A method for monitoring fetal health status based on electrophysiological signals

Abstract

The present invention relates to the field of medical prenatal examination, and specifically to a method for monitoring the health status of a fetus based on electrophysiological signals. The method includes: extracting electrocardiogram signals; real-time monitoring; performing credibility scoring; calculating the fetal health index; obtaining electrocardiogram waveform analysis results; and final evaluation. The present invention calculates and compares the maternal characteristics of each interception time period, and then performs credibility assessment on the fetal electrocardiogram signal; the maternal electrocardiogram signal can provide an important basis for the credibility assessment of the fetal electrocardiogram signal, thereby effectively eliminating the interference of maternal physiological changes on the fetal signal and improving the accuracy of the assessment; by dividing the total signal acquisition time into multiple interception time periods of the same length, an independent credibility score is performed for the fetal electrocardiogram signal in each time period; this method makes the monitoring results more detailed and can reduce the impact of maternal characteristics on the monitoring of the health status of the fetus.

Description

Technical Field

[0001] The present invention relates to the field of medical prenatal examination, and in particular to a method for monitoring the health status of a fetus based on electrophysiological signals. Background Art

[0002] Fetal health monitoring occupies an important position in obstetric medicine. Traditional methods mainly rely on fetal heart rate monitoring and ultrasonic examinations. Although these methods can provide certain fetal health information, they often have problems such as lack of real-time performance, single monitoring indicators and limited evaluation accuracy. With the advancement of technology, monitoring methods based on electrophysiological signals have gradually attracted attention, especially the analysis of fetal electrocardiogram signals (FECG).

[0003] However, existing fetal ECG signal monitoring methods often lack sufficient consideration of maternal physiological characteristics when evaluating the health status of the fetus; this leads to large differences in the credibility assessment of fetal ECG signals under different maternal conditions, which in turn affects the final health assessment results; in addition, many methods only focus on a single heart rate or heart rate variability (HRV), and fail to comprehensively consider multiple physiological parameters, resulting in insufficient comprehensiveness of the monitoring results; based on the above problems, the present invention proposes a fetal health status monitoring method based on electrophysiological signals. Summary of the invention

[0004] The present invention calculates and compares the maternal characteristics of each intercepted time period, and then evaluates the credibility of the fetal electrocardiogram signal; the maternal electrocardiogram signal can provide an important basis for the credibility evaluation of the fetal electrocardiogram signal, thereby effectively eliminating the interference of maternal physiological changes on the fetal signal and improving the accuracy of the evaluation; by dividing the total signal acquisition time into multiple intercepted time periods of equal length, an independent credibility score is performed on the fetal electrocardiogram signal in each time period; this method makes the monitoring results more detailed and can reduce the impact of maternal characteristics on the monitoring of the fetal health status.

[0005] A method for monitoring fetal health status based on electrophysiological signals, comprising:

[0006] Collecting superimposed signals, the superimposed signals include maternal electrocardiogram signals and fetal electrocardiogram signals; extracting fetal electrocardiogram signals and maternal electrocardiogram signals from the superimposed signals respectively using an independent component analysis algorithm;

[0007] Based on the fetal ECG signal, the fetal heart rate is obtained, and the fetal heart rate variability is obtained at the end of each preset time window; for each first monitoring time point, if the obtained fetal heart rate is within the preset fetal heart rate normal range to , no operation; if the fetal heart rate is to or to Then determine whether all fetal heart rates obtained within the current preset buffer time are within to or to If yes, an alarm is triggered; if no, no action is taken; if the fetal heart rate is greater than or less than The alarm is triggered immediately;

[0008] For each second monitoring time point, determine whether the obtained fetal heart rate variability is within a preset normal range of fetal heart rate variability, if yes, do nothing; if no, trigger an alarm;

