Ventilation data processing method, electronic equipment and anesthesia machine

Abstract

The invention discloses a ventilation data processing method, electronic equipment and an anaesthesia machine, and relates to the technical field of computers. The method comprises the following steps: generating a to-be-identified waveform without amplitude information according to original ventilation data of a current acquisition period; calculating the similarity degree of the waveform shape between the waveform to be identified and each target standard waveform, wherein each target standard waveform is a waveform which is generated according to ventilation parameter standard values of different disease features and does not contain amplitude information; and determining whether the to-be-identified waveform is abnormal or not according to the similarity degree. According to the method, the problems that tiny pathological changes cannot be actively perceived and a user needs to continuously and frequently monitor pathological characteristic waveforms in related technologies can be solved, and abnormal waveform data can be accurately and efficiently recognized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computers, and in particular to a ventilation data processing method, an electronic device and an anesthesia machine. BACKGROUND

[0002] The anesthesia machine assists the patient to complete exhalation by outputting anesthetic gas (such as sevoflurane, desflurane, etc.) and oxygen or air according to preset tidal volume, respiratory rate and other parameters through a breathing machine. In order to ensure that gas exchange proceeds normally, the anesthesia machine monitors ventilation parameter data and passively prompts an alarm when a single parameter exceeds the corresponding threshold. However, this ventilation parameter data monitoring method cannot identify abnormal conditions before the threshold, cannot actively perceive small pathological changes, and the user needs to continuously and frequently monitor the pathological characteristic waveform.

[0003] In view of this, accurately and efficiently identifying abnormal waveform data is a technical problem that needs to be solved by those skilled in the art.

[0004] It should be noted that the information disclosed in the above background section is only used to strengthen the understanding of the background of the present application, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art.

[0005] SUMMARY

[0006] The present application provides a ventilation data processing method, an electronic device and an anesthesia machine, which accurately and efficiently identify abnormal waveform data, can identify waveform morphological change trends with pathological significance, do not need artificial continuous and frequent monitoring of pathological characteristic waveforms, and improve user experience.

[0007] To solve the above technical problems, the present application provides the following technical solutions:

[0008] The present application provides a ventilation data processing method, an electronic device and an anesthesia machine, which accurately and efficiently identify abnormal waveform data, can identify waveform morphological change trends with pathological significance, do not need artificial continuous and frequent monitoring of pathological characteristic waveforms, and improve user experience.

[0009] According to the original ventilation data of the current collection period, a to-be-identified waveform without amplitude information is generated; the similarity of the waveform shape between the to-be-identified waveform and each target standard waveform is calculated, and each target standard waveform is a waveform without amplitude information generated according to the ventilation parameter standard value of different disease characteristics; and whether the to-be-identified waveform is abnormal is determined according to the similarity.

[0010] For example, according to the original ventilation data of the current collection period, a to-be-identified waveform without amplitude information is generated, including:

[0011] From the original ventilation data of the current acquisition cycle, the maximum ventilation parameter maximum value and the minimum ventilation parameter minimum value with the maximum value in the amplitude interval range are selected; the amplitude range information is determined according to the ventilation parameter maximum value and the ventilation parameter minimum value; for each data point of the original ventilation data, the difference value information between the current data point and the bottom of the amplitude interval range is determined according to the value of the current data point in the amplitude interval range and the ventilation parameter minimum value, and the new value of the current data point in the amplitude interval range is determined according to the difference value information and the amplitude range information; the new value of each data point of the original ventilation data is generated to generate the to-be-identified waveform without amplitude information.

[0012] Exemplarily, before the to-be-identified waveform without amplitude information is generated, the method further comprises:

[0013] The total number of waveform points of each target standard waveform is obtained; if the total number of data points of the original ventilation data is less than the total number of waveform points, the original ventilation data is subjected to interpolation processing so that the total number of data points of the original ventilation data is the same as the total number of waveform points of the target standard waveform; if the total number of data points of the original ventilation data is greater than the total number of waveform points, the original ventilation data is subjected to filtering processing, and the original ventilation data subjected to filtering processing is resampled according to the total number of waveform points.

[0014] Exemplarily, the determining whether the to-be-identified waveform is abnormal according to the similarity degree comprises:

[0015] If at least one similarity degree quantization value is greater than a first threshold value, the to-be-identified waveform is abnormal; if at least one target similarity degree quantization value is greater than or equal to a second threshold value and less than or equal to the first threshold value, log information is generated according to the to-be-identified waveform, the target similarity degree quantization value and the target standard waveform corresponding to the target similarity degree quantization value; if there is no similarity degree quantization value greater than the first threshold value, the to-be-identified waveform is normal.

[0016] Exemplarily, before the to-be-identified waveform without amplitude information is generated according to the original ventilation data of the current acquisition cycle, the method further comprises:

[0017] When a detection interval setting instruction is received, the number of exhalation cycles is extracted from the detection interval setting instruction; if the current breathing cycle statistical number reaches the number of exhalation cycles, the acquisition operation of the original ventilation data is triggered.

[0018] Exemplarily, before the similarity degree of the waveform shape between the to-be-identified waveform and each target standard waveform is calculated, the method further comprises:

[0019] Upon receiving a standard waveform construction instruction, the system extracts the disease name and corresponding ventilation parameter standard values ​​based on the disease characteristics from the instruction; performs amplitude reduction processing on the data of each standard point of the ventilation parameter standard value within the amplitude range; generates a standard waveform based on each standard point after amplitude reduction processing, using the ventilation parameter standard value as the vertical axis and the sampling frequency as the horizontal axis, and sets at least a disease name label for the standard waveform; and stores the standard waveform in the waveform template library.

[0020] For example, the step of performing amplitude reduction processing on the data of each standard point of the ventilation parameter standard value within the amplitude range includes:

[0021] If the numerical difference between adjacent standard points of the ventilation parameter standard value is greater than a preset discrete threshold, the ventilation parameter standard value is interpolated; if the ventilation parameter standard value includes a target standard point whose numerical difference with all standard points in the preset neighborhood is greater than the preset discrete threshold, the target standard point is deleted from the ventilation parameter standard value; from the current ventilation parameter standard values, the largest standard value and the smallest standard value within the amplitude range are selected; based on the largest and smallest standard values, standard amplitude range information is determined; for each standard point, based on the current standard point's value within the amplitude range and the smallest standard value, the standard deviation information between the current standard point and the bottom of the amplitude range is determined, and based on the standard deviation information and the standard amplitude range information, a new standard value for the current standard point within the amplitude range is determined, so as to generate a standard waveform without amplitude information based on the new standard values ​​of each standard point.

[0022] For example, calculating the similarity of the waveform shape between the waveform to be identified and each target standard waveform includes:

[0023] A local cost matrix is ​​constructed based on the squared differences between each data point of the waveform to be identified and the standard points of the current target standard waveform. The matrix dimension of the local cost matrix is ​​determined based on the total number of data points of the waveform to be identified and the number of standard points of the current target standard waveform. A cumulative cost matrix is ​​constructed by taking the sum of the smallest element among the surrounding elements of each element of the local cost matrix as a new matrix element. Based on the cumulative cost matrix, the optimal alignment value between the waveform to be identified and the current target standard waveform is determined. The sum of the squared differences of each optimal alignment value is calculated, and the initial similarity value between the waveform to be identified and the current target standard waveform is determined based on the ratio of the sum to the total logarithm of the optimal alignment values. The quantification value of the similarity between the waveform to be identified and the current target standard waveform is determined based on the difference between the initial similarity value and the preset benchmark value.

[0024] This application also provides an electronic device, including a memory and a processor, wherein the processor is used to implement the steps of any of the ventilation data processing methods described above when executing a computer program stored in the memory.

[0025] Finally, this application also provides an anesthesia machine, including the aforementioned electronic device, which is connected to an anesthesia machine sensor, an alarm, and a human-machine interface device; the human-machine interface device includes at least a waveform parameter input page; the processor of the electronic device communicates with the anesthesia machine sensor through a target hardware interface, and the alarm triggers an audible and visual alarm upon receiving an alarm signal from the processor.

[0026] The advantage of the technical solution provided in this application lies in its ability to perform multi-parameter and holistic waveform feature analysis on the currently collected ventilation data and ventilation parameters with different disease characteristics. This analysis can identify waveform morphology change trends with pathological significance and detect dangerous situations where the values ​​are normal but the waveform morphology is abnormal before the parameters exceed the limits. This greatly reduces missed diagnoses caused by improper threshold settings or special disease conditions. Compared with early warning based on a single static threshold, this approach achieves a shift from "passive alarm" to "active early warning." By comparing the waveform shape with different disease characteristics, the pathological feature information reflected in the current ventilation data can be accurately understood, and abnormal wavelengths can be accurately identified. This eliminates the need for continuous and frequent manual monitoring of pathological characteristic waveforms, thus improving the user experience.

[0027] Furthermore, this application also provides corresponding electronic devices and anesthesia machines for implementing the ventilation data processing method, further making the method more practical. The electronic devices and anesthesia machines have corresponding advantages.

[0028] The technical features mentioned above, those to be mentioned below, and those shown individually in the accompanying drawings can be arbitrarily combined, as long as the combined technical features are not contradictory. All feasible combinations of features are the technical content explicitly described in this application. Any one of the multiple sub-features contained in the same statement can be applied independently, without necessarily being applied together with other sub-features.

[0029] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this application. Attached Figure Description

[0030] To more clearly illustrate the technical solutions of this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 A flowchart illustrating a ventilation data processing method provided in this application;

[0032] Figure 2 A schematic diagram illustrating the acquisition of the original ventilation data provided in this application;

[0033] Figure 3 A schematic diagram of the standard waveform of airway pressure provided in this application;

[0034] Figure 4 A schematic diagram illustrating the determination of the optimal alignment value provided in this application;

[0035] Figure 5 A structural framework diagram of an exemplary embodiment of the ventilation data processing device provided in this application;

[0036] Figure 6 A structural diagram of an exemplary embodiment of the electronic device provided in this application;

[0037] Figure 7 This is a structural diagram of an exemplary embodiment of the anesthesia machine provided in this application. Detailed Implementation

[0038] To enable those skilled in the art to better understand the technical solutions of this application, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. The terms "first," "second," "third," "fourth," etc., used in the specification and the aforementioned drawings are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. The term "exemplary" means "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as superior to or better than other embodiments.

[0039] With the continuous development of anesthesia depth monitoring technology and ventilator technology, mainstream clinical ventilation parameter alarm systems generally adopt a fixed threshold triggering mechanism, that is, based on airway pressure, end-expiratory time, etc. Individual ventilation parameters such as airway concentration and expiratory flow rate are preset with fixed normal range limits. When a parameter exceeds or falls below this limit, the system triggers an alarm. For example, when airway pressure exceeds a preset safe upper limit, or when end-expiratory pressure exceeds a certain limit... When the concentration exceeds the normal clinical reference range, the alarm system will be activated to provide a notification.

[0040] However, alarms based solely on the numerical threshold of a single ventilation parameter, without analyzing the overall shape and trend of the waveform, make it difficult to identify subtle pathological changes where "parameter values ​​are within the normal range but the waveform morphology is abnormal." For many progressive diseases (such as bronchospasm, pulmonary embolism, and diaphragmatic elevation during laparoscopic surgery), the waveform "shape" may have already undergone characteristic changes before reaching the threshold, such as the "shark fin" expiratory waveform of bronchospasm, making it very easy to miss. Furthermore, ventilation waveforms are updated in real-time at high frequencies (e.g., continuously updated according to the sampling frequency of the anesthesia machine).

[0041] Therefore, this application performs multi-parameter and holistic waveform feature analysis on currently collected ventilation data and ventilation parameters of different disease characteristics. This enables the identification of pathologically significant waveform morphology change trends, and detects dangerous situations where values ​​are normal but waveform morphology is abnormal even before parameters exceed limits. It can identify pathologically significant waveform morphology change trends without requiring continuous and frequent manual monitoring of pathological waveforms, thus improving the user experience. After introducing the technical solution of this application, the various non-limiting embodiments of this application are described in detail below with reference to the accompanying drawings and specific implementation methods.

[0042] Please see first. Figure 1 , Figure 1 This is a flowchart illustrating a ventilation data processing method provided in this embodiment. This embodiment may include the following:

[0043] S101: Generate a waveform to be identified without amplitude information based on the raw ventilation data of the current acquisition cycle.

[0044] The current data acquisition cycle is the preset duration of a single ventilation cycle on the anesthesia machine, such as one respiratory cycle (approximately 3 seconds, matching the patient's normal respiratory rate of 20 breaths / minute). The raw ventilation data consists of airway pressure and end-expiratory temperature readings collected in real-time by the anesthesia machine's sensors. Ventilation-related parameters such as concentration and expiratory flow rate, with raw units, such as airway pressure. The expiratory flow rate is measured in L / min. For example, the raw ventilation data consists of airway pressure data collected by the anesthesia machine's sensors over one respiratory cycle at a sampling frequency of 50Hz (one data point every 0.02 seconds), totaling 180 data points, with a data range of 8-35. (Includes complete waveform data of inspiratory pressure rise, plateau phase, and expiratory pressure fall). These 180 raw airway pressure data points constitute the raw ventilation data. Examples of some key data points are shown below. Figure 2 As shown, the waveform to be identified without amplitude information refers to the original ventilation data, such as original airway pressure data, after amplitude reduction processing to eliminate 8-35. The amplitude difference is removed, and only the waveform shape features are retained. The generated waveform has data point values ​​mapped to the [0, 1] interval. This embodiment eliminates the interference of the original data amplitude difference through amplitude removal processing, accurately identifies abnormal waveforms that are highly similar to the standard waveform, and the similarity quantization value can reach at least 0.93. The alarm response delay is <100ms (meeting the needs of clinical real-time monitoring).

[0045] S102: Calculate the similarity of the waveform shape between the waveform to be identified and each target standard waveform. Each target standard waveform is a waveform without amplitude information generated based on the standard values ​​of ventilation parameters according to different disease characteristics.

[0046] The waveform to be identified is the waveform generated after amplitude reduction and preprocessing of the original ventilation data, retaining only the waveform shape characteristics and not amplitude information. The target standard waveform is, for example, a waveform generated by clinical experts based on standard values ​​of ventilation parameters for different disease characteristics, stored in a waveform template library after amplitude reduction processing, and each waveform can be labeled with a corresponding disease name. The similarity of waveform shapes refers to the degree of similarity between two waveforms determined by any similarity algorithm, and can be represented by a quantitative value (i.e., a specific value within the range of 0-1). The closer the value is to 1, the more similar the waveform to be identified is to the target standard waveform. It can also be qualitatively represented, such as very similar, not very similar, and not similar. For example, the shape characteristics of the standard waveform for ARDS are: slow rise in inspiratory pressure, persistently high pressure during the plateau phase, and gradual decrease in expiratory pressure; the shape characteristics of the standard waveform for bronchospasm are: a sudden rise in inspiratory pressure and a "sawtooth" fluctuation in expiratory pressure.

[0047] S103: Determine whether the waveform to be identified is abnormal based on the degree of similarity.

[0048] A judgment threshold can be preset. Based on clinical needs, the first threshold can be set to 0.9 (high similarity threshold, judged as abnormal) and the second threshold to 0.8 (medium similarity threshold, recorded in the log). If the similarity quantification value (0.93) between the waveform to be identified and the standard waveform is greater than the first threshold (0.9), the condition of "at least one similarity quantification value is greater than the first threshold" is met. Finally, if the waveform to be identified is abnormal, the anesthesia machine's electronic equipment immediately sends a signal to the alarm, triggering an audible and visual alarm, and the message "Suspected ARDS, it is recommended to check the patient's lung compliance" is highlighted on the human-machine interface. By judging the similarity of waveform shape, early pathological signals of "airway pressure values ​​not exceeding the upper limit but waveform morphology abnormal" can be captured (such as waveform changes caused by early ARDS lung compliance decline), effectively reducing the risk of missed diagnosis and buying time for clinical intervention.

[0049] In the technical solution provided in this application embodiment, the currently collected ventilation data and ventilation parameters of different disease characteristics are subjected to multi-parameter and holistic waveform feature analysis. This can identify the waveform morphology change trend with pathological significance, and detect dangerous situations where the values ​​are normal but the waveform morphology is abnormal before the parameters exceed the standard. This greatly reduces the missed diagnosis caused by improper threshold settings or special conditions. Compared with early warning based on a single static threshold, it realizes the transformation from "passive alarm" to "active early warning". By comparing the waveform shape with different disease characteristics, the pathological feature information reflected by the current ventilation data can be accurately understood, and abnormal wavelengths can be accurately identified. There is no need for continuous and frequent manual monitoring of pathological characteristic waveforms, which improves the user experience.

[0050] Considering that amplitude differences in the raw ventilation data can interfere with waveform shape comparison, making it impossible to accurately determine whether the waveform is abnormal due to the disease, in the above embodiments, this embodiment also provides an exemplary method for generating the waveform to be identified, which may include the following:

[0051] From the raw ventilation data of the current acquisition cycle, select the maximum value and minimum value of the ventilation parameter that are within the amplitude range. Based on the maximum and minimum values ​​of the ventilation parameters, determine the amplitude range information. For each data point in the raw ventilation data, determine the difference between the current data point and the bottom of the amplitude range based on the current data point's value within the amplitude range and the minimum value of the ventilation parameter. Based on the difference information and the amplitude range information, determine the new value of the current data point within the amplitude range. Based on the new values ​​of each data point in the raw ventilation data, generate a waveform to be identified that does not contain amplitude information.

[0052] The amplitude range refers to the range within which the values ​​in the raw ventilation data or standard values ​​of ventilation parameters fall, used to determine the maximum, minimum, and amplitude range of the data. Taking a Cartesian coordinate system as an example, each data point of the waveform corresponds to a value on the xy-axis. The amplitude range can be the y-axis, and the maximum and minimum values ​​of the ventilation parameters can be the maximum and minimum values ​​of the entire waveform on the y-axis. The amplitude range information can be the difference between the maximum and minimum values ​​of the ventilation parameters. For example, if the anesthesia machine sensor collects raw airway pressure ventilation data within the current acquisition cycle, its value range might be 5-30. That is, the amplitude range is [5, 30]. From this raw ventilation data, the ventilation parameter with the largest value, 30, was selected. And the minimum value of the ventilation parameter with the smallest value is 5. The amplitude range information is calculated based on the maximum and minimum values, i.e., 30-5=25. For a specific data point in the raw ventilation data, the value is 15. Calculate the difference between this data point and the bottom of the amplitude range, i.e., 15-5=10. Divide the difference information by the amplitude range information to obtain the new value of the data point, i.e., 10 / 25 = 0.4. Following the above steps, calculate the new values ​​of all data points in the original ventilation data, and generate the waveform to be identified without amplitude information based on these new values.

[0053] For example, the raw ventilation data consists of 180 raw airway pressure data points. From these 180 raw data points, the maximum value of the ventilation parameter within the amplitude range [8, 35] is selected (35... ) and minimum value (8) ); Calculate the amplitude range information: 35-8=27 For each original data point, calculate the new value using the formula: New Value = (Current Data Point Value - Minimum Value) / Amplitude Range, with a sampling time of 0.50 seconds and 25... For example, the new value = (25-8) / 27≈0.63; based on the new values ​​of 180 data points (all in the range of [0,1]), generate a waveform to be identified that retains only shape features and does not contain amplitude information (the horizontal axis is the sampling time and the vertical axis is the new value).

[0054] As can be seen from the above, this embodiment eliminates the influence of amplitude differences on waveform shape judgment by performing amplitude removal processing on the original ventilation data, so that the subsequent similarity calculation is based solely on waveform shape, thereby improving the accuracy of abnormal waveform identification.

[0055] Considering that the total number of data points between the original ventilation data and the target standard waveform is inconsistent, making it impossible to accurately calculate the similarity between the two, based on the above embodiments, the present invention also provides the following data processing implementation method, which may include the following:

[0056] Obtain the total number of waveform points for each target standard waveform; if the total number of data points in the original ventilation data is less than the total number of waveform points, then interpolate the original ventilation data to make the total number of data points in the original ventilation data the same as the total number of waveform points in the target standard waveform; if the total number of data points in the original ventilation data is greater than the total number of waveform points, then filter the original ventilation data and resample the filtered original ventilation data according to the total number of waveform points.

[0057] Interpolation refers to a process that smooths the waveform and makes the data more complete when the total number of data points in the original ventilation data or standard values ​​of ventilation parameters is insufficient or when there are points with excessive dispersion. Filtering refers to a process used to remove noise interference from the original ventilation data or to filter data when the total number of data points is too large.

[0058] Obtain the total number of waveform points for each target standard waveform. Assume a target standard waveform has 200 waveform points. If the total number of data points in the raw ventilation data of the current acquisition period is 150, which is less than the target standard waveform's total of 200, then perform linear interpolation on the raw ventilation data to add 50 data points, bringing the total number of data points in the processed raw ventilation data to 200. If the total number of data points in the raw ventilation data of the current acquisition period is 250, which is greater than the target standard waveform's total of 200, then first perform Butterworth low-pass filtering on the raw ventilation data to remove noise interference. Then, resample the filtered raw ventilation data according to the total number of 200 waveform points to obtain ventilation data with 200 data points.

[0059] As can be seen from the above, this embodiment uses interpolation or filtering resampling to ensure that the total number of data points of the original ventilation data is consistent with the total number of data points of the target standard waveform, thus providing a data basis for the accurate calculation of the similarity in the future.

[0060] The above embodiments do not limit the similarity judgment criteria, and cannot reasonably process waveforms in different situations according to the similarity, resulting in inaccurate alarms or omission of abnormal information. The present invention also provides an exemplary implementation method, which may include the following:

[0061] If at least one similarity quantization value is greater than the first threshold, the waveform to be identified is abnormal; if at least one target similarity quantization value is greater than or equal to the second threshold and less than or equal to the first threshold, log information is generated based on the waveform to be identified, the target similarity quantization value, and the target standard waveform corresponding to the target similarity quantization value; if there is no similarity quantization value greater than the first threshold, the waveform to be identified is normal.

[0062] The similarity quantification value can be calculated using any similarity calculation algorithm, such as traditional dynamic time warping. This quantification index measures the similarity between the waveform to be identified and the target standard waveform, ranging from 0 to 1, with values ​​closer to 1 indicating higher similarity. A preset first threshold is, for example, 0.9 ± a small fluctuation value, and a second threshold is 0.8 ± a small fluctuation value. The similarity quantification value between the waveform to be identified and each target standard waveform is calculated. If the calculated similarity quantification value between the waveform to be identified and a target standard waveform is 0.95, which is greater than the first threshold of 0.9, the waveform to be identified is determined to be abnormal, and an audible and visual alarm is immediately triggered, with a highlighted message on the screen. If the similarity quantification value between the waveform to be identified and a target standard waveform is 0.85, which is greater than or equal to the second threshold of 0.8 and less than or equal to the first threshold of 0.9, log information is generated based on the waveform to be identified, the similarity quantification value, and the corresponding target standard waveform, recording relevant data for later review. If the similarity quantification value between the waveform to be identified and all target standard waveforms is 0.7, which is less than the first threshold of 0.9, the waveform to be identified is determined to be normal, and no alarm operation is triggered.

[0063] As can be seen from the above, this embodiment improves the accuracy of alarms by setting thresholds to divide similarity levels and adopting different processing methods for different levels, while recording information of intermediate levels to facilitate subsequent clinical analysis.

[0064] Considering that physicians need to continuously monitor the screen for waveform changes, prolonged operation can easily lead to visual fatigue, increasing the risk of missing abnormal waveforms. Furthermore, the fixed detection trigger frequency (e.g., once every 5 seconds) cannot be customized according to clinical needs (e.g., once every 3 respiratory cycles), which can easily lead to errors due to cross-respiratory-cycle detection or increase system redundancy due to excessively high detection frequency. To address the problem that the fixed detection interval prevents adjustment of the raw ventilation data acquisition frequency according to actual clinical needs, thus failing to meet the monitoring requirements of specific diseases, this application also provides the following solution:

[0065] Before generating the waveform to be identified without amplitude information, when a detection interval setting instruction is received, the number of expiratory cycles is extracted from the detection interval setting instruction; if the current respiratory cycle count reaches the number of expiratory cycles, the acquisition operation of raw ventilation data is triggered.

[0066] The detection interval can be customized by the physician, specifying the time or respiratory cycle condition for triggering the raw ventilation data acquisition operation. The physician sends the detection interval setting command through the anesthesia machine's human-machine interface, specifying 5 expiratory cycles. Upon receiving this command, the electronic device extracts the 5 expiratory cycles and continuously counts the current respiratory cycles. When the count reaches 5, the electronic device triggers the raw ventilation data acquisition operation, beginning the collection of raw ventilation data for that cycle.

[0067] As can be seen from the above, this embodiment supports users to customize the detection interval according to clinical needs, making the collection of raw ventilation data more flexible and better adaptable to monitoring scenarios of different diseases.

[0068] To address the lack of a customizable standard waveform construction mechanism, which prevents the creation of targeted standard waveforms for specific diseases based on clinical experience, resulting in weak waveform recognition, and further, the absence of a waveform template library that clinical experts can customize and associate with disease information, makes it impossible to link templates to specific diseases (such as "..."). The implementation of the "increased slope - bronchospasm" model is based on the above embodiments. Furthermore, the template management only supports basic CRUD operations and lacks structured storage and grouping capabilities. Therefore, this invention also provides an exemplary implementation method, which may include the following:

[0069] Upon receiving a standard waveform construction instruction, extract the disease name and corresponding ventilation parameter standard values ​​for the disease characteristics from the instruction; perform amplitude reduction processing on the data of each standard point of the ventilation parameter standard value within the amplitude range; generate a standard waveform based on each standard point after amplitude reduction processing, using the ventilation parameter standard value as the vertical axis and the sampling frequency as the horizontal axis, and at least set a disease name label for the standard waveform; and store the standard waveform in the waveform template library.

[0070] The standard waveform construction command is issued by the user through the human-computer interaction interface, while the waveform template library is a database used to store processed standard waveforms, supporting management operations such as adding, deleting, modifying, and querying waveforms. For example, a clinical expert can send a standard waveform construction command through the waveform parameter input page of the human-computer interaction device. The command includes the disease name "bronchospasm" and the standard values ​​of ventilation parameters corresponding to the characteristics of this disease. After receiving the command, the electronic device extracts the disease name "bronchospasm" and the corresponding standard values ​​of ventilation parameters. The data of each standard point within the amplitude range of this standard value of ventilation parameters is processed to remove amplitude information, resulting in standard point data without amplitude information. Using the standard values ​​of ventilation parameters as the vertical axis and the sampling frequency of the anesthesia machine as the horizontal axis, a standard waveform is generated based on the standard points after amplitude removal, and the disease name label "bronchospasm" is set for this standard waveform. The generated standard waveform is stored in the waveform template library for subsequent similarity calculations. Figure 3 The figure shows a smooth standard waveform corresponding to airway pressure.

[0071] As can be seen from the above, this embodiment supports clinical experts in customizing and constructing standard waveforms, and managing their labels and storage, which enriches the content of the waveform template library and improves the targeting and flexibility of waveform recognition.

[0072] Considering that the standard values ​​of ventilation parameters may contain discrete points or be unsmooth, affecting the accuracy of the standard waveform and thus reducing the accuracy of subsequent similarity calculations, this invention also provides the following implementation method, which may include:

[0073] If the numerical difference between adjacent standard points of the ventilation parameter standard value is greater than a preset discrete threshold, the ventilation parameter standard value is interpolated. If the ventilation parameter standard value includes a target standard point whose numerical difference with all standard points in the preset neighborhood is greater than the preset discrete threshold, the target standard point is deleted from the ventilation parameter standard value. From the current ventilation parameter standard values, the largest standard value and the smallest standard value within the amplitude range are selected. Based on the largest and smallest standard values, the standard amplitude range information is determined. For each standard point, based on the current standard point's value within the amplitude range and the smallest standard value, the standard deviation information between the current standard point and the bottom of the amplitude range is determined. Based on the standard deviation information and the standard amplitude range information, the new standard value of the current standard point within the amplitude range is determined. Based on the new standard values ​​of each standard point, a standard waveform without amplitude information is generated.

[0074] In this embodiment, for example, the standard values ​​of ventilation parameters corresponding to the symptom of "bronchospasm" are obtained, assuming that the amplitude range is [8, 28], and the preset discrete threshold is 5. Check the adjacent standard points of this ventilation parameter standard value. If the difference between two adjacent standard points is 7... Greater than the discrete threshold 5 Then, linear interpolation is performed on the standard values ​​of ventilation parameters between these two points to supplement data points and make the numerical transition smoother. If there is a target standard point in the standard values ​​of ventilation parameters, the difference between its value and the values ​​of all standard points in the preset neighborhood is 8. Greater than the discrete threshold 5 If the target standard point is not found, it will be removed from the ventilation parameter standard values. From the processed ventilation parameter standard values, the maximum standard value 28 within the amplitude range [8, 28] is selected. and minimum standard value 8 The calculated standard amplitude range information is 28-8=20. For each standard point, such as a standard point with a value of 18... The standard deviation between the value and the bottom of the amplitude range is calculated to be 18-8=10. Dividing this difference information by the standard amplitude range information yields a new standard value of 10 / 20 = 0.5 for the standard point. Based on the new standard values ​​of all standard points, a standard waveform for the symptom of "bronchospasm" without amplitude information is generated.

[0075] As can be seen from the above, this embodiment generates a smooth standard waveform without amplitude information by performing discrete point processing and amplitude removal processing on the standard values ​​of ventilation parameters, thereby improving the accuracy of the standard waveform and providing a reliable basis for comparison in similarity calculation.

[0076] Considering that traditional similarity calculation methods cannot accurately measure the shape similarity between waveforms with different total numbers of data points, and the calculation results have large errors, this embodiment also provides a method for calculating waveform shape similarity, which may include the following:

[0077] A local cost matrix is ​​constructed based on the squared differences between each data point of the waveform to be identified and the standard points of the current target standard waveform. The matrix dimension of the local cost matrix is ​​determined based on the total number of data points of the waveform to be identified and the number of standard points of the current target standard waveform. A cumulative cost matrix is ​​constructed by taking the sum of the smallest element among the surrounding elements of each element of the local cost matrix as a new matrix element. Based on the cumulative cost matrix, the optimal alignment value between the waveform to be identified and the current target standard waveform is determined. The sum of the squared differences of each optimal alignment value is calculated, and the initial similarity value between the waveform to be identified and the current target standard waveform is determined based on the ratio of the sum to the total logarithm of the optimal alignment values. The quantification value of the similarity between the waveform to be identified and the current target standard waveform is determined based on the difference between the initial similarity value and the preset benchmark value.

[0078] In this embodiment, the waveform to be identified and any target standard waveform can be represented as follows: It can be based on the relational formula Calculate the squared difference between all data points, and then... Organized into an N*M matrix, we obtain the local cost matrix. Here, X and Y represent two sequences. Sequence X has N points, each with a value from x1, x2 to xn. Sequence Y has M points, each with a value from y1, y2 to ym. xi represents any point in sequence X, so i ranges from 1 to N; yj represents any point in sequence Y, so j ranges from 1 to M. Since the amplitude effects of the two waveforms have been removed, xi and yj are both in the range of 0 to 1. di,j represents the squared difference between any point in sequence X and any point in sequence Y, which can be defined as the cost. There are N*M possible values ​​for di,j. Through the relational formula... Using boundary conditions as To construct a cumulative cost matrix D of the same size, This is the total squared difference between the two sequences after alignment along the optimal path. D is the cumulative cost matrix, also N*M, but its meaning differs from di,j. Each cell in the diagram represents a Di,j. Di,j is the sum of the minimum costs from point i,j back to point 1,1. Therefore, Di,j is actually the cost di,j of the current point (i,j) plus the minimum value among the three cumulative costs to its left, bottom left, and bottom. Thus, DN,M is the minimum cost sum, which in practice represents the optimal alignment value of the two waveforms. Starting from (N,M), backtracking along the direction that minimizes the previous state to (1,1) yields the alignment path index sequence. Where K is the path length (number of pairs of points after alignment), that is, after obtaining the D matrix, we can follow the path from the top right corner Dm,n to D1,1, thus obtaining a series of point sets {ik,jk}. The curly braces {} represent an ordered sequence, that is, each pair (ik,jk) inside the braces is strung together from 1 to K according to k, and (ik,jk) is the pair inside the parentheses. In the diagram, a black point represents a pair (ik,jk), as shown below. Figure 4 As shown, k represents the k-th alignment point. In the diagram, k equals 20 and consists of two integers: ik is the selected index (ik-th sample) in the sequence X = (x1, x2, ..., xN); jk is the selected index (jk-th sample) in the sequence Y = (y1, y2, ..., yM). In other words, this pairing represents considering xik and yjk as "aligned" together. The subscripts k = 1 to K indicate that there are K pairs in the sequence, with k increasing from 1 to K. The length of the K-path is the total number of alignment pairs. Once the optimal alignment value is determined, it can be determined according to... The sum of the squared differences of each optimal alignment value is calculated, and `Draw` is the sum of the distances between the selected points. This means that the sum of the squared differences of all pairs along the path is used as the total mismatch cost for that path. Because it has been normalized, the values ​​of `di` and `j` range from 0 to 1, so the maximum sum of the k costs is k. `Dnorm` is the average squared error of all paired points along the DTW alignment path (i.e., the sum of the squared differences at each step divided by the number of steps K). Its value ranges from 0 to 1. Intuitively, the smaller the value, the more similar the points are.

[0079] To ensure that the distance is independent of the sequence length while remaining within a certain range, a relational expression can be used. Dividing the total squared difference by the path length generally indicates a higher similarity score, suggesting a better match. This can be based on a relational formula. Then perform a reverse transformation. Here, S replaces "distance / error" with a similarity score, ranging from 0 to 1, where 1 represents complete similarity and 0 represents the greatest degree of dissimilarity.

[0080] For example, assuming the total number of data points for the waveform to be identified is 180, and the total number of standard points for the current target standard waveform is 200, a 180×200 local cost matrix is ​​constructed based on the squared differences between each data point of the waveform to be identified and the standard points of the current target standard waveform. A cumulative cost matrix is ​​constructed by using the sum of the smallest element among the surrounding elements of each element of the local cost matrix as the new matrix element. Based on the cumulative cost matrix, the optimal alignment value between the waveform to be identified and the current target standard waveform is determined by tracing back from the bottom right element to the top left element. The sum of the squared differences of all optimal alignment values ​​is calculated; assuming the sum is 36 and the total number of optimal alignment values ​​is 180, the initial similarity value is determined to be 0.2 based on the ratio of the sum to the total number of alignment values ​​(36 / 180 = 0.2). The preset baseline value is 0.8. Based on the difference between the initial similarity value and the baseline value (0.8 - 0.2 = 0.6), the quantified similarity value between the waveform to be identified and the current target standard waveform is determined to be 0.6.

[0081] As can be seen from the above, this embodiment can effectively handle the problem of waveform similarity calculation with different total number of data points, improve the accuracy of similarity quantification values, and provide a reliable basis for waveform anomaly judgment.

[0082] It should be noted that there is no strict order of execution for the steps in this application. As long as they conform to a logical order, these steps can be executed simultaneously or in a certain preset order. Figure 1 This is just an illustrative example and does not mean that this is the only possible execution order.

[0083] This application also provides a corresponding apparatus for the ventilation data processing method, further enhancing the practicality of the method. The apparatus can be described from both a functional module perspective and a hardware perspective. The ventilation data processing apparatus provided in this application is described below. This apparatus is used to implement the ventilation data processing method provided in this application. In this embodiment, the ventilation data processing apparatus may include or be divided into one or more program modules. These one or more program modules are stored in a storage medium and executed by one or more processors to complete the ventilation data processing method disclosed in Embodiment 1. The program module referred to in this embodiment refers to a series of computer program instruction segments capable of performing specific functions, which are more suitable than the program itself for describing the execution process of the ventilation data processing apparatus in the storage medium. The following description will specifically introduce the functions of each program module in this embodiment. The ventilation data processing apparatus described below can be referred to in correspondence with the ventilation data processing method described above.

[0084] From the perspective of functional modules, see Figure 5 , Figure 5 This is a structural diagram of the ventilation data processing device provided in this embodiment under one specific implementation. The device may include:

[0085] The waveform generation module 501 is used to generate a waveform to be identified without amplitude information based on the raw ventilation data of the current acquisition cycle.

[0086] The similarity recognition module 502 is used to calculate the degree of similarity between the waveform to be identified and each target standard waveform. Each target standard waveform is a waveform without amplitude information generated based on the standard values ​​of ventilation parameters according to different disease characteristics.

[0087] The waveform anomaly identification module 503 is used to determine whether the waveform to be identified is abnormal based on the degree of similarity.

[0088] For example, in some embodiments of this embodiment, the waveform generation module 501 can also be used to: select the maximum value of the ventilation parameter with the largest value and the minimum value of the ventilation parameter with the smallest value from the original ventilation data of the current acquisition cycle; determine the amplitude range information based on the maximum value and the minimum value of the ventilation parameter; for each data point of the original ventilation data, determine the difference information between the current data point and the bottom of the amplitude range based on the value of the current data point within the amplitude range and the minimum value of the ventilation parameter, and determine the new value of the current data point within the amplitude range based on the difference information and the amplitude range information; and generate a waveform to be identified without amplitude information based on the new values ​​of each data point of the original ventilation data.

[0089] For example, in some other embodiments of this embodiment, the above-mentioned apparatus may further include a data preprocessing module, which is used to: obtain the total number of waveform points of each target standard waveform; if the total number of data points of the original ventilation data is less than the total number of waveform points, then perform interpolation processing on the original ventilation data so that the total number of data points of the original ventilation data is the same as the total number of waveform points of the target standard waveform; if the total number of data points of the original ventilation data is greater than the total number of waveform points, then perform filtering processing on the original ventilation data, and resample the filtered original ventilation data according to the total number of waveform points.

[0090] For example, in some other embodiments of this embodiment, the waveform anomaly identification module 503 can also be used to: if at least one similarity quantization value is greater than the first threshold, then the waveform to be identified is abnormal; if at least one target similarity quantization value is greater than or equal to the second threshold and less than or equal to the first threshold, then log information is generated based on the waveform to be identified, the target similarity quantization value, and the target standard waveform corresponding to the target similarity quantization value; if there is no similarity quantization value greater than the first threshold, then the waveform to be identified is normal.

[0091] For example, in some other embodiments of this embodiment, the above-mentioned device may further include a detection interval configuration module, which is used to: when a detection interval setting instruction is received, extract the number of expiratory cycles from the detection interval setting instruction; if the current respiratory cycle count reaches the number of expiratory cycles, trigger the acquisition operation of raw ventilation data.

[0092] For example, in some other embodiments of this embodiment, the above-mentioned device may further include a database construction module, which is used to: when a standard waveform construction instruction is received, extract the disease name and the standard values ​​of ventilation parameters corresponding to the corresponding disease characteristics from the standard waveform construction instruction; perform amplitude reduction processing on the data of each standard point of the standard value of ventilation parameters within the amplitude range; generate a standard waveform based on each standard point after amplitude reduction processing, with the standard value of ventilation parameters as the vertical axis value and the sampling frequency as the horizontal axis value, and set at least a disease name label for the standard waveform; and store the standard waveform in a waveform template library.

[0093] As an exemplary implementation of the above embodiments, the database construction module can also be used to: if the numerical difference between adjacent standard points of the ventilation parameter standard value is greater than a preset discrete threshold, then perform interpolation processing on the ventilation parameter standard value; if the ventilation parameter standard value includes a target standard point whose numerical difference with the standard points in the preset neighborhood is greater than the preset discrete threshold, then delete the target standard point from the ventilation parameter standard value; select the maximum standard value and the minimum standard value that are the largest and smallest values ​​within the amplitude range from the current ventilation parameter standard values; determine the standard amplitude range information based on the maximum and minimum standard values; for each standard point, determine the standard deviation information between the current standard point and the bottom of the amplitude range based on the current standard point's value within the amplitude range and the minimum standard value, and determine the new standard value of the current standard point within the amplitude range based on the standard deviation information and the standard amplitude range information, so as to generate a standard waveform without amplitude information based on the new standard values ​​of each standard point.

[0094] For example, in some other embodiments of this embodiment, the similarity recognition module 502 can also be used to: construct a local cost matrix based on the squared differences between each data point of the waveform to be recognized and the standard points of the current target standard waveform; the matrix dimension of the local cost matrix is ​​determined based on the total number of data points of the waveform to be recognized and the number of standard points of the current target standard waveform; construct a cumulative cost matrix by taking the sum of each element of the local cost matrix and the minimum element value among the surrounding elements as a new matrix element; determine the optimal alignment value between the waveform to be recognized and the current target standard waveform based on the cumulative cost matrix; calculate the sum of the squared differences of each optimal alignment value, and determine the initial similarity value between the waveform to be recognized and the current target standard waveform based on the ratio of the sum to the total logarithm of the optimal alignment values; determine the quantification value of the similarity between the waveform to be recognized and the current target standard waveform based on the difference between the initial similarity value and the preset benchmark value.

[0095] The ventilation data processing device mentioned above is described from the perspective of functional modules. Furthermore, this application also provides an electronic device, which is described from the perspective of hardware. Figure 6 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present application. The electronic device includes a memory and a processor. The memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the ventilation data processing method embodiments described above.

[0096] like Figure 6 As shown, the electronic device includes a memory 60 for storing a computer program; and a processor 61 for executing the computer program to implement the steps of the ventilation data processing method as described in any of the above embodiments.

[0097] The processor 61 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 61 may also be a controller, microcontroller, microprocessor, or other data processing chip. The processor 61 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 61 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 61 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 61 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0098] The memory 60 may include one or more computer non-volatile storage media, which may be non-transitory. The memory 60 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the memory 60 may be an internal storage unit of an electronic device, such as a server hard drive. In other embodiments, the memory 60 may be an external storage device of an electronic device, such as a plug-in hard drive on a server, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. Furthermore, the memory 60 may include both internal and external storage units of the electronic device. The memory 60 can be used not only to store application software and various types of data installed on the electronic device, such as code for programs executing the ventilation data processing method, but also to temporarily store data that has been output or will be output. In this embodiment, the memory 60 is used to store at least the following computer program 601, which, after being loaded and executed by the processor 61, is capable of implementing the relevant steps of the ventilation data processing method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 60 may also include an operating system 602 and data 603, and the storage method may be temporary storage or permanent storage. The operating system 602 may include Windows, Unix, Linux, etc. The data 603 may include, but is not limited to, data corresponding to ventilation data processing results.

[0099] In some embodiments, the above-mentioned electronic device may further include a display screen 62, an input / output interface 63, a communication interface 64 (or network interface), a power supply 65, and a communication bus 66. The display screen 62 and the input / output interface 63, such as a keyboard, are user interfaces. Exemplary user interfaces may also include standard wired interfaces, wireless interfaces, etc. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a display screen or display unit, used to display information processed in the electronic device and to display a visual user interface. The communication interface 64 may exemplary include wired and / or wireless interfaces, such as a Wi-Fi interface, a Bluetooth interface, etc., typically used to establish communication connections between the electronic device and other electronic devices. The communication bus 66 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. This bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0100] Those skilled in the art will understand that Figure 6 The structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, such as sensors 67 that perform various functions.

[0101] It is understood that if the ventilation data processing method in the above embodiments is implemented as a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, or all or 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 executes all or part of the steps of the methods in the various embodiments of this application. The aforementioned storage medium includes, but is not limited to, various media capable of storing program code, such as: USB flash drive, mobile hard disk, read-only memory (ROM), random access memory (RAM), electrically erasable programmable ROM, register, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, removable disk, CD-ROM, magnetic disk, or optical disk. Based on this, this application also provides a non-volatile storage medium storing a computer program, which, when executed by a processor, performs the steps of the ventilation data processing method as described in any of the above embodiments.

[0102] It is understood that if the ventilation data processing method in the above embodiments is implemented as a software functional unit and sold or used as an independent product, the computer software product may not need to be stored in a physical storage medium. For example, it can be directly transmitted to a computer or other device with information processing capabilities via a wired or wireless network to execute all or part of the steps of the methods in the various embodiments of this application. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, or all or part of the technical solution, can be embodied in the form of a software product. Based on this, this application also provides a computer program product, which stores a computer program, and when the computer program is executed by a processor, it performs the steps of the ventilation data processing method as described in any of the above embodiments.

[0103] Finally, this application also provides an anesthesia machine, such as Figure 7 As shown, it includes at least the electronic device 701 of the above embodiments. The electronic device 701 can communicate with the anesthesia machine sensor 702, the alarm 703 and the human-machine interaction device 704 respectively. The human-machine interaction device 704 includes at least a waveform parameter input page. The processor of the electronic device 701 communicates with the anesthesia machine sensor 702 through a target hardware interface. After receiving the alarm signal from the processor, the alarm 704 triggers an audible and visual alarm.

[0104] To make the technical solution of this application clearer to those skilled in the art, this invention also provides an implementation of an anesthesia machine. The embedded processor of the anesthesia machine at least encapsulates a waveform generation and template management module for waveform generation and template management, a real-time data acquisition and preprocessing module for real-time data acquisition and preprocessing, a similarity calculation and alarm decision module for similarity calculation and alarm decision, and a detection scheduling module and storage management module for detection scheduling and storage management. The processor is connected to the anesthesia machine sensors, alarm, and storage unit. Hardware interface: A standard CAN interface is used to communicate with the anesthesia machine sensors; the alarm triggers an audible and visual alarm via a relay; Software platform: An embedded Linux system, with a kernel module developed in C / C++ and an upper-level interface implemented in Qt (graphical user interface). Simulation tests and preclinical trials were conducted on multiple mainstream anesthesia machines to verify that the matching algorithm accuracy is >95% and the alarm response latency is <100ms.

[0105] The waveform generation and template management module supports experience data entry, waveform reconstruction, and template naming and grouping functions, and also includes a template library. The experience data entry function allows experienced doctors to input end-tidal values ​​representing symptoms through a dedicated interface. The waveform reconstruction function involves interpolating and filtering the input discrete parameters to generate a smooth, standard waveform. Interpolation and filtering are post-processing steps on the doctor's input data. For example, if the doctor inputs a value of 1 followed by 100, the system will prompt the user to interpolate and smoothly transition from 1 to 100. Alternatively, it may assume the user entered the wrong value and ignore the 100. The template naming and grouping function allows doctors to name each waveform template (e.g., "..."). Increased slope - bronchospasm), setting description and alarm prompt method; Figure 3 For example, the x-axis represents time, and the time interval between points is the machine's sampling interval, such as sampling once every 0.1 seconds. In practice, users do not need to input the x-axis value because the machine's sampling frequency is fixed. Taking airway pressure as an example, the Y-axis represents the pressure value. In practice, users do not need to input the x-axis value, but only the complete y-axis value for one cycle. The number of Y-values ​​input is determined by the user; 100, 200, or even fewer are all acceptable. However, the more points, the more comprehensively the parameter changes within this respiratory cycle can be described, leading to more accurate judgments. The template library stores all templates locally as structured data (timestamp-parameter pairs), which can be added, deleted, modified, and queried.

[0106] Among them, the real-time data acquisition and preprocessing module can support real-time acquisition of airway pressure according to the same usage cycle as the anesthesia machine. Original data such as concentration and flow rate are preprocessed by any one or any combination of the following: data denoising (Kalman filtering), normalization, and then resampled according to the template sampling rate to ensure subsequent matching accuracy. Among them, the resampling process is the interpolation process and the process of deleting values, that is, upsampling and downsampling. This operation is not for the data entered into the template library, but for the machine data actually obtained. For example, if M is the number of waveform points entered in the template, and the actual number of collected source points < M (upsampling): perform linear or spline interpolation on the actual collected source sequence (linear is preferred, cubic / Catmull-Rom is smoother, and the user can actually be allowed to choose). If the actual number of collected source points > M (downsampling): first perform low-pass / anti-aliasing filtering (moving average or Butterworth), and then resample according to M points.

[0107] Among them, the similarity calculation and alarm decision module incorporates the waveform similarity calculation algorithm described in the above embodiments, matches the real-time waveform sequence with each target standard waveform in the template library, can match multiple disease templates simultaneously, or can also be evaluated in order of priority. Finally, a similarity score of 0–1 is output. In addition, a hierarchical alarm strategy is also incorporated. When the similarity > 0.9: high warning, immediately trigger an audible and visual alarm and highlight the prompt on the screen; similarity 0.8 - 0.9: record the log; similarity < 0.8: ignore. Among them, the detection scheduling module is used to support custom detection intervals. Doctors can specify in the setting interface to trigger a complete match once according to the "number of respiratory cycles" or time interval; Scheduler: real-time count the respiratory cycles, call the matching module when the set value is reached, and end this detection at the end of the next exhalation to avoid cross-cycle errors.

[0108] For example, in an operating room of a certain hospital, a doctor is performing general anesthesia surgery on a patient, and uses an anesthesia machine equipped with the electronic device of the present invention to monitor the patient's ventilation, and it is necessary to focus on monitoring whether the patient has bronchospasm.

[0109] Standard waveform construction and storage: Clinical experts input the standard values of ventilation parameters corresponding to the "bronchospasm" disease through the man-machine interaction device of the anesthesia machine. After the electronic device receives the standard waveform construction instruction, it processes the standard value of the parameter. First, check the difference between adjacent standard point values, perform interpolation processing on those with a difference greater than the discrete threshold, and delete isolated discrete target standard points; then determine the standard amplitude range information, perform amplitude removal processing on each standard point, generate a standard waveform without amplitude information, add a "bronchospasm" label to it, and store it in the waveform template library.

[0110] Detection interval setting: The doctor sets the detection interval to 5 respiratory cycles through the man-machine interaction device according to the patient's condition. After the electronic device extracts the number of exhalation cycles, it counts the patient's respiratory cycles in real time.

[0111] Raw ventilation data acquisition and preprocessing: When the number of respiratory cycles reaches 5, the electronic device triggers the acquisition operation, acquiring raw airway pressure ventilation data for that cycle through sensors. The total number of data points is 150, while the total number of data points for the target standard waveform of "bronchospasm" is 200. The electronic device performs linear interpolation on the raw ventilation data, supplementing it with 50 data points to bring the total to 200. Then, the maximum value of 30 from the raw ventilation data is selected. and minimum value 5 The calculation range is 25. The amplitude of each data point is removed to generate the waveform to be identified.

[0112] Similarity calculation: The electronic device uses an improved DTW algorithm to construct the local cost matrix and cumulative cost matrix of the waveform to be identified and the target standard waveform of "bronchospasm", determine the optimal alignment value, and calculate the similarity quantization value to be 0.92.

[0113] Anomaly detection and handling: Since the similarity quantification value of 0.92 is greater than the first threshold of 0.9, the electronic device determines that the waveform to be identified is abnormal and immediately sends an alarm signal to the alarm. The alarm triggers an audible and visual alarm and simultaneously highlights a message on the human-computer interaction device screen indicating that the patient may have bronchospasm, reminding the doctor to take timely intervention measures.

[0114] The above provides a detailed description of a ventilation data processing method, electronic device, and anesthesia machine provided in this application. The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Whether the units and algorithm steps of the various examples described in the disclosed embodiments are executed by electronic hardware or computer software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, and such implementations should not be considered beyond the scope of this application. Several improvements and modifications can be made to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A ventilation data processing method, characterized in that, include: Based on the raw ventilation data of the current acquisition cycle, generate a waveform to be identified that does not contain amplitude information; Calculate the similarity of the waveform shape between the waveform to be identified and each target standard waveform, wherein each target standard waveform is a waveform without amplitude information generated based on standard values ​​of ventilation parameters according to different disease characteristics; Based on the degree of similarity, it is determined whether the waveform to be identified is abnormal.

2. The ventilation data processing method according to claim 1, characterized in that, The step of generating a waveform to be identified without amplitude information based on the raw ventilation data of the current acquisition cycle includes: From the raw ventilation data of the current acquisition cycle, select the maximum value of the ventilation parameter that is the largest within the amplitude range, and the minimum value of the ventilation parameter that is the smallest. The amplitude range information is determined based on the maximum and minimum values ​​of the ventilation parameters; For each data point of the original ventilation data, the difference information between the current data point and the bottom of the amplitude range is determined based on the value of the current data point within the amplitude range and the minimum value of the ventilation parameter. Then, the new value of the current data point within the amplitude range is determined based on the difference information and the amplitude range information. Based on the new values ​​of each data point in the original ventilation data, a waveform to be identified without amplitude information is generated.

3. The ventilation data processing method according to claim 1, characterized in that, Before generating the waveform to be identified without amplitude information, the process also includes: Obtain the total number of waveform points for each target standard waveform; If the total number of data points in the original ventilation data is less than the total number of waveform points, then the original ventilation data is interpolated to make the total number of data points in the original ventilation data the same as the total number of waveform points in the target standard waveform. If the total number of data points in the original ventilation data is greater than the total number of waveform points, the original ventilation data is filtered, and the filtered original ventilation data is resampled according to the total number of waveform points.

4. The ventilation data processing method according to claim 1, characterized in that, The step of determining whether the waveform to be identified is abnormal based on the similarity includes: If at least one similarity quantization value is greater than the first threshold, then the waveform to be identified is abnormal; If at least one target similarity quantization value is greater than or equal to the second threshold and less than or equal to the first threshold, then log information is generated based on the waveform to be identified, the target similarity quantization value, and the target standard waveform corresponding to the target similarity quantization value. If there is no similarity quantization value greater than the first threshold, then the waveform to be identified is normal.

5. The ventilation data processing method according to claim 1, characterized in that, Before generating the waveform to be identified without amplitude information based on the raw ventilation data of the current acquisition cycle, the following steps are also included: When a detection interval setting instruction is received, the number of expiratory cycles is extracted from the detection interval setting instruction; If the current respiratory cycle count reaches the expiratory cycle count, the raw ventilation data acquisition operation is triggered.

6. The ventilation data processing method according to any one of claims 1 to 5, characterized in that, Before calculating the similarity of waveform shape between the waveform to be identified and each target standard waveform, the method further includes: When a standard waveform construction instruction is received, the standard values ​​of ventilation parameters corresponding to the disease name and the corresponding disease characteristics are extracted from the standard waveform construction instruction. The data of each standard point of the ventilation parameter standard value within the amplitude range are subjected to amplitude reduction processing; Using standard values ​​of ventilation parameters as the vertical axis and sampling frequency as the horizontal axis, standard waveforms are generated based on each standard point after amplitude removal processing, and at least a disease name label is set for the standard waveform. The standard waveform is stored in the waveform template library.

7. The ventilation data processing method according to claim 6, characterized in that, The step of removing amplitude values ​​from the data at each standard point of the ventilation parameter standard value within the amplitude range includes: If the numerical difference between adjacent standard points of the ventilation parameter standard value is greater than a preset discrete threshold, then the ventilation parameter standard value is interpolated. If the ventilation parameter standard value includes a target standard point whose numerical difference from the standard point in the preset neighborhood is greater than a preset discrete threshold, then the target standard point is deleted from the ventilation parameter standard value. From the current standard values ​​of ventilation parameters, select the maximum standard value with the largest value and the minimum standard value with the smallest value that are within the amplitude range; Based on the maximum standard value and the minimum standard value, determine the standard amplitude range information; For each standard point, the standard deviation information between the current standard point and the bottom of the amplitude interval is determined based on the value of the current standard point within the amplitude interval and the minimum standard value. Then, the new standard value of the current standard point within the amplitude interval is determined based on the standard deviation information and the standard amplitude range information. Based on the new standard values ​​of each standard point, a standard waveform without amplitude information is generated.

8. The ventilation data processing method according to any one of claims 1 to 5, characterized in that, The calculation of the similarity in waveform shape between the waveform to be identified and each target standard waveform includes: A local cost matrix is ​​constructed based on the squared differences between each data point of the waveform to be identified and the standard point of the current target standard waveform; the matrix dimension of the local cost matrix is ​​determined based on the total number of data points of the waveform to be identified and the number of standard points of the current target standard waveform. The cumulative cost matrix is ​​constructed by taking the sum of each element of the local cost matrix and the minimum value of the surrounding elements as a new matrix element; Based on the cumulative cost matrix, the optimal alignment value between the waveform to be identified and the current target standard waveform is determined; Calculate the sum of the squared differences of each optimal alignment value, and determine the initial similarity value between the waveform to be identified and the current target standard waveform based on the ratio of the sum to the total logarithm of the optimal alignment values; Based on the difference between the initial similarity value and the preset benchmark value, a quantitative value of the similarity between the waveform to be identified and the current target standard waveform is determined.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the ventilation data processing method as described in any one of claims 1 to 8 when executing the computer program.

10. An anesthesia machine, characterized in that, Includes the electronic device as described in claim 9; the electronic device is respectively connected to the anesthesia machine sensor, alarm, and human-machine interaction device; The human-computer interaction device includes at least a waveform parameter input page; The processor of the electronic device communicates with the anesthesia machine sensor through a target hardware interface. After receiving the alarm signal from the processor, the alarm triggers an audible and visual alarm.

Images