EIT-based method and device for monitoring respiration in an anaesthesia scenario
By combining flexible curved electrode array design with multiple signal processing algorithms, the problem of signal distortion caused by contact resistance fluctuations and interference signals in respiratory monitoring of electrical impedance imaging devices under anesthesia scenarios was solved, and accurate identification of respiratory disorders was achieved.
Patent Information
- Application Number
- CN202610630953.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-09
- Publication Date
- 2026-08-04
AI Technical Summary
Existing electrical impedance imaging devices suffer from distortion of lung respiratory signals caused by contact resistance fluctuations in respiratory monitoring under anesthesia, making it impossible to accurately capture the true ventilation dynamics and respiratory disturbance characteristics of the lungs during sleep, resulting in low accuracy in respiratory disturbance identification.
The system employs a flexible curved electrode array design, a micro-bump array structure, a miniature piezoresistive pressure sensor and an impedance detection chip. It combines Kalman filtering and wavelet packet decomposition algorithms to remove heartbeat and body movement interference signals. The system uses the U-Net++ semantic segmentation algorithm to segment the lung region, extracts core respiratory feature parameters, and uses a proportional-integral-derivative PID control algorithm to compensate for contact resistance.
It improves the continuity of signal acquisition and data accuracy of respiratory monitoring under anesthesia, enhances the accuracy of capturing core respiratory features, and solves the problem of low accuracy in respiratory disorder identification.
Smart Images

Figure CN122498810A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of electrical impedance tomography (EIT) image data processing technology, and in particular to a respiratory monitoring method and device based on EIT in anesthesia scenarios. Background Technology
[0002] Currently, Electrical Impedance Tomography (EIT) is a medical imaging technique based on the difference in electrical conductivity of biological tissues. It reconstructs an image of the body's impedance distribution by applying a safe current through electrodes on the body surface and measuring the response voltage. This technique is suitable for anesthesia-related applications such as surgery. Examples include postoperative respiratory monitoring and adjusting ventilation parameters during surgery based on EIT results; it can also be used to target elderly individuals, who are at relatively high risk for anesthetic complications.
[0003] However, existing electrical impedance tomography (EIT) devices suffer from distorted respiratory signals in anesthesia due to contact resistance fluctuations. Specifically, the rigid planar electrodes (thickness > 0.5 mm) of traditional EIT do not conform well to the patient's chest wall curve during sleep under anesthesia. The electrodes are prone to displacement when the patient changes position, such as turning over or lying on their side. The contact resistance fluctuation range is > 20 Ω. This is compounded by interference signals such as body movement and heartbeat (accounting for > 15%), making it impossible to accurately capture the true ventilation dynamics and respiratory disturbance characteristics of the lungs during sleep under anesthesia. This directly affects the accuracy of respiratory disturbance identification, resulting in low accuracy of EIT monitoring for respiratory disturbance identification under anesthesia. Summary of the Invention
[0004] The purpose of this invention is to provide a respiratory monitoring method and device based on EIT in anesthesia scenarios, so as to solve the technical problem of low accuracy of EIT monitoring in identifying respiratory disorders in anesthesia scenarios.
[0005] Firstly, this application provides a respiratory monitoring method based on EIT in an anesthesia setting, the method comprising: Initial EIT data is acquired using an electrical impedance imaging (EIT) monitoring device; the EIT data includes a sequence of EIT images. Based on the initial EIT data, the peak frequency of heartbeats in the initial EIT data is extracted through frequency domain analysis, and adaptive matching is performed in the time domain using a Kalman filter algorithm to remove interference signals of heartbeats and retain the low-frequency signal of respiration in the signal spectrum corresponding to the initial EIT data, thus obtaining the first EIT data; the dynamic adjustment range of the filter cutoff frequency corresponding to the Kalman filter algorithm is a preset range; the signal is the pixel time sequence signal in the EIT image sequence; The high-frequency interference signal of body motion interference in the signal spectrum corresponding to the first EIT data is removed by using the wavelet packet decomposition algorithm to obtain the second EIT data; Based on the preset EIT lung image data, the U-Net++ semantic segmentation algorithm is used to segment the lung region in the enhanced image corresponding to the second EIT data, and output the binary mask of the lung region. Based on the binary mask of the lung region, the core respiratory feature parameters are extracted in a directional manner. Then, based on the core respiratory feature parameters, the features in the second EIT data are filtered for key features and the dimensionality of non-key features is reduced to obtain the final EIT data.
[0006] In one possible implementation, the substrate of the electrical impedance imaging (EIT) monitoring device is liquid silicone; the conductive layer of the electrical impedance imaging (EIT) monitoring device is a composite graphene conductive layer of a preset thickness; the electrode array corresponding to the electrical impedance imaging (EIT) monitoring device is set as a segmented arc structure according to preset adult thoracic anatomy data; and multiple independent electrode units in the electrical impedance imaging (EIT) monitoring device are connected in series by flexible connecting strips to adapt to changes in thoracic morphology under different body positions and ensure the adhesion between the electrodes and the skin. The electrode surface of the electrical impedance imaging (EIT) monitoring device adopts a micro-bump array structure to reduce the contact area between the electrode and the skin and reduce the pressure. The electrode edge of the electrode surface is rounded, and the back of the electrode surface is provided with an acrylic low-sensitivity pressure-sensitive adhesive and a breathable non-woven fabric fixing patch.
[0007] In one possible implementation, each electrode unit is equipped with a miniature piezoresistive pressure sensor and an impedance detection chip for real-time acquisition of contact pressure and contact resistance data. The optimal range of the contact resistance data corresponds to a preset resistance range, and the contact resistance is dynamically compensated using a proportional-integral-derivative (PID) control algorithm according to the following formula: ; in, This represents the compensation voltage at time k, used to adjust the potential of the conductive layer on the electrode surface to achieve contact resistance compensation. This represents the proportionality coefficient, used to respond to resistance deviation; This represents the integral coefficient, used to eliminate static bias; This represents the differential coefficient, used to suppress and regulate oscillations; This represents the resistance deviation at time k. , The measured contact resistance at time k is... Let be the target resistance, and represent the midpoint of the optimal interval. Indicates the sampling period.
[0008] In one possible implementation, the step of using a wavelet packet decomposition algorithm to remove high-frequency interference signals of bodily motion interference from the signal spectrum corresponding to the first EIT data to obtain the second EIT data includes: Using the db4 wavelet as the basis function and decomposing the signal of the first EIT data into 5 layers, the following decomposition formula is obtained: A small wave of buns: ;in, , indicating the number of decomposition levels. For sub-sequence numbers, This is the parent function of the db4 wavelet; express A small wave of buns; The energy of each sub-band is calculated based on the 32 wavelet packet sub-bands, and the sub-bands whose energy exceeds a preset energy threshold are identified as motion interference sub-bands; the preset energy threshold is determined based on the average energy of all sub-bands. The body motion interference sub-band is processed using a soft thresholding function to obtain the processed sub-band; wherein, the soft threshold corresponding to the soft thresholding function is . , For signal length, The standard deviation of noise; The processed subband is subjected to inverse wavelet packet transform to obtain the pure respiratory signal recovered after body motion interference removal; A respiratory signal continuity verification model is constructed based on the pure respiratory signal, and the amplitude of the filtered respiratory signal is calculated using the respiratory signal continuity verification model. When the amplitude of the filtered respiratory signal is greater than or equal to a preset amplitude threshold, it is determined to be a valid respiratory signal. When the amplitude of the filtered respiratory signal is less than the preset amplitude threshold, a secondary verification is performed in combination with body motion sensor data to obtain the second EIT data that eliminates body motion interference signals and passes the interference verification.
[0009] In one possible implementation, after removing high-frequency interference signals of bodily motion interference from the signal spectrum corresponding to the first EIT data using a wavelet packet decomposition algorithm to obtain the second EIT data, the method further includes: Using a low-noise operational amplifier, the input noise voltage is amplified, and the signal of the second EIT data is amplified to a preset voltage range value using the following formula: ;in, The signal after amplification. Indicates magnification factor; This indicates DC bias, used to keep the signal within the effective range of analog-to-digital conversion; The amplitude of the respiratory signal corresponding to the second EIT data; The amplified signal is attenuated by an electromagnetic interference (EMI) layer wrapped around the electrode leads to ensure the penetration of the shielded electric field strength; wherein, the shielding layer is a shielding layer composed of a metal braided mesh and a graphite conductive coating. Based on the amplified signal, the temporal characteristics of lung ventilation changes are enhanced using the inter-frame difference method to obtain the third EIT data after the respiratory signal is enhanced. The difference formula corresponding to the inter-frame difference method is: ; This is the result of inter-frame difference; The amplified signal of the current frame; The amplified signal of the frame following the current frame; when The system is used to determine effective ventilation changes and enhance dynamic changes in electrical impedance caused by respiratory movements, as well as weaken static interference signals, in order to improve the identification of lung ventilation characteristics after signal enhancement.
[0010] In one possible implementation, the step of segmenting the lung region in the enhanced image corresponding to the second EIT data using the U-Net++ semantic segmentation algorithm based on preset EIT lung image data and outputting a binary mask of the lung region includes: The grayscale image of the third EIT data is input into the improved U-Net++ semantic segmentation algorithm model to obtain a binary mask of the lung region; wherein, the improved U-Net++ semantic segmentation algorithm model is a segmentation network trained using the preset EIT lung image data; Based on the binary mask of the lung region, the 4x, 8x, and 16x downsampled features are fused using the following formula to obtain the multi-scale feature fusion result: in, The images are feature maps after downsampling by 4x, 8x, and 16x, respectively. The fusion coefficient; Based on the multi-scale feature fusion results, the network parameters are optimized using the Dice loss function to obtain the optimized network parameters. The segmentation loss function of the Dice loss function is expressed as follows: ;in, This is the segmentation loss function of the Dice loss function; To avoid smooth terms with a denominator of 0; This is a grayscale image of the third EIT data; This is a binary mask for the lung region; Based on the optimized network parameters, non-ventilated areas are eliminated through morphological erosion and dilation operations to remove interference areas, resulting in the final lung region segmentation result after removing interference areas. The corrosion calculation formula corresponding to the morphological corrosion and dilation calculation is as follows: , This is a binary mask for the lung region. It is a 3×3 square structural element. The result of the corrosion operation is shown; the dilation operation formula corresponding to the morphological corrosion and dilation operations is as follows: ,in, It is a 3×3 square structural element. This is the result of the erosion calculation.
[0011] In one possible implementation, the core respiratory characteristic parameters include at least one of the following: rate of change of ventilation volume, respiratory cycle stability, percentage of local ventilation defect area, amplitude of electrical impedance mutation at airway obstruction site, and alveolar expansion delay time. The step of performing key feature filtering and non-key feature dimensionality reduction on the features in the second EIT data based on the respiratory core feature parameters to obtain the final EIT data includes: Based on the aforementioned respiratory core feature parameters, the feature weights are calculated using the Relief F algorithm. Based on the feature weights, multiple key features with weights greater than the preset weights are selected, and non-key features other than the multiple key features are reduced in dimensionality by principal component analysis (PCA) to ensure that key data is retained while improving recognition efficiency, thus obtaining the final EIT data.
[0012] In one possible implementation, the electrical impedance imaging (EIT) monitoring device corresponds to an adapted sampling frequency. ,in, The highest frequency of the respiratory signal; the sampling data bit width of the electrical impedance imaging (EIT) monitoring device is 16 bits, and the data volume per frame is... Number of electrode channels, power consumption in acquisition state for: ;in, Supply voltage to lithium batteries, ; This is the average operating current. .
[0013] Secondly, this application provides a respiratory monitoring device based on EIT in an anesthesia setting, comprising: The acquisition module is used to acquire initial EIT data through an electrical impedance imaging (EIT) monitoring device; the EIT data includes an EIT image sequence; The first elimination module is used to extract the peak frequency of heartbeats from the initial EIT data through frequency domain analysis based on the initial EIT data, and then perform adaptive matching in the time domain using a Kalman filter algorithm to eliminate interference signals of heartbeats and retain the low-frequency signal of respiration in the signal spectrum corresponding to the initial EIT data, thereby obtaining the first EIT data; wherein, the dynamic adjustment range of the filter cutoff frequency corresponding to the Kalman filter algorithm is a preset range; the signal is the pixel time sequence signal in the EIT image sequence; The second elimination module is used to eliminate high-frequency interference signals of body motion interference in the signal spectrum corresponding to the first EIT data using a wavelet packet decomposition algorithm to obtain the second EIT data. The segmentation module is used to segment the lung region in the enhanced image corresponding to the second EIT data using the U-Net++ semantic segmentation algorithm based on the preset EIT lung image data, and output the binary mask of the lung region. The extraction module is used to extract respiratory core feature parameters based on the binary mask of the lung region, and to perform key feature screening and non-key feature dimensionality reduction on the features in the second EIT data according to the respiratory core feature parameters to obtain the final EIT data.
[0014] Thirdly, this application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the method described in the first aspect above.
[0015] Fourthly, this application also provides a computer-readable storage medium storing computer-executable instructions that, when invoked and executed by a processor, cause the processor to perform the method described in the first aspect above.
[0016] This application brings the following beneficial effects: This application provides a respiratory monitoring method and device based on electrical impedance tomography (EIT) in an anesthesia setting. It acquires initial EIT data using an EIT monitoring device; based on the initial EIT data, it extracts the peak frequency of cardiac pulsation from the initial EIT data through frequency domain analysis and performs adaptive matching in the time domain using a Kalman filter algorithm to remove interference signals from cardiac pulsation and retain the low-frequency respiratory signal in the signal spectrum corresponding to the initial EIT data, obtaining first EIT data; the dynamic adjustment range of the filter cutoff frequency corresponding to the Kalman filter algorithm is a preset range; a wavelet packet decomposition algorithm is used to remove high-frequency interference signals of body movement interference in the signal spectrum corresponding to the first EIT data, obtaining second EIT data; based on the preset EIT lung image data, the U-Net++ semantic segmentation algorithm is used to segment the lung region in the enhanced image corresponding to the second EIT data, outputting a binary lung region. Masking; Based on the binary mask of the lung region, the core respiratory feature parameters are extracted in a directional manner, and the features in the second EIT data are filtered for key features and the dimensionality of non-key features is reduced according to the core respiratory feature parameters to obtain the final EIT data; In this scheme, by combining adaptive Kalman filtering with wavelet packet decomposition dual-modal filtering for interference removal and deep integration of U-Net++ segmentation and multi-dimensional feature extraction algorithms, the problem of respiratory signal distortion caused by contact resistance fluctuations and interference signals in sleep state monitoring under anesthesia is solved. The amplitude of contact resistance fluctuation is controlled within a certain range, the continuity of signal acquisition is improved, and the data distortion rate is reduced. After interference removal and feature enhancement, the accuracy of capturing core respiratory features is improved, providing accurate and real EIT signal support for respiratory disorder identification, and solving the technical problem of low accuracy of EIT monitoring for respiratory disorder identification under anesthesia.
[0017] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0019] Figure 1 A flowchart illustrating the respiratory monitoring method under anesthesia based on EIT provided in this application embodiment; Figure 2Another schematic diagram of the respiratory monitoring method under anesthesia based on EIT provided in the embodiments of this application; Figure 3 A schematic diagram of a respiratory monitoring device based on EIT in an anesthesia setting provided in this application embodiment; Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation
[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0021] The terms "comprising" and "having," and any variations thereof, used in the embodiments of this application, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.
[0022] Currently, existing Electrical Impedance Tomography (EIT) devices lack specific electrode adaptation and signal anti-interference design for sleep states under anesthesia. This results in distortion of lung respiratory signals caused by contact resistance fluctuations during sleep state monitoring under anesthesia. Consequently, they cannot accurately capture the true ventilation dynamics and respiratory disturbance characteristics of the lungs during sleep states under anesthesia, directly affecting the accuracy of respiratory disturbance identification and leading to low accuracy in respiratory disturbance identification during sleep state EIT monitoring under anesthesia.
[0023] Based on this, this application provides a respiratory monitoring method and device based on EIT in anesthesia scenarios. This method can solve the technical problem of low accuracy of EIT monitoring in identifying respiratory disorders in anesthesia scenarios.
[0024] The embodiments of the present invention will be further described below with reference to the accompanying drawings.
[0025] Figure 1 This is a flowchart illustrating a respiratory monitoring method based on EIT (Extra-Intensive Therapy) in an anesthesia setting, as provided in an embodiment of this application. Figure 1 As shown, the method includes: Step S110: Acquire initial EIT data using electrical impedance imaging (EIT) monitoring equipment.
[0026] The EIT data includes EIT image sequences. In one possible implementation, the substrate of the electrical impedance tomography (EIT) monitoring device is liquid silicone; the conductive layer of the EIT monitoring device is a composite graphene conductive layer of a preset thickness; the electrode array corresponding to the EIT monitoring device is set as a segmented arc structure according to preset adult thoracic anatomy data; multiple independent electrode units in the EIT monitoring device are connected in series by flexible connecting strips to adapt to changes in thoracic morphology under different body positions and ensure the fit between the electrodes and the skin; the electrode surface of the EIT monitoring device adopts a micro-bump array structure to reduce the contact area between the electrodes and the skin and reduce the degree of pressure; the electrode edges on the electrode surface are rounded, and the back of the electrode surface is provided with an acrylic low-sensitivity pressure-sensitive adhesive and a breathable non-woven fabric fixing patch. Therefore, the effect of an adaptive flexible curved electrode array in the sleep state under anesthesia is achieved.
[0027] For the integrated electrode structure design, for example, the substrate and conductive layer: medical-grade liquid silicone (Shore hardness 30A) is used as the substrate, with a composite graphene conductive layer (thickness 50μm). The overall electrode thickness is controlled within 0.2mm, combining flexibility and conductive stability. Curved surface adaptation design: based on adult thoracic anatomy data (anteroposterior diameter 25-30cm, transverse diameter 30-35cm), the electrode array is designed as a segmented arc-shaped structure (each segment with a radius of curvature of 28-32cm), consisting of 8 independent electrode units (each unit size 15mm×20mm) connected in series by flexible connecting strips. This adapts to changes in thoracic morphology under different body positions, ensuring an electrode-skin fit of ≥95%. Skin-friendly optimization: The electrode surface adopts a micro-bump array design (bump diameter 0.5mm, height 0.1mm, spacing 0.8mm) to reduce the contact area between the electrode and the skin (contact area reduced by 30%), reducing pressure. The electrode edges are rounded with 0.5mm corners to avoid friction damage to the skin during sleep under anesthesia. The back is coated with medical low-sensitivity pressure-sensitive adhesive (acrylic material, adhesion duration ≥8 hours, peel strength ≤0.6N / cm²), and is equipped with a breathable non-woven fabric fixing patch (breathability ≥3000g / (m²·24h)) to balance fixation and comfort.
[0028] In some embodiments, each of the above-mentioned electrode units is provided with a miniature piezoresistive pressure sensor and an impedance detection chip for real-time acquisition of contact pressure and contact resistance data. The optimal range of contact resistance corresponding to the contact resistance data is a preset resistance range. Dynamic compensation of contact resistance is performed using a proportional-integral-derivative (PID) control algorithm according to the following formula: ; in, This represents the compensation voltage at time k, used to adjust the potential of the conductive layer on the electrode surface to achieve contact resistance compensation. This represents the proportionality coefficient, used to respond to resistance deviation; This represents the integral coefficient, used to eliminate static bias; This represents the differential coefficient, used to suppress and regulate oscillations; This represents the resistance deviation at time k. , The measured contact resistance at time k is... Let be the target resistance, and represent the midpoint of the optimal interval. Indicates the sampling period.
[0029] For an adaptive stabilization system for contact resistance, an exemplary feature is the built-in sensing module: each electrode unit incorporates a miniature piezoresistive pressure sensor (range 0-1 N / cm², accuracy 0.01 N / cm²) and an impedance detection chip (detection range 1-100 Ω, accuracy 0.1 Ω) to collect contact pressure and contact resistance data in real time. The dynamic adjustment mechanism sets the optimal contact resistance range to [5 Ω, 10 Ω] and uses a proportional-integral-derivative (PID) control algorithm to achieve dynamic compensation of the contact resistance. The specific formula is as follows: ;in: The compensation voltage at time k (unit: V) is used to adjust the potential of the conductive layer on the electrode surface to achieve contact resistance compensation. : Proportional coefficient, ranging from 0.8 to 1.2, used for fast response to resistance deviation; Integral coefficient, with a value of 0.05-0.1, is used to eliminate static bias; : Differential coefficient, with a value of 0.1-0.2, used to suppress regulation oscillations; The resistance deviation at time k. , The measured contact resistance at time k (unit: Ω) The target resistance is set at 7Ω, which is the midpoint of the optimal range. Sampling period, valued at 0.01s (100Hz sampling frequency). Adjust the response time to ≤50ms to ensure the contact resistance quickly stabilizes within the target range; As an optional implementation, the aforementioned electrical impedance imaging (EIT) monitoring device also includes a loosening warning mechanism: when the contact pressure is <0.1 N / cm², the device locally triggers a weak vibration alert (vibration intensity ≤0.3 g), and simultaneously sends an electrode inspection reminder to the associated terminal. The vibration triggering condition formula is as follows: ;in, The measured contact pressure at time k (unit: N / cm²) The pressure threshold, This indicates that a vibration warning has been triggered.
[0030] In one possible implementation, the electrical impedance imaging (EIT) monitoring device corresponds to an appropriate sampling frequency. ,in, This represents the highest frequency of the respiratory signal; the sampling data bit width of the electrical impedance imaging (EIT) monitoring device is 16 bits, and the data volume per frame is... Number of electrode channels, power consumption in acquisition state for: ;in, Supply voltage to lithium batteries, ; This is the average operating current. .
[0031] For supporting data processing and transmission assurance, an example is low-power data acquisition control: sampling parameter adaptation: sampling frequency. (satisfies the Nyquist sampling theorem, , (Highest frequency of respiratory signal), data width 16 bits, single frame data volume (Number of electrode channels) Power consumption during data acquisition The power consumption calculation formula is: ;in, (Lithium battery power supply voltage). (Average operating current). Cache mechanism: Employs a double-buffered design, cache capacity... Data retention time The amount of cached data meets the requirements. It supports data backtracking in abnormal scenarios.
[0032] For secure and stable data transmission from electrical impedance tomography (EIT) monitoring equipment, an exemplary transmission protocol is used: Bluetooth 5.2 BLE protocol, with data transmitted in packets, and packet length... Transmission rate Transmission distance Batch transmission cycle under non-early warning conditions Transmission delay under warning status , Indicates the number of lost data packets. This represents the total number of data packets sent. For data encryption: end-to-end encryption using the AES-256 algorithm is employed, with the encryption key... Generated through two-way authentication between the device and the terminal. The device's proprietary key. The terminal key is also secured by an authentication hash value; the transmission process uses the TLS1.3 protocol to encrypt the link, and the cloud uses the national cryptographic SM4 algorithm for backup, which complies with medical data compliance requirements.
[0033] Step S120: Based on the initial EIT data, extract the peak frequency of heartbeats in the initial EIT data through frequency domain analysis, and perform adaptive matching in the time domain using the Kalman filter algorithm to remove interference signals of heartbeats and retain the low-frequency signal of respiration in the corresponding signal spectrum of the initial EIT data to obtain the first EIT data.
[0034] The dynamic adjustment range of the filter cutoff frequency corresponding to the above Kalman filter algorithm is a preset range; the signal is the pixel time sequence signal in the EIT image sequence.
[0035] The Kalman filter algorithm uses a preset range for dynamically adjusting the filter cutoff frequency. For example, with a dynamic adjustment range of [0.3Hz, 2.5Hz], the algorithm adaptively matches the peak heartbeat frequency (1-2Hz) in the signal spectrum in real time, eliminating heartbeat interference while retaining the low-frequency components of the respiratory signal, resulting in a 40% improvement in the signal-to-noise ratio after filtering.
[0036] For precise removal of interference signals during sleep under anesthesia, an exemplary dual-modal filtering algorithm optimization is used: Low-frequency interference removal (cardiac pulsation interference): An adaptive Kalman filter algorithm is employed, with the state equation and observation equation as follows: State equation: ; Observation equation: ; Filtering update formula: Prior estimate: Prior covariance: Kalman gain: Posterior estimation: Posterior covariance: ;in: Represents the state vector. The actual respiratory signal amplitude at time k (unit: V). The rate of change of the signal (unit: V / s); Here is the state transition matrix. (Corresponding to a 15Hz sampling frequency); The noise is the process noise, which follows a mean of 0 and a variance of 0. Gaussian distribution; The signal observed at time k (including respiratory signals and cardiac pulsation interference, unit: V); The observation matrix; To observe the noise, it follows a mean of 0 and a variance of 0. The Gaussian distribution.
[0037] Step S130: Use wavelet packet decomposition algorithm to remove high-frequency interference signals of body motion interference in the signal spectrum corresponding to the first EIT data to obtain the second EIT data.
[0038] As one possible implementation, the above-mentioned wavelet packet decomposition algorithm is used to remove high-frequency interference signals of body motion interference from the signal spectrum corresponding to the first EIT data to obtain the second EIT data. Specifically, this may include the following steps: Using the db4 wavelet as the basis function and decomposing the signal of the first EIT data into 5 layers, the following decomposition formula is obtained: A small wave of buns: ;in, , indicating the number of decomposition levels. For sub-sequence numbers, It is the generating function of the db4 wavelet; express A small wave of buns; The energy of each sub-band is calculated based on 32 wavelet packet sub-bands, and sub-bands with energy exceeding a preset energy threshold are identified as motion interference sub-bands; the preset energy threshold is determined based on the average energy of all sub-bands. The motion disturbance subband is processed using a soft thresholding function to obtain the processed subband; where the soft threshold corresponding to the soft thresholding function is... , For signal length, The noise standard deviation is used; the processed subband is subjected to wavelet packet inverse transform to obtain the pure respiratory signal recovered after body motion interference removal; A respiratory signal continuity verification model is constructed based on the pure respiratory signal, and the amplitude of the filtered respiratory signal is calculated using the respiratory signal continuity verification model. When the amplitude of the filtered respiratory signal is greater than or equal to a preset amplitude threshold, it is determined to be a valid respiratory signal. When the amplitude of the filtered respiratory signal is less than the preset amplitude threshold, a second verification is performed in combination with body motion sensor data to obtain the second EIT data that eliminates body motion interference signals and passes the interference verification.
[0039] For high-frequency interference removal (body motion interference), wavelet packet decomposition algorithm is used as an example, and the specific steps are as follows: Wavelet packet decomposition: The db4 wavelet is selected as the basis function to decompose the EIT signal. Decomposed into 5 layers, resulting in The decomposition formula for wavelet packet bands is: ;in, (Number of decomposition layers) The sub-band number is (0≤m≤31). It is the mother function of the db4 wavelet; Interference subband identification: Calculate the energy of each subband Set energy threshold , The average energy of all subbands, (Empirical coefficient), energy exceeds The subband was identified as a motion interference subband; Thresholding: A soft thresholding function is used to process the interference subbands. in, Soft threshold, Given the signal length, the noise standard deviation can be 0.6745; Finally, signal reconstruction is performed: inverse wavelet packet transform is applied to the processed subbands to restore the pure respiratory signal. The motion interference rejection rate is ≥92%.
[0040] For interference verification mechanisms, for example, a respiratory signal continuity verification model is constructed, and the similarity of five adjacent frames of signals is calculated. : ;in, This represents the amplitude of the filtered respiratory signal. A respiratory signal is considered valid when the preset amplitude threshold is reached. At the same time, the signal integrity is ensured by combining the body motion sensor data for secondary verification (the body motion signal amplitude is determined to be a valid signal when it is <0.1g).
[0041] As an optional implementation, after using the wavelet packet decomposition algorithm to remove high-frequency interference signals of body motion interference from the signal spectrum corresponding to the first EIT data to obtain the second EIT data, the method may further include the following steps: Using a low-noise operational amplifier, the input noise voltage is amplified, and the signal of the second EIT data is amplified to a preset voltage range using the following formula: ;in, The signal after amplification. Indicates magnification factor; This indicates DC bias, used to keep the signal within the effective range of analog-to-digital conversion; The respiratory signal amplitude corresponding to the second EIT data; The amplified signal is attenuated by an electromagnetic interference (EMI) effect through a shielding layer wrapped around the electrode leads to ensure the penetration of the shielded electric field. The shielding layer is a metal braided mesh combined with a graphite conductive coating. Based on the amplified signal, the temporal characteristics of lung ventilation changes are enhanced using the inter-frame difference method to obtain the third EIT data after the respiratory signal is enhanced. The difference formula corresponding to the inter-frame difference method is: ; This is the result of inter-frame difference; This is the amplified signal of the current frame; This is the amplified signal of the frame following the current frame; when The system is used to determine effective ventilation changes and enhance dynamic changes in electrical impedance caused by respiratory movements, as well as weaken static interference signals, in order to improve the identification of lung ventilation characteristics after signal enhancement.
[0042] For respiratory signal enhancement processing, for example, signal amplification and noise reduction: using a low-noise operational amplifier (input noise voltage ≤1nV / √Hz), the original EIT signal (amplitude range 1-10mV) is amplified to 1-5V according to the following formula: ;in, The signal is amplified (unit: V). (Magnification) (DC bias ensures the signal remains within the effective range of analog-to-digital conversion); the electrode leads are wrapped with a shielding layer (metal braided mesh + graphite conductive coating). The intensity of the external interference electric field. The electric field strength after shielding.
[0043] For temporal feature enhancement: the inter-frame difference method is used to enhance the lung ventilation change characteristics. The difference formula is: ;when When the signal is enhanced, it is determined to be an effective ventilation change, highlighting the dynamic changes in electrical impedance caused by respiratory movements, weakening static interference signals, and improving the identification of lung ventilation characteristics by 50% after signal enhancement.
[0044] Step S140: Based on the preset EIT lung image data, use the U-Net++ semantic segmentation algorithm to segment the lung region in the enhanced image corresponding to the second EIT data, and output the binary mask of the lung region.
[0045] In one possible implementation, the above-mentioned image segmentation of the lung region in the enhanced image corresponding to the second EIT data using the U-Net++ semantic segmentation algorithm based on the preset EIT lung image data, and the output of the lung region binary mask, may specifically include the following steps: The grayscale image of the third EIT data is input into the improved U-Net++ semantic segmentation algorithm model to obtain a binary mask of the lung region; wherein, the improved U-Net++ semantic segmentation algorithm model is a segmentation network trained using the preset EIT lung image data; Based on the binary mask of the lung region, the 4x, 8x, and 16x downsampled features are fused using the following formula to obtain the multi-scale feature fusion result: in, The images are feature maps after downsampling by 4x, 8x, and 16x, respectively. The fusion coefficient; Based on the multi-scale feature fusion results, the network parameters are optimized using the Dice loss function to obtain the optimized network parameters. The segmentation loss function of the Dice loss function is expressed as follows: ;in, The segmentation loss function is the Dice loss function. To avoid smooth terms with a denominator of 0; A grayscale image of the third EIT data; A binary mask for the lung region; Based on the optimized network parameters, non-ventilated areas are eliminated through morphological erosion and dilation operations to remove interference areas, resulting in the final lung region segmentation result after removing interference areas. Among them, the corrosion calculation formulas corresponding to morphological corrosion and dilation calculations are as follows: , For the binary mask of the lung region, It is a 3×3 square structural element. This is the result of the corrosion operation; the dilation operation formula corresponding to the morphological corrosion and dilation operations is: ,in, It is a 3×3 square structural element. This is the result of the erosion calculation.
[0046] For accurate extraction and adaptation optimization of respiratory features, an exemplary model for accurate lung region segmentation is the improved U-Net++ semantic segmentation algorithm, trained on 5000 clinical EIT lung image data, with EIT grayscale images as network input. (Size 256×256 pixels), output as a binary mask of the lung region. ( Indicates the lung region. (Representing non-lung regions); Multi-scale feature fusion module: fuses 4x, 8x, and 16x downsampling features. The feature fusion formula is: ;in, The images are feature maps after downsampling by 4x, 8x, and 16x, respectively. For fusion weights (satisfying) For the segmentation loss function: the Dice loss function is used to optimize the network parameters, and the formula is: ;in, To avoid smoothing terms with a denominator of 0, the segmentation accuracy should be ≥96%.
[0047] For eliminating interference areas, for example, non-ventilated areas can be removed using morphological corrosion-dilation calculations. The corrosion calculation formula is as follows: The expansion formula is: ,in It is a 3×3 square structural element.
[0048] Step S150: Based on the binary mask of the lung region, the respiratory core feature parameters are extracted in a directional manner, and the key features in the second EIT data are screened and the non-key features are reduced in dimensionality according to the respiratory core feature parameters to obtain the final EIT data.
[0049] In one optional implementation, the core respiratory characteristic parameters include at least one of the following: rate of change of ventilation, respiratory cycle stability, percentage of area of local ventilation deficit, amplitude of electrical impedance mutation at airway obstruction site, and alveolar expansion delay time. The above-mentioned key feature selection and non-key feature dimensionality reduction of the features in the second EIT data based on respiratory core feature parameters yield the final EIT data, as shown below. Figure 2 As shown, the specific steps may include the following: Step S210: Calculate feature weights using the Relief F algorithm based on the core respiratory feature parameters; Step S220: Select multiple key features with weights greater than the preset weights based on feature weights, and perform dimensionality reduction on non-key features other than the multiple key features through principal component analysis (PCA) to ensure that recognition efficiency is improved while retaining key data, and obtain the final EIT data.
[0050] For targeted extraction of respiratory core features, an exemplary feature parameter system (including calculation formulas and accuracy requirements) includes: ventilation rate of change : Maximum voluntary ventilation during a respiratory cycle (unit: mL). Minimum ventilation rate (unit: mL). Respiratory cycle (unit: s), accuracy ±2%; respiratory cycle stability : The standard deviation is the sum of the standard deviations over 10 consecutive respiratory cycles. For the average respiratory cycle, the following is required Percentage of area with localized ventilation defects : Area of the pulmonary ventilation defect (unit: pixel). Total lung area (pixel), accuracy ±1%; amplitude of electrical impedance mutation at airway obstruction sites. : ,in The impedance value of the blocking site (unit: Ω). Mean impedance of the surrounding normal area (unit: Ω), threshold ≥5Ω; alveolar expansion delay time : ,in The time of onset of alveolar expansion (unit: s). The time to expand to the maximum volume (unit: s) is accurate to ±0.1s.
[0051] For feature selection and dimensionality reduction, for example, the Relief F algorithm is used to calculate feature weights. ,filter The eight key features are then used for dimensionality reduction through principal component analysis (PCA). These are the eigenvalues of the characteristic covariance matrix. The number of principal components after dimensionality reduction ( This ensures that core information is preserved while improving recognition efficiency.
[0052] In this embodiment, by combining adaptive Kalman filtering with wavelet packet decomposition dual-modal filtering for interference removal and deep fusion of U-Net++ segmentation and multi-dimensional feature extraction algorithms, the problem of respiratory signal distortion caused by contact resistance fluctuations and interference signals in sleep state monitoring under anesthesia is solved. The contact resistance fluctuation amplitude is controlled within ±2Ω, the continuity of signal acquisition is improved, and the data distortion rate is reduced. After interference removal and feature enhancement, the accuracy of capturing respiratory core features is improved, providing accurate and realistic EIT signal support for respiratory disorder identification, and solving the technical problem of low accuracy of EIT monitoring for respiratory disorder identification under anesthesia.
[0053] Furthermore, through the deep integration of adaptive electrode design for sleep states under anesthesia, PID-controlled contact resistance stabilization algorithm, dual-modal filtering (adaptive Kalman + wavelet packet decomposition) interference removal technology, and improved U-Net++ segmentation and multi-dimensional feature extraction algorithm, the problem of respiratory signal distortion caused by contact resistance fluctuations and interference signals in sleep state monitoring under anesthesia has been completely solved. Specifically, the contact resistance fluctuation amplitude is controlled within ±2Ω, the signal acquisition continuity reaches 99.5%, and the data distortion rate is reduced to below 2%. After interference removal and feature enhancement, the accuracy rate of capturing core respiratory features is ≥98%, providing accurate and realistic EIT signal support for respiratory disorder identification. This improves the accuracy rate of sleep state disorder identification under anesthesia to over 95%, which is 40% higher than the traditional EIT sleep state monitoring solution under anesthesia. At the same time, the patient's VAS score for sleep state comfort under anesthesia is ≥9.0 (out of 10), achieving a balance between signal accuracy and wearing comfort in sleep state monitoring under anesthesia, and significantly enhancing the clinical application value of EIT technology in the field of respiratory monitoring.
[0054] The technical solutions provided in this application can be applied to anesthesia-related application scenarios such as surgery. For example, postoperative respiratory warning and intraoperative adjustment of ventilation parameters based on EIT results are applications related to anesthesia; it can also be applied to elderly people, who are at relatively high risk of anesthesia complications.
[0055] Figure 3 A schematic diagram of a respiratory monitoring device based on EIT in an anesthesia setting is provided. Figure 3 As shown, the EIT-based respiratory monitoring device 300 for anesthesia includes: Acquisition module 301 is used to acquire initial EIT data through electrical impedance imaging (EIT) monitoring equipment; the EIT data includes EIT image sequences; The first elimination module 302 is used to extract the peak frequency of heartbeats from the initial EIT data through frequency domain analysis based on the initial EIT data, and perform adaptive matching in the time domain using a Kalman filter algorithm to eliminate interference signals of heartbeats and retain the low-frequency signal of respiration in the signal spectrum corresponding to the initial EIT data, thereby obtaining the first EIT data; wherein, the dynamic adjustment range of the filter cutoff frequency corresponding to the Kalman filter algorithm is a preset range; the signal is the pixel time sequence signal in the EIT image sequence; The second elimination module 303 is used to eliminate the high-frequency interference signal of body motion interference in the signal spectrum corresponding to the first EIT data using the wavelet packet decomposition algorithm to obtain the second EIT data. The segmentation module 304 is used to segment the lung region in the enhanced image corresponding to the second EIT data according to the preset EIT lung image data using the U-Net++ semantic segmentation algorithm, and output the binary mask of the lung region. Extraction module 305 is used to extract respiratory core feature parameters based on the binary mask of the lung region, and to perform key feature screening and non-key feature dimensionality reduction on the features in the second EIT data according to the respiratory core feature parameters to obtain the final EIT data.
[0056] The respiratory monitoring device based on EIT in the anesthesia scenario provided in this application embodiment has the same technical features as the respiratory monitoring method based on EIT in the anesthesia scenario provided in the above embodiment, so it can also solve the same technical problems and achieve the same technical effects.
[0057] An electronic device provided in this application embodiment, such as Figure 4 As shown, the electronic device 400 includes a processor 402 and a memory 401. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the steps of the method provided in the above embodiments.
[0058] See Figure 4 The electronic device also includes a bus 403 and a communication interface 404. The processor 402, the communication interface 404 and the memory 401 are connected via the bus 403. The processor 402 is used to execute executable modules, such as computer programs, stored in the memory 401.
[0059] The memory 401 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 404 (which can be wired or wireless), such as the Internet, wide area network, local area network, or metropolitan area network.
[0060] Bus 403 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.
[0061] The memory 401 is used to store programs. After receiving an execution instruction, the processor 402 executes the program. The method executed by the apparatus defined by the process disclosed in any of the preceding embodiments of this application can be applied to the processor 402 or implemented by the processor 402.
[0062] Processor 402 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 402 or by instructions in software form. The processor 402 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 gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory 401, and processor 402 reads the information from memory 401 and, in conjunction with its hardware, completes the steps of the above method.
[0063] Corresponding to the above-described EIT-based respiratory monitoring method under anesthesia, this application embodiment also provides a computer-readable storage medium storing computer-executable instructions. When the computer-executable instructions are called and run by a processor, the computer-executable instructions cause the processor to perform the steps of the above-described EIT-based respiratory monitoring method under anesthesia.
[0064] The respiratory monitoring device based on EIT in anesthesia scenarios provided in this application embodiment can be specific hardware on the device or software or firmware installed on the device. The implementation principle and technical effects of the device provided in this application embodiment are the same as those in the foregoing method embodiments. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the foregoing method embodiments. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can all be referred to the corresponding processes in the above method embodiments, and will not be repeated here.
[0065] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.
[0066] For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.
[0067] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0068] In addition, the functional units in the embodiments provided in this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0069] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the EIT-based respiratory monitoring method under anesthesia described in the various embodiments of this application. The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.
[0070] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.
[0071] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The protection scope of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application. All should be covered within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.
Claims
1. A respiratory monitoring method based on EIT in an anesthesia setting, characterized in that, The method includes: Initial EIT data is acquired using an electrical impedance imaging (EIT) monitoring device; the EIT data includes a sequence of EIT images. Based on the initial EIT data, the peak frequency of heartbeats in the initial EIT data is extracted through frequency domain analysis, and adaptive matching is performed in the time domain using a Kalman filter algorithm to remove interference signals of heartbeats and retain the low-frequency signal of respiration in the signal spectrum corresponding to the initial EIT data, thus obtaining the first EIT data; the dynamic adjustment range of the filter cutoff frequency corresponding to the Kalman filter algorithm is a preset range; the signal is the pixel time sequence signal in the EIT image sequence; The high-frequency interference signal of body motion interference in the signal spectrum corresponding to the first EIT data is removed by using the wavelet packet decomposition algorithm to obtain the second EIT data; Based on the preset EIT lung image data, the U-Net++ semantic segmentation algorithm is used to segment the lung region in the enhanced image corresponding to the second EIT data, and output the binary mask of the lung region. Based on the binary mask of the lung region, the core respiratory feature parameters are extracted in a directional manner. Then, based on the core respiratory feature parameters, the features in the second EIT data are filtered for key features and the dimensionality of non-key features is reduced to obtain the final EIT data.
2. The method according to claim 1, characterized in that, The substrate of the electrical impedance imaging (EIT) monitoring device is liquid silicone; the conductive layer of the electrical impedance imaging (EIT) monitoring device is a composite graphene conductive layer of a preset thickness; the electrode array corresponding to the electrical impedance imaging (EIT) monitoring device is set as a segmented arc structure according to preset adult thoracic anatomy data; multiple independent electrode units in the electrical impedance imaging (EIT) monitoring device are connected in series by flexible connecting strips to adapt to changes in thoracic morphology under different body positions and ensure the adhesion between the electrodes and the skin. The electrode surface of the electrical impedance imaging (EIT) monitoring device adopts a micro-bump array structure to reduce the contact area between the electrode and the skin and reduce the pressure. The electrode edge of the electrode surface is rounded, and the back of the electrode surface is provided with an acrylic low-sensitivity pressure-sensitive adhesive and a breathable non-woven fabric fixing patch.
3. The method according to claim 2, characterized in that, Each electrode unit is equipped with a miniature piezoresistive pressure sensor and an impedance detection chip for real-time acquisition of contact pressure and contact resistance data. The optimal range of the contact resistance data corresponds to a preset resistance range. The contact resistance is dynamically compensated using a proportional-integral-derivative (PID) control algorithm according to the following formula: ; in, This represents the compensation voltage at time k, used to adjust the potential of the conductive layer on the electrode surface to achieve contact resistance compensation. This represents the proportionality coefficient, used to respond to resistance deviation; This represents the integral coefficient, used to eliminate static bias; This represents the differential coefficient, used to suppress and regulate oscillations; This represents the resistance deviation at time k. , The measured contact resistance at time k is... Let be the target resistance, and represent the median of the optimal interval. Indicates the sampling period.
4. The method according to claim 1, characterized in that, The process involves using wavelet packet decomposition to remove high-frequency interference signals from the signal spectrum corresponding to the first EIT data, resulting in the second EIT data, including: Using the db4 wavelet as the basis function and decomposing the signal of the first EIT data into 5 layers, the following decomposition formula is obtained: A small wave of buns: ;in, , indicating the number of decomposition levels. For sub-sequence numbers, This is the parent function of the db4 wavelet; express A small wave of buns; The energy of each sub-band is calculated based on the 32 wavelet packet sub-bands, and the sub-bands whose energy exceeds a preset energy threshold are identified as motion interference sub-bands; the preset energy threshold is determined based on the average energy of all sub-bands. The motion interference sub-band is processed using a soft thresholding function to obtain the processed sub-band; wherein the soft threshold corresponding to the soft thresholding function is . , For signal length, The standard deviation of noise; The processed subband is subjected to wavelet packet inverse transform to obtain the pure respiratory signal recovered after body motion interference removal; A respiratory signal continuity verification model is constructed based on the pure respiratory signal, and the amplitude of the filtered respiratory signal is calculated using the respiratory signal continuity verification model. When the amplitude of the filtered respiratory signal is greater than or equal to a preset amplitude threshold, it is determined to be a valid respiratory signal. When the amplitude of the filtered respiratory signal is less than the preset amplitude threshold, a secondary verification is performed in combination with body motion sensor data to obtain the second EIT data that eliminates body motion interference signals and passes the interference verification.
5. The method according to claim 4, characterized in that, After removing high-frequency interference signals of body motion interference from the signal spectrum corresponding to the first EIT data using the wavelet packet decomposition algorithm to obtain the second EIT data, the method further includes: Using a low-noise operational amplifier, the input noise voltage is amplified, and the signal of the second EIT data is amplified to a preset voltage range value using the following formula: ;in, The signal after amplification. Indicates magnification factor; This indicates DC bias, used to keep the signal within the effective range of analog-to-digital conversion; The amplitude of the respiratory signal corresponding to the second EIT data; The amplified signal is attenuated by an electromagnetic interference (EMI) layer wrapped around the electrode leads to ensure the penetration of the shielded electric field strength; wherein, the shielding layer is a shielding layer composed of a metal braided mesh and a graphite conductive coating. Based on the amplified signal, the temporal characteristics of lung ventilation changes are enhanced using the inter-frame difference method to obtain the third EIT data after the respiratory signal is enhanced. The difference formula corresponding to the inter-frame difference method is: ; This is the result of inter-frame difference; The amplified signal of the current frame; The amplified signal of the frame following the current frame; when The system is used to determine effective ventilation changes and enhance dynamic changes in electrical impedance caused by respiratory movements, as well as weaken static interference signals, in order to improve the identification of lung ventilation characteristics after signal enhancement.
6. The method according to claim 5, characterized in that, The step of segmenting the lung region in the enhanced image corresponding to the second EIT data using the U-Net++ semantic segmentation algorithm based on the preset EIT lung image data and outputting a binary mask of the lung region includes: The grayscale image of the third EIT data is input into the improved U-Net++ semantic segmentation algorithm model to obtain a binary mask of the lung region; wherein, the improved U-Net++ semantic segmentation algorithm model is a segmentation network trained using the preset EIT lung image data; Based on the binary mask of the lung region, the 4x, 8x, and 16x downsampled features are fused using the following formula to obtain the multi-scale feature fusion result: in, The images are feature maps after downsampling by 4x, 8x, and 16x, respectively. The fusion coefficient; Based on the multi-scale feature fusion results, the network parameters are optimized using the Dice loss function to obtain the optimized network parameters. The segmentation loss function of the Dice loss function is expressed as follows: ;in, This is the segmentation loss function of the Dice loss function; To avoid smooth terms with a denominator of 0; This is a grayscale image of the third EIT data; This is a binary mask for the lung region; Based on the optimized network parameters, non-ventilated areas are eliminated through morphological erosion and dilation operations to remove interference areas, resulting in the final lung region segmentation result after removing interference areas. The corrosion calculation formula corresponding to the morphological corrosion and dilation calculation is as follows: , This is a binary mask for the lung region. It is a 3×3 square structural element. The result of the corrosion operation is shown; the dilation operation formula corresponding to the morphological corrosion and dilation operations is as follows: ,in, It is a 3×3 square structural element. This is the result of the erosion calculation.
7. The method according to claim 1, characterized in that, The core respiratory characteristic parameters include at least one of the following: ventilation rate of change, respiratory cycle stability, area of local ventilation defect, electrical impedance mutation amplitude of airway obstruction site, and alveolar expansion delay time. The step of performing key feature filtering and non-key feature dimensionality reduction on the features in the second EIT data based on the respiratory core feature parameters to obtain the final EIT data includes: Based on the aforementioned respiratory core feature parameters, the feature weights are calculated using the Relief F algorithm. Based on the feature weights, multiple key features with weights greater than the preset weights are selected, and non-key features other than the multiple key features are reduced in dimensionality by principal component analysis (PCA) to ensure that key data is retained while improving recognition efficiency, thus obtaining the final EIT data.
8. The method according to claim 1, characterized in that, The electrical impedance tomography (EIT) monitoring device is adapted to the following sampling frequency. ,in, The highest frequency of the respiratory signal; the sampling data bit width of the electrical impedance imaging (EIT) monitoring device is 16 bits, and the data volume per frame is... Number of electrode channels, power consumption in acquisition state for: ;in, Supply voltage to lithium batteries, ; This is the average operating current. .
9. A respiratory monitoring device based on EIT in an anesthesia setting, characterized in that, include: The acquisition module is used to acquire initial EIT data through the electrical impedance imaging EIT monitoring device; The EIT data includes EIT image sequences; The first elimination module is used to extract the peak frequency of heartbeats from the initial EIT data through frequency domain analysis based on the initial EIT data, and then perform adaptive matching in the time domain using a Kalman filter algorithm to eliminate interference signals of heartbeats and retain the low-frequency signal of respiration in the signal spectrum corresponding to the initial EIT data, thereby obtaining the first EIT data; wherein, the dynamic adjustment range of the filter cutoff frequency corresponding to the Kalman filter algorithm is a preset range; the signal is the pixel time sequence signal in the EIT image sequence; The second elimination module is used to eliminate high-frequency interference signals of body motion interference in the signal spectrum corresponding to the first EIT data using a wavelet packet decomposition algorithm to obtain the second EIT data. The segmentation module is used to segment the lung region in the enhanced image corresponding to the second EIT data using the U-Net++ semantic segmentation algorithm based on the preset EIT lung image data, and output the binary mask of the lung region. The extraction module is used to extract respiratory core feature parameters based on the binary mask of the lung region, and to perform key feature screening and non-key feature dimensionality reduction on the features in the second EIT data according to the respiratory core feature parameters to obtain the final EIT data.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when invoked and executed by a processor, cause the processor to perform the method according to any one of claims 1 to 8.