[0009] After the fetal ECG signal acquisition is completed, the total signal acquisition time is divided into m interception time periods of equal length based on the total signal acquisition time, and the intercepted maternal ECG signals and intercepted fetal ECG signals corresponding to each interception time period are extracted; based on all the extracted intercepted maternal ECG signals, each intercepted fetal ECG signal is scored for credibility; based on all the intercepted fetal ECG signals and the corresponding credibility scores, the fetal health index is calculated and obtained;

[0010] The fetal ECG signal is converted into a fetal ECG using a fetal ECG monitoring device, and part of the fetal ECG is selected as the input of the ECG waveform analysis model to obtain the ECG waveform analysis results; the ECG waveform analysis results and the fetal health index are combined to make a final assessment of the fetal health status.

[0011] Preferably, based on all extracted intercepted maternal ECG signals, each intercepted fetal ECG signal is scored for credibility, and the specific operation is as follows:

[0012] Extract the maternal ECG signal and use the formula Calculate the long-term maternal heart rate variability value; represents the i-th heartbeat interval in the maternal ECG signal, i=1, 2, …, N; is the average value of N heartbeat intervals; at the same time, using the formula Calculate and obtain the overall value of the mother's heart rate;

[0013] For any interception time period, use the formula Calculate and obtain the short-term maternal heart rate variability value of the current interception time period; n represents the number of heartbeat intervals in the current interception time period; where Indicates the jth heartbeat interval in the current interception time period, j=1, 2, ..., n; is the average value of all heartbeat intervals in the current interception time period; using the formula Calculate the heart rate variability credibility score of the current intercepted time period; at the same time, use the formula Calculate the short-term value of the mother's heart rate in the current interception time period; by formula Calculate and obtain the heart rate credibility score of the current intercepted time period;

[0014] Based on the heart rate variability credibility score and heart rate credibility score obtained above, the sum of the heart rate variability credibility score and the heart rate credibility score is calculated, and the maternal psychological state scores of the m intercepted maternal ECG signals are obtained respectively. The lower the maternal psychological state score, the higher the degree of abnormal psychological state of the mother; the m maternal psychological state scores obtained are used as the credibility scores of the m intercepted fetal ECG signals , k=1, 2,…, m.

[0015] Preferably, based on all intercepted fetal electrocardiogram signals and corresponding credibility scores, the fetal health index is calculated and obtained, and the specific operations are as follows:

[0016] For m interception time periods, use the formula Calculate the weight of the kth intercept time period using the formula Calculate the fetal comprehensive heart rate value, where is the fetal heart rate corresponding to the kth interception time period; using the formula Calculate the fetal comprehensive heart rate variability value, where is the fetal heart rate variability corresponding to the kth interception time period; using the formula Calculate the ideal fetal heart rate, where and are the maximum and minimum values ​​of the normal range of fetal heart rate respectively; using the formula Calculate the ideal fetal heart rate variability, where and are the maximum and minimum values ​​of the normal range of fetal heart rate variability; Calculate the fetal health index.

[0017] Preferably, the total signal acquisition time is divided into m interception time periods of equal length, wherein the value of m is obtained by a genetic algorithm.

[0018] Preferably, the value of m is obtained by a genetic algorithm, and the specific operation is as follows:

[0019] Step 1: Set the population size, randomly generate x candidate m values, form the randomly generated x candidate m values ​​into an initial population, and set the maximum number of iterations;

[0020] Step 2: For any candidate m value, use the formula The fitness value U of the candidate m value is calculated, where V represents the size of the candidate m value, and W represents the standard deviation of the fetal heart rate variability corresponding to each intercepted time period when the candidate m value is used to segment the total signal acquisition time. and are the reciprocals of the candidate m values. and the weight coefficient of W. This formula indicates that the optimization goal of the candidate m value is to minimize the size of the m value and maximize the difference in fetal heart rate variability in each interception time period; the higher the calculated fitness value, the better the performance of the candidate m value;

[0021] Step 3: Apply the roulette wheel selection method to select the corresponding number of candidate m values ​​with the best performance according to the preset selection ratio, and form the selected candidate m values ​​into a selection individual set;

[0022] Step 4: Apply the preset crossover probability to perform a crossover operation on the candidate m values ​​in the selected individual set to generate a corresponding number of new candidate m values, and form the obtained new candidate m values ​​into a crossover individual set; randomly select candidate m values ​​from the crossover individual set, and make a small adjustment to the candidate m values ​​according to the preset mutation probability. The adjusted crossover individual set is the mutation individual set;

[0023] Step 5: Merge the selected individual set and the mutated individual set to obtain the merged individual set, calculate the fitness of all candidate m values ​​in the merged individual set, and take the x candidate m values ​​with the best fitness performance as the new generation population;

[0024] Step 6: Repeat steps 3 to 5 to iteratively update the population until the maximum number of iterations is reached. The population with the best fitness performance in the last generation is the optimal m value.

[0025] Preferably, a portion of the fetal electrocardiogram is selected as the input of the electrocardiogram waveform analysis model, and the specific operation is as follows:

[0026] A credibility threshold is set, and an intercepted fetal electrocardiogram corresponding to an intercepted fetal electrocardiogram signal having a credibility score higher than the credibility threshold is selected as an input of the electrocardiogram waveform analysis model.

[0027] Preferably, the ECG waveform analysis model is established based on the CNN model, including an input layer, a convolution layer, a pooling layer, a fully connected layer and an output layer: the input layer is used to input the preprocessed ECG; the convolution layer is used to extract the features present in the ECG; the pooling layer is used to reduce the data dimension by downsampling, reduce the amount of calculation, reduce the risk of overfitting, and retain the most important features; the fully connected layer is used to quantize the features output by the pooling layer, integrate the image features, and associate them with the final classification results; the output layer is used to generate the probability distribution of each category according to the output of the fully connected layer, and give the final ECG waveform analysis result.

[0028] Preferably, the training of the ECG waveform analysis model is specifically performed as follows:

[0029] Obtain several ECG samples annotated with ECG waveform analysis results; divide all ECG samples into a training set and a test set, use the training set to train the ECG waveform analysis model, and then input the test set into the ECG waveform analysis model to obtain the ECG waveform analysis test results; set the training conditions, if the ECG waveform analysis test results meet the training conditions, output the trained ECG waveform analysis model; otherwise, continue to use the training set to train the ECG waveform analysis model.

[0030] The present invention has the following advantages:

[0031] The present invention calculates and compares the maternal characteristics of each intercepted time period, and then evaluates the credibility of the fetal electrocardiogram signal; the maternal electrocardiogram signal can provide an important basis for the credibility evaluation of the fetal electrocardiogram signal, thereby effectively eliminating the interference of maternal physiological changes on the fetal signal and improving the accuracy of the evaluation; by dividing the total signal acquisition time into multiple intercepted time periods of equal length, an independent credibility score is performed on the fetal electrocardiogram signal in each time period; this method makes the monitoring results more detailed and can reduce the impact of maternal characteristics on the monitoring of the fetal health status.

[0032] The present invention can provide a comprehensive fetal health assessment by calculating a comprehensive fetal health index; this index not only includes traditional heart rate and heart rate variability indicators, but also combines the physiological characteristics of the mother and the signal credibility scores of each time period, making the assessment results more comprehensive and in-depth; this comprehensive assessment helps doctors better understand the health status of the fetus, discover potential health problems in a timely manner, and take corresponding intervention measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a flow chart of a method for monitoring fetal health status based on electrophysiological signals adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0035] A method for monitoring fetal health status based on electrophysiological signals, such as Figure 1 As shown, including:

[0036] Collecting superimposed signals, which contain maternal ECG signals and fetal ECG signals. Collecting superimposed signals is the starting point of the monitoring method and provides raw data for subsequent signal processing and analysis. This step is necessary because the ECG signals of the fetus are usually weak and easily interfered by the maternal ECG signals when the fetus is in the mother's body, so it is necessary to capture these signals through sensors. The fetal ECG signals and maternal ECG signals are extracted from the superimposed signals respectively using the independent component analysis algorithm. Independent component analysis (ICA) is a signal processing technology used to separate independent signal sources from multi-source signals. In the present invention, the ICA algorithm is used to process the collected superimposed signals in order to separate the fetal ECG signals from the maternal ECG signals. The ICA algorithm can more clearly identify and extract the fetal ECG signals, which is crucial for subsequent analysis because clear signals can provide more accurate health information. At the same time, the ICA algorithm helps to eliminate or reduce the interference of maternal signals on fetal signals, which is very necessary for accurately assessing the health status of the fetus. This step is directly completed by the fetal ECG monitor.

[0037] Based on the fetal ECG signal, the fetal heart rate is obtained, and the fetal heart rate variability is obtained at the end of each preset time window. The fetal ECG signal is an important physiological signal reflecting the fetal heart activity; by analyzing these signals, the frequency of the fetal heartbeat, that is, the fetal heart rate, can be detected; the heart rate is one of the key indicators for assessing the health of the fetus, because it can reflect the physiological state of the fetus and possible abnormalities; obtaining the fetal heart rate is the basis for monitoring the health of the fetus; the normal heart rate range usually indicates that the fetus is in a healthy state, while an abnormal heart rate (too high or too low) may indicate certain health problems, such as hypoxia, heart abnormalities, etc.; heart rate variability is Refers to the natural fluctuation of the heartbeat interval time; heart rate variability is an important indicator to measure the regulatory ability of the cardiac autonomic nervous system, and is closely related to heart health and overall health status; in the present invention, by calculating heart rate variability at the end of a preset time window, the cardiac regulatory ability and health status of the fetus can be evaluated; the analysis of heart rate variability helps to have a more comprehensive understanding of the fetal heart function and the activity of the autonomic nervous system; normal heart rate variability indicates that the fetal heart regulation mechanism is normal, while abnormal heart rate variability may indicate potential health problems; for each first monitoring time point, if the fetal heart rate obtained is within the preset fetal heart rate normal range to , no operation; if the fetal heart rate is to or to , then determine whether all fetal heart rates obtained within the current preset buffer time are within to or to If yes, an alarm is triggered; if no, no action is taken; if the fetal heart rate is greater than or less than , an alarm is triggered immediately; this real-time monitoring and response mechanism can promptly detect and deal with possible fetal health problems, thereby improving the safety and health of the fetus;

[0038] For each second monitoring time point, determine whether the obtained fetal heart rate variability is within the preset normal range of fetal heart rate variability. If yes, no operation is performed; if not, this may indicate that the fetus's autonomic nervous system regulation ability is impaired or there are other health problems, and an alarm is triggered;

[0039] When the fetal ECG signal acquisition is completed, based on the total signal acquisition time, the total signal acquisition time is divided into m interception time periods of equal length, and the intercepted maternal ECG signal and the intercepted fetal ECG signal corresponding to each interception time period are extracted; after the fetal ECG signal acquisition is completed, in order to analyze the signal more carefully, the entire acquisition cycle is divided into multiple time periods of equal length; this can more accurately capture the physiological changes of the fetus in different time periods; by dividing the long-term signal into multiple short time periods, the impact of noise and interference that may exist in the long-term signal can be reduced, and the accuracy of signal analysis can be improved; at the same time, this method helps to identify short-term changes in the fetal physiological state, which is helpful for timely discovery and handling of potential problems. Provide support; by extracting maternal and fetal signals separately, the physiological state of the fetus can be analyzed more accurately, while taking into account the impact of the maternal state on the signal; based on all extracted intercepted maternal ECG signals, each intercepted fetal ECG signal is scored for credibility; the credibility scoring mechanism can identify and exclude fetal signal abnormalities caused by abnormal maternal emotions or states, ensuring the accuracy of the evaluation results; at the same time, this also helps to reduce the impact of abnormal maternal data on the difficulty of signal extraction and improve the efficiency of signal processing; based on all intercepted fetal ECG signals and the corresponding credibility scores, the fetal health index is calculated; by calculating the fetal health index, a comprehensive fetal health assessment tool can be provided for medical staff. This index not only takes into account the physiological parameters of the fetus, but also the credibility of the signal, making the evaluation results more comprehensive and reliable;

[0040] Fetal ECG signals are converted into fetal ECGs using fetal ECG monitoring equipment. ECGs are an important tool for analyzing cardiac electrophysiological activity. By converting ECG signals into ECGs, medical staff can visually observe fetal cardiac activity, including heart rate, heart rhythm, and possible abnormal waveforms. For example, the P wave represents the electrical activity of the atrium, which is a rounded wavelet under normal circumstances. If the P wave is abnormally tall, wide, or split, it may indicate atrial enlargement or atrial conduction abnormality. Part of the fetal ECG is selected as the input of the ECG waveform analysis model to obtain the ECG waveform analysis results. Selecting representative ECG segments can improve the efficiency and accuracy of the analysis. By focusing on those segments that can provide the most information, the fetus can be identified more effectively. The health status of the heart is evaluated by combining the ECG waveform analysis results and the fetal health index to reduce unnecessary data processing; the final assessment of the fetal health status is made by combining the ECG waveform analysis results and the fetal health index; after obtaining the ECG waveform analysis results, these results are combined with the previously calculated fetal health index to make a comprehensive health status assessment; the fetal health index is a comprehensive indicator that reflects the overall health status of the fetus, while the ECG waveform analysis results provide specific information on heart function; by comprehensively considering the ECG waveform analysis results and the fetal health index, a comprehensive and accurate assessment of the fetal health status can be made; this comprehensive assessment helps medical staff make more accurate diagnoses, develop more appropriate treatment plans, and predict the health status of the fetus.

[0041] Based on all the extracted intercepted maternal ECG signals, the credibility of each intercepted fetal ECG signal is scored. The specific operation is as follows:

[0042] Extract the maternal ECG signal and use the formula Calculate the long-term maternal heart rate variability value; represents the i-th heartbeat interval in the maternal ECG signal, i=1, 2, …, N; is the average value of N heartbeat intervals; at the same time, using the formula Calculate and obtain the overall value of the mother's heart rate;

[0043] For any interception time period, use the formula Calculate and obtain the short-term maternal heart rate variability value of the current interception time period; n represents the number of heartbeat intervals in the current interception time period; where Indicates the jth heartbeat interval in the current interception time period, j=1, 2, ..., n; is the average value of all heartbeat intervals in the current interception time period; using the formula Calculate the heart rate variability credibility score of the current intercepted time period; at the same time, use the formula Calculate the short-term value of the mother's heart rate in the current interception time period; by formula Calculate and obtain the heart rate credibility score of the current intercepted time period;

[0044] Based on the heart rate variability credibility score and heart rate credibility score obtained above, the sum of the heart rate variability credibility score and the heart rate credibility score is calculated, and the maternal psychological state scores of the m intercepted maternal ECG signals are obtained respectively. The lower the maternal psychological state score, the higher the degree of abnormal psychological state of the mother; the m maternal psychological state scores obtained are used as the credibility scores of the m intercepted fetal ECG signals , k=1, 2,…, m.

[0045] Based on all intercepted fetal ECG signals and the corresponding credibility scores, the fetal health index is calculated and obtained. The specific operations are as follows:

[0046] For m interception time periods, use the formula Calculate the weight of the kth intercept time period using the formula Calculate the fetal comprehensive heart rate value, where is the fetal heart rate corresponding to the kth interception time period; using the formula Calculate the fetal comprehensive heart rate variability value, where is the fetal heart rate variability corresponding to the kth interception time period; using the formula Calculate the ideal fetal heart rate, where and are the maximum and minimum values ​​of the normal range of fetal heart rate respectively; using the formula Calculate the ideal fetal heart rate variability, where and are the maximum and minimum values ​​of the normal range of fetal heart rate variability; Calculate the fetal health index.

[0047] The total signal acquisition time is divided into m interception time periods of equal length, where the value of m is obtained through a genetic algorithm, which is a search heuristic algorithm that simulates the biological evolution process in nature. It evolves the potential solutions to the problem by simulating natural genetic mechanisms (such as selection, crossover, mutation, etc.) in order to find or approach the optimal solution. The genetic algorithm starts with a set of randomly generated candidate solutions (called a population) and continuously optimizes these solutions through an iterative process. Each generation of the population selects excellent individuals for reproduction by evaluating their fitness. At the same time, crossover and mutation operations are introduced to maintain the diversity of the population and prevent the algorithm from converging to a local optimal solution too early.

[0048] The value of m is obtained through genetic algorithm. The specific operations are as follows:

[0049] Step 1: Set the population size, randomly generate x candidate m values, form the randomly generated x candidate m values ​​into an initial population, and set the maximum number of iterations;

[0050] Step 2: For any candidate m value, use the formula The fitness value U of the candidate m value is calculated, where V represents the size of the candidate m value, and W represents the standard deviation of the fetal heart rate variability corresponding to each intercepted time period when the candidate m value is used to segment the total signal acquisition time. and are the reciprocals of the candidate m values. and the weight coefficient of W. This formula indicates that the optimization goal of the candidate m value is to minimize the size of the m value and maximize the difference in fetal heart rate variability in each interception time period; the higher the calculated fitness value, the better the performance of the candidate m value;

[0051] Step 3: Apply the roulette wheel selection method to select the corresponding number of candidate m values ​​with the best performance according to the preset selection ratio, and form the selected candidate m values ​​into a selection individual set;

[0052] Step 4: Apply the preset crossover probability to perform a crossover operation on the candidate m values ​​in the selected individual set to generate a corresponding number of new candidate m values, and form the obtained new candidate m values ​​into a crossover individual set; randomly select candidate m values ​​from the crossover individual set, and make a small adjustment to the candidate m values ​​according to the preset mutation probability. The adjusted crossover individual set is the mutation individual set;

[0053] Step 5: Merge the selected individual set and the mutated individual set to obtain the merged individual set, calculate the fitness of all candidate m values ​​in the merged individual set, and take the x candidate m values ​​with the best fitness performance as the new generation population;

[0054] Step 6: Repeat steps 3 to 5 to iteratively update the population until the maximum number of iterations is reached. The population with the best fitness performance in the last generation is the optimal m value.

[0055] Select part of the fetal electrocardiogram as the input of the electrocardiogram waveform analysis model. The specific operations are as follows:

[0056] A credibility threshold is set, and an intercepted fetal electrocardiogram corresponding to an intercepted fetal electrocardiogram signal having a credibility score higher than the credibility threshold is selected as an input of the electrocardiogram waveform analysis model.

[0057] The ECG waveform analysis model is established based on the CNN model, including input layer, convolution layer, pooling layer, fully connected layer and output layer: the input layer is used to input the preprocessed ECG; the convolution layer is used to extract the features in the ECG; the pooling layer is used to reduce the data dimension by downsampling, reduce the amount of calculation, reduce the risk of overfitting, and retain the most important features; the fully connected layer is used to quantize the features output by the pooling layer, integrate the image features, and associate them with the final classification results; the output layer is used to generate the probability distribution of each category according to the output of the fully connected layer, and give the final ECG waveform analysis results.

[0058] The specific operations for training the ECG waveform analysis model are as follows:

[0059] Obtain several ECG samples annotated with ECG waveform analysis results; divide all ECG samples into a training set and a test set, use the training set to train the ECG waveform analysis model, and then input the test set into the ECG waveform analysis model to obtain the ECG waveform analysis test results; set the training conditions, if the ECG waveform analysis test results meet the training conditions, output the trained ECG waveform analysis model; otherwise, continue to use the training set to train the ECG waveform analysis model.

[0060] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A method for monitoring fetal health status based on electrophysiological signals, characterized in that: include: Collecting superimposed signals, the superimposed signals include maternal electrocardiogram signals and fetal electrocardiogram signals; The independent component analysis algorithm is used to extract the fetal ECG signal and the maternal ECG signal from the superimposed signal. Based on the fetal ECG signal, the fetal heart rate is obtained, and the fetal heart rate variability is obtained at the end of each preset time window; for each first monitoring time point, if the obtained fetal heart rate is within the preset fetal heart rate normal range to , no operation; if the fetal heart rate is to or to , then determine whether all fetal heart rates obtained within the current preset buffer time are within to or to If yes, an alarm is triggered; if no, no action is taken; if the fetal heart rate is greater than or less than , then the alarm is triggered immediately; For each second monitoring time point, determine whether the obtained fetal heart rate variability is within a preset normal range of fetal heart rate variability, if yes, do nothing; if no, trigger an alarm; After the fetal ECG signal acquisition is completed, the total signal acquisition time is divided into m interception time periods of equal length based on the total signal acquisition time, and the intercepted maternal ECG signals and intercepted fetal ECG signals corresponding to each interception time period are extracted; based on all the extracted intercepted maternal ECG signals, each intercepted fetal ECG signal is scored for credibility; based on all the intercepted fetal ECG signals and the corresponding credibility scores, the fetal health index is calculated and obtained; Using a fetal ECG monitoring device to convert the fetal ECG signal into a fetal ECG, selecting part of the fetal ECG as the input of the ECG waveform analysis model to obtain the ECG waveform analysis results; combining the ECG waveform analysis results with the fetal health index to make a final assessment of the fetal health status; Based on all the extracted intercepted maternal ECG signals, the credibility of each intercepted fetal ECG signal is scored. The specific operation is as follows: Extract maternal ECG signals and use the formula Calculate the long-term maternal heart rate variability value; represents the i-th heartbeat interval in the maternal ECG signal, i=1, 2, …, N; is the average value of N heartbeat intervals; at the same time, using the formula Calculate and obtain the overall value of the mother's heart rate; For any interception time period, use the formula Calculate and obtain the short-term maternal heart rate variability value of the current interception time period; n represents the number of heartbeat intervals in the current interception time period; where Indicates the jth heartbeat interval in the current interception time period, j=1, 2, ..., n; is the average value of all heartbeat intervals in the current interception time period; using the formula Calculate the heart rate variability credibility score of the current intercepted time period; at the same time, use the formula Calculate the short-term value of the mother's heart rate in the current interception time period; by formula Calculate and obtain the heart rate credibility score of the current intercepted time period; Based on the heart rate variability credibility score and heart rate credibility score obtained above, the sum of the heart rate variability credibility score and the heart rate credibility score is calculated, and the maternal psychological state scores of the m intercepted maternal ECG signals are obtained respectively. The lower the maternal psychological state score, the higher the degree of abnormal psychological state of the mother; the m maternal psychological state scores obtained are used as the credibility scores of the m intercepted fetal ECG signals , k=1,2,…,m; Based on all intercepted fetal ECG signals and the corresponding credibility scores, the fetal health index is calculated and obtained. The specific operations are as follows: For m interception time periods, use the formula Calculate the weight of the kth intercept time period using the formula Calculate the fetal comprehensive heart rate value, where is the fetal heart rate corresponding to the kth interception time period; using the formula Calculate the fetal comprehensive heart rate variability value, where is the fetal heart rate variability corresponding to the kth interception time period; using the formula Calculate the ideal fetal heart rate, where and They are respectively the maximum and minimum values ​​of the normal range of fetal heart rate; Using the formula Calculate the ideal fetal heart rate variability, where and They are the maximum and minimum values ​​of the preset normal range of fetal heart rate variability; By formula Calculate the fetal health index.

2. The method for monitoring the health status of a fetus based on electrophysiological signals according to claim 1, characterized in that: The total signal acquisition time is divided into m interception time periods of equal length, wherein the value of m is obtained through a genetic algorithm.

3. The method for monitoring the health status of a fetus based on electrophysiological signals according to claim 2, characterized in that: The value of m is obtained through genetic algorithm. The specific operations are as follows: Step 1: Set the population size, randomly generate x candidate m values, form the randomly generated x candidate m values ​​into an initial population, and set the maximum number of iterations; Step 2: For any candidate m value, use the formula The fitness value U of the candidate m value is calculated, where V represents the size of the candidate m value, and W represents the standard deviation of the fetal heart rate variability corresponding to each intercepted time period when the candidate m value is used to segment the total signal acquisition time. and are the reciprocals of the candidate m values. and the weight coefficient of W. This formula indicates that the optimization goal of the candidate m value is to minimize the size of the m value and maximize the difference in fetal heart rate variability in each interception time period; the higher the calculated fitness value, the better the performance of the candidate m value; Step 3: Apply the roulette wheel selection method to select the corresponding number of candidate m values ​​with the best performance according to the preset selection ratio, and form the selected candidate m values ​​into a selection individual set; Step 4: Apply the preset crossover probability to perform a crossover operation on the candidate m values ​​in the selected individual set to generate a corresponding number of new candidate m values, and form the obtained new candidate m values ​​into a crossover individual set; randomly select candidate m values ​​from the crossover individual set, and make a small adjustment to the candidate m values ​​according to the preset mutation probability. The adjusted crossover individual set is the mutation individual set; Step 5: Merge the selected individual set and the mutated individual set to obtain the merged individual set, calculate the fitness of all candidate m values ​​in the merged individual set, and take the x candidate m values ​​with the best fitness performance as the new generation population; Step 6: Repeat steps 3 to 5 to iteratively update the population until the maximum number of iterations is reached. The population with the best fitness performance in the last generation is the optimal m value.

4. The method for monitoring fetal health status based on electrophysiological signals according to claim 3, characterized in that: Select part of the fetal electrocardiogram as the input of the electrocardiogram waveform analysis model. The specific operations are as follows: A credibility threshold is set, and an intercepted fetal electrocardiogram corresponding to an intercepted fetal electrocardiogram signal having a credibility score higher than the credibility threshold is selected as an input of the electrocardiogram waveform analysis model.

5. The method for monitoring fetal health status based on electrophysiological signals according to claim 4, characterized in that: The ECG waveform analysis model is established based on the CNN model, including input layer, convolution layer, pooling layer, fully connected layer and output layer: the input layer is used to input the preprocessed ECG; the convolution layer is used to extract the features in the ECG; the pooling layer is used to reduce the data dimension by downsampling, reduce the amount of calculation, reduce the risk of overfitting, and retain the most important features; the fully connected layer is used to quantize the features output by the pooling layer, integrate the image features, and associate them with the final classification results; the output layer is used to generate the probability distribution of each category according to the output of the fully connected layer, and give the final ECG waveform analysis results.

6. The method for monitoring fetal health status based on electrophysiological signals according to claim 5, characterized in that: The specific operations for training the ECG waveform analysis model are as follows: Obtain several electrocardiogram samples with electrocardiogram waveform analysis results annotated thereon; divide all electrocardiogram samples into a training set and a test set, use the training set to train the electrocardiogram waveform analysis model, and then input the test set into the electrocardiogram waveform analysis model to obtain the electrocardiogram waveform analysis test results; Set the training conditions. If the ECG waveform analysis test results meet the training conditions, output the trained ECG waveform analysis model; otherwise, continue to use the training set to train the ECG waveform analysis model.

Images