A remote monitoring system for obstetrics and gynecology patients

Abstract

The present invention relates to the field of remote metering technology, specifically to a remote monitoring system for obstetrics and gynecology patients, the system comprising: a fetal heart rate data acquisition module to collect fetal heart rate signals, integrate signal timestamps and signal values, establish a fetal heart rate time series data set, adjust the dynamic scale factor of the fetal heart rate time series data set, and perform time series analysis to obtain time series analysis results. In the present invention, by establishing a time series data set and adjusting the dynamic scale factor, data accuracy and processing efficiency are effectively enhanced. The system improves the depth of data analysis and strengthens the ability to recognize periodic patterns by performing multi-scale entropy calculations and performing periodic pattern recognition, which is crucial for early identification of pregnancy problems. The detection of abnormal patterns and comparison with historical health data further enhance the sensitivity and preventiveness of monitoring, which helps to promptly discover and handle abnormal situations.

Description

Technical Field

[0001] The present invention relates to the technical field of remote metering, and in particular to a remote monitoring system for gynecological and obstetric patients. Background Art

[0002] Remote metering technology is a technology that enables remote data collection and monitoring through a communication network. This technology is widely used in the fields of healthcare, