end, a digitally controlled LED emits light onto the skin, and a photodiode (PD) detects the reflected signal. The raw signal is processed by a transimpedance amplifier (TIA), converted to a digital signal by an analog-to-digital converter (ADC), and transmitted to a digital device via Bluetooth. Figure 2The lower part shows the data processing flow: LWT is used to decompose the signal into frequency bands. The results of LWT can be further input into a convolutional neural network (CNN) to extract features and classify and identify respiratory events (e.g., normal breathing, breath holding, and apnea events). CNN is used to further process the signal to determine the respiratory rate and detect apnea. In addition, LWT can also isolate the frequency bands of SpO2 and heart rate (HR) at the same time, making this system an ideal candidate for low-power, multi-functional wearable biosensor integration.

[0052] In a possible implementation, the original PPG data is subjected to signal decomposition to extract frequency band data related to breathing, including: dividing the original PPG data into even index samples and odd index samples; performing signal decomposition based on the even index samples and odd index samples to extract frequency band data related to breathing. It can be seen that the wavelet transform module classifies the original PPG data according to even and odd indexes through advanced algorithms and processes them separately. This method effectively stratifies the changing characteristics of the time series signal by frequency and extracts the frequency bands related to breathing in a more targeted manner.

[0053] Specifically, performing signal decomposition based on the even-indexed samples and the odd-indexed samples to extract frequency band data related to breathing includes:

[0054] The first step is to calculate the difference between the odd-indexed samples and the even-indexed samples to obtain the detail coefficients. This step captures high-frequency components, including some noise and rapidly changing signal components. This is a particularly effective method for analyzing short-term dynamic changes.

[0055] The second step: Calculate the average of the even-indexed samples and the odd-indexed samples to obtain the approximate coefficient, which represents the low-frequency component of the signal. The key is to capture longer-cycle physiological changes, such as the breathing cycle.

[0056] Step 3: The approximation coefficients are further decomposed into even indexes and odd indexes, and the previous details and approximation calculations are repeated layer by layer until a sufficient decomposition depth is reached to extract the frequency band data related to breathing. This multi-level decomposition can progressively extract the specific frequency band features related to different breathing states from the PPG data layer by layer.

[0057] It can be understood that the wavelet transform module, namely LWT, performs frequency band decomposition on the original PPG signal and accurately extracts frequency band data related to breathing. The module first divides the original PPG data into even-indexed samples and odd-indexed samples. Then, by calculating the difference and mean between samples, two coefficients are obtained: detail coefficient and approximate coefficient. The detail coefficient represents higher-frequency changes and mainly captures rapid physiological changes, such as fluctuations caused by heartbeats; while the approximate coefficient represents low-frequency changes, such as slow blood volume changes caused by breathing. It can be seen that by using LWT to process the PPG signal, while removing high-frequency noise, fast and low-power extraction of blood oxygen, heart rate and breathing signals is achieved.

[0058] This decomposition process is iterative. In each iteration, the system further decomposes the approximate coefficients, gradually refining them to lower frequency bands until accurate data reflecting the breathing-related frequency bands are obtained. This series of decompositions ultimately forms a set of frequency bands from high to low resolution, each of which reflects changes in physiological activities on different time scales. In addition, the multi-resolution analysis properties of wavelet transform are particularly suitable for processing physiological signals with non-stationary properties, such as PPG signals. This method not only reduces data redundancy, but also retains the key features of the signal, which is helpful for subsequent feature extraction and analysis.

[0059] The processing flow of the wavelet transform module includes signal decomposition, step-by-step processing, and frequency band selection. When the signal is decomposed, Figure 2 As shown in the figure, the original PPG data is decomposed at multiple levels using the wavelet transform module, namely LWT. Each level of decomposition divides the signal into different frequency bands: low frequency band (approximation coefficient) and high frequency band (detail coefficient). The approximation coefficients of each level can be further decomposed to extract more detailed low-frequency components, while the high-frequency components contain detailed information related to heartbeat and breathing. The frequency range of each decomposition level in LWT can be calculated based on the sampling frequency Fs. The Nyquist frequency is F Nyquist =Fs / 2, the frequency range of the detail coefficient at level L is:

[0060]

[0061] Among them, High represents high frequency and Low represents low frequency. Finally, specific frequency bands (such as specific levels of detail coefficients) are selected, which are considered to be closely related to the breathing pattern because they contain signals of periodic blood volume changes caused by breathing.

[0062] Figure 3 is a schematic diagram comparing signals collected by a breathing sensor and a PPG sensor in an exemplary embodiment of the present application, Figure 3 Figure 2 shows the real breathing signal collected from the breathing sensor electrodes (top), and the raw data collected from the NIR channel of the PPG sensor (bottom), corresponding to Figure 3 The piezoelectric chip data and PPG infrared channel data in the figure. The carrier wave in the signal is generated by blood pulsation and is usually a more dominant signal, while the envelope wave is the modulation of the blood pulse signal caused by breathing.

[0063] In another specific implementation, the respiratory rate prediction module includes a convolution layer, a pooling layer, and a fully connected layer. Feature extraction and respiratory rate prediction are performed on the frequency band information related to breathing to obtain a target detection result, including: extracting features from the frequency band information related to breathing through the convolution layer to obtain a feature map. Specifically, the convolution layer processes the frequency band data after wavelet transformation, extracts key time-frequency features through multiple filters, and generates a feature map. The feature map is pooled by the pooling layer, and the pooling layer reduces the spatial size of the feature map, which helps to reduce overfitting and improve the generalization ability of the model. The fully connected layer further integrates the pooled data, uses the activation function to increase the nonlinear processing capability, and finally outputs the predicted respiratory rate. In other words, the results after the pooling operation are integrated and analyzed through the fully connected layer to obtain a respiratory rate prediction value, and the respiratory rate prediction value is used as the target detection result, and the target detection result can be normal breathing, breath holding and suffocation events, etc.

[0064] It can be understood that the breathing rate prediction module further processes the frequency band data obtained through wavelet transform and uses convolutional neural network (CNN) for deep learning analysis. First, the convolution layer of this module extracts features from the frequency band information through multiple filters to generate feature maps, which focus on key time-frequency features that may contain breathing rate information. Next, the pooling layer downsamples the feature map to reduce the data dimension and computational complexity, while enhancing the model's insensitivity to small changes in the input data, which helps improve prediction accuracy and prevent overfitting. Finally, the fully connected layer integrates all features and adds nonlinearity through activation functions such as ReLU to output the final predicted breathing rate.

[0065] Overall, through the design and implementation of this system, accurate detection of respiratory rate can be achieved, and the system has low power consumption, high accuracy and good user portability. For long-term monitoring, such as sleep quality analysis and daily health management, such a system provides a very practical solution.

[0066] In order to verify the technical effect of the breathing detection system described in this application, this application also conducted a series of experimental tests. Figure 4 This is a schematic diagram of the experimental test environment setting. Figure 4The configuration of the test environment is shown, including a dual-channel PPG sensor located on the forehead (red) and a reference sensor placed at the nostril (blue), with the ground electrode fixed on the chin. In the experiment, a total of six participants (four males and two females) were tested. Each participant performed four trials. In each trial, the participant was asked to breathe normally for 30 seconds, hold his breath for 30 seconds, and then breathe again for one minute while data was collected. The experiment was conducted by wearing a PPG sensor and a piezoelectric pressure-based respiration sensor on the forehead, which senses changes in air pressure at the nostrils while breathing. The data of the PPG sensor is collected at 232Hz, while the data of the respiration sensor is collected at 1000Hz. In order to enhance comfort and reduce the impact of movement, Bluetooth modules are implemented in both the PPG sensor and the respiration sensor, and a soft bandage is used to firmly adhere the sensors to the skin.

[0067] A total of 24 trials were collected in the experiment, with an average of 4 trials per participant. Four channels of PPG data were collected for each trial, including two red and two near infrared (NIR) channels, and one data channel from the pressure sensor (left and right nostril). The experimental results were collected and analyzed using LWT. The PPG sensor used for breathing detection consists of four channels, using two sets of LEDs, each including one red (660nm) and one NIR (940nm). A piezoelectric pressure sensor was placed under the nose as a reference breathing sensor. It can measure the breathing rate by detecting small pressure changes caused by air flow during inhalation and exhalation. The experiment was verified by neural network training and reference data for results.

[0068] Figure 5 is a schematic diagram of an experimental result in an exemplary embodiment of the present application, such as Figure 5 As shown in the figure, the results of the PPG sensor and the breathing sensor show a significant drop in the power of the breathing signal between 25 seconds and 65 seconds, which is the period when the participants were asked to hold their breath. The power detected by the PPG sensor is not completely consistent with the readings of the dedicated breathing sensor. This difference occurs because the breathing signal in the PPG data is ultimately demodulated from the mixed signal, in which the blood pulse signal dominates. In addition, PPG data is more sensitive to noise due to its complex biological characteristics. Therefore, deep learning methods need to be introduced to better extract information from these data.

[0069] Therefore, 18 trials were randomly selected for training, 3 for each participant to increase variability. Each trial contained 4 sets of data from four PPG sensor channels, each containing 120 seconds of testing at 232Hz (about 27,840 data points). To further increase the potential of the training data, the data was enhanced by introducing jitter, amplitude modulation, and harmonic modulation. Six data sets were retained for each participant for validation.

[0070] To demonstrate the system’s ability to monitor respiration, its performance was evaluated using three metrics. First, Breathing Rate Per Minute (BRM), defined as the number of breaths per minute, was used as a gold standard metric for comparison with state-of-the-art wearable respiratory sensors. Second, for the specific application of sleep monitoring and apnea detection, successful breath-hold detection measures the effectiveness of the algorithm in identifying breath-holds within a 30 to 60 second window. A breath-hold is defined as the signal amplitude falling below a threshold, which is set to half the mean of the absolute signal values. If the duration is 20 seconds or more, it is considered 100% successful; shorter durations receive a success percentage based on the detection time relative to 20 seconds. The third metric, false apnea detection, is designed to evaluate instances of false breath-hold detections outside the window. A false detection occurs when the signal remains below the threshold for more than 5 consecutive seconds. Each instance outside the specified window is counted as a false detection, indicating that the low activity rate is incorrectly interpreted as a breath-hold. These conditions allow for the identification of true breath-hold events as well as false detections independent of the reference or true signal. As the threshold is reduced, condition 1 will decrease (not ideal) and condition 2 will also decrease (ideal), and vice versa.

[0071] For example Figure 5 As shown, in this experiment, the signal was collected at a frequency of 232Hz, so the theoretical optimal range for calculating respiration is the 8th to 9th level of the detail coefficient, corresponding to the frequency range of 0.45Hz to 0.91Hz for the 8th level and 0.23Hz to 0.45Hz for the 9th level, respectively. However, in actual operation, the lower frequencies are susceptible to motion artifacts, while the higher frequencies are dominated by the blood pulsation signal modulated by the respiratory signal and may contain more useful information. Through waveform inspection and comparison of the cross-correlation between each level (individual and combined layer summation) and the reference signal, it was found that the summation of detail coefficients of layers 5 to 7 is the most robust in our setting. Since each layer of the LWT reduces the total number of samples by half, the lower layers are upsampled to align with the 5th layer. Compared with the original PPG data, the summation output is reduced to 1 / 32 of the original size while retaining the critical low-frequency information required for analyzing the breathing pattern.

[0072] In addition to using high-level information to extract low-frequency signals of breathing, we also use the middle layer (3-6 layers) to extract information about heart rate and blood oxygen. This range is somewhat intertwined with the range of breathing detection, mainly to leave a certain margin (headroom), because the subsequent calculation steps themselves can handle a small amount of noise. The amplitude of the heart rate signal is also the amplitude of the AC signal of PPG, which is defined as the change in reflected light intensity caused by pulse changes. The amplitude of the breathing signal extracted previously is also the DC signal of PPG, which is the reflected light caused by non-pulse such as skeletal muscle and skin, and is also modulated by low-frequency signals such as breathing. After the heart rate frequency band is extracted, we mainly confirm the frequency of this signal by moving average and calculating the peaks and valleys, thereby determining the heart rate. In addition, we use the formula R = (ACr / DCr) / (ACir / DCir) to calculate the R value, and calculate the arterial blood oxygen SPO2 by fitting. Among them, ACr is the AC signal of red light, DCr is the DC signal of red light, and ACir and DCir are the AC and DC signals of infrared.

[0073] It can be seen that by integrating PPG sensing technology with wavelet transform and deep learning analysis, the breathing detection system provided by this application can continuously monitor the breathing state without contacting the user's body, reduce user discomfort, and improve the continuity of data collection. In particular, wavelet transform not only reduces data redundancy, but also retains the key features of the signal, which is helpful for subsequent feature extraction and analysis, thereby ensuring the accuracy of subsequent PPG vital signal detection. The low-power design and wireless data transmission function of this system make it have a wide range of application prospects, especially suitable for medical and health fields that require long-term monitoring.

[0074] In some embodiments of the present application, a breathing detection method is also provided, such as Figure 6 As shown, the method includes: S1, obtaining original PPG data; S2, performing signal decomposition on the original PPG data to extract frequency band data related to breathing; S3, performing feature extraction and respiratory rate prediction on the frequency band information related to breathing to obtain a target detection result. The respiratory detection method is a hybrid method that enhances the measurement of respiratory power through a wearable PPG device and integrates a wavelet transform (LWT) and a three-layer convolutional neural network (CNN).

[0075] The S2 step, i.e., LWT, provides a computationally efficient, multi-resolution, and hardware-compatible method for processing raw PPG data while significantly reducing the data load of the neural network. The training of the CNN enhances the system's ability to produce robust results, ultimately improving the accuracy of respiratory power measurement. The S2 step includes: dividing the raw PPG data into even-indexed samples and odd-indexed samples; performing signal decomposition based on the even-indexed samples and the odd-indexed samples, and extracting frequency band data related to breathing. Performing signal decomposition based on the even-indexed samples and the odd-indexed samples, extracting frequency band data related to breathing, includes: a first step, calculating the difference between the odd-indexed samples and the even-indexed samples to obtain detail coefficients; a second step, calculating the average of the odd-indexed samples and the even-indexed samples to obtain approximate coefficients; a third step, dividing the approximate coefficients into even-indexed samples and odd-indexed samples, and repeating the first and second steps in sequence until the frequency band data related to breathing is extracted.

[0076] Through the above multi-level decomposition, the frequency band data related to breathing can be effectively extracted from the PPG signal. This decomposition can distinguish the physiological fluctuations of different frequencies caused by heartbeat and breathing, and in particular, can accurately capture the low-frequency changes caused by breathing.

[0077] In the feature extraction and respiratory rate prediction stage, a three-layer convolutional neural network (CNN) was used. Through its deep learning capabilities, CNN can learn complex time-frequency features from LWT-processed data, which are key to measuring respiratory power. The CNN structure includes multiple convolutional layers and pooling layers, which can automatically detect and focus on the signal features that are most meaningful for respiratory rate prediction. In this way, not only the accuracy of the prediction is improved, but also the model's ability to generalize to data under different physiological and environmental conditions is enhanced.

[0078] This method effectively combines sensor acquisition, signal processing and machine learning prediction, which not only improves the accuracy of respiratory power measurement, but also greatly expands the application scope of monitoring technology due to its high efficiency and strong adaptability. Users can monitor their own respiratory status without the operation of professional medical personnel and effectively manage their respiratory health. In addition, through continuous monitoring, the system can help identify potential respiratory-related health problems, such as sleep apnea, early.

[0079] Please refer to the following Figure 7 , which shows a schematic diagram of an electronic device provided by some embodiments of the present application. Figure 7As shown, the electronic device 2 includes: a processor 200, a memory 201, a bus 202 and a communication interface 203, wherein the processor 200, the communication interface 203 and the memory 201 are connected via the bus 202; the memory 201 stores a computer program that can be run on the processor 200, and the processor 200 executes the breathing detection method provided in any of the aforementioned embodiments of the present application when running the computer program, wherein the method includes: acquiring original PPG data; performing signal decomposition on the original PPG data to extract frequency band data related to breathing; performing feature extraction and respiratory rate prediction on the frequency band information related to breathing to obtain a target detection result.

[0080] The memory 201 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 203 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0081] The bus 202 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 201 is used to store programs, and the processor 200 executes the programs after receiving execution instructions. The control method disclosed in any implementation of the above-mentioned embodiment of the present application may be applied to the processor 200, or implemented by the processor 200.

[0082] The processor 200 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 200. The above processor 200 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 201, and the processor 200 reads the information in the memory 201 and completes the steps of the breathing detection method in combination with its hardware.

[0083] An embodiment of the present application further provides a computer-readable storage medium corresponding to the breathing detection method provided in the aforementioned embodiment, on which a computer program is stored. When the computer program is executed by a processor, the breathing detection method provided in any of the aforementioned embodiments is executed.

[0084] In addition, examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0085] In addition, an embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the breathing detection method provided by any of the aforementioned embodiments, the method including: acquiring original PPG data; performing signal decomposition on the original PPG data to extract frequency band data related to breathing; performing feature extraction and respiratory rate prediction on the frequency band information related to breathing to obtain a target detection result.

[0086] Those skilled in the art will appreciate that the various component embodiments of the present application may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all functions of some or all components of the virtual machine creation device according to the embodiments of the present application.

[0087] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A breathing detection system, characterized in that: The system comprises: PPG sensor, used to obtain raw PPG data; A wavelet transform module, used to perform signal decomposition on the original PPG data and extract frequency band data related to breathing; The breathing rate prediction module is used to perform feature extraction and breathing rate prediction on the frequency band information related to breathing to obtain a target detection result.

2. The breathing detection system according to claim 1, characterized in that: The PPG sensor includes a light source component, a photodetector, a signal amplifier, and an analog-to-digital converter.

3. The breathing detection system according to claim 2, characterized in that: The obtaining of raw PPG data comprises: The light source assembly emits light, and the photoelectric detector receives a reflection signal of the emitted light; amplifying the reflected signal by the signal amplifier; The amplified reflected signal is converted into digital form by the analog-to-digital converter to obtain original PPG data.

4. The breathing detection system according to claim 1, characterized in that: The method of performing signal decomposition on the original PPG data and extracting frequency band data related to breathing includes: Dividing the original PPG data into even-indexed samples and odd-indexed samples; Signal decomposition is performed based on the even-numbered index samples and the odd-numbered index samples to extract frequency band data related to breathing.

5. The breathing detection system according to claim 4, characterized in that: The step of performing signal decomposition based on the even-indexed samples and the odd-indexed samples to extract frequency band data related to breathing includes: The first step is to calculate the difference between the odd-indexed sample and the even-indexed sample to obtain the detail coefficient; The second step is to calculate the average value of the odd index samples and the even index samples to obtain the approximate coefficient; In the third step, the approximation coefficients are divided into even-indexed samples and odd-indexed samples, and the first step and the second step are sequentially repeated until the frequency band data related to breathing is extracted.

6. The breathing detection system according to claim 1, characterized in that: The breathing rate prediction module includes a convolutional layer, a pooling layer and a fully connected layer.

7. The breathing detection system according to claim 6, characterized in that: The step of extracting features and predicting the breathing rate of the frequency band information related to breathing to obtain a target detection result includes: Extracting features of the frequency band information related to breathing through the convolution layer to obtain a feature map; Performing a pooling operation on the feature map through a pooling layer; The results after the pooling operation are integrated and analyzed through the fully connected layer to obtain a respiratory rate prediction value, which is used as the target detection result.

8. A breathing detection method, characterized in that: The method comprises: Get raw PPG data; Performing signal decomposition on the original PPG data to extract frequency band data related to breathing; Feature extraction and respiratory rate prediction are performed on the frequency band information related to breathing to obtain a target detection result.

9. An electronic device, comprising a memory and a processor, characterized in that: The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the breathing detection method as claimed in claim 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the breathing detection method according to claim 8 is implemented.

Images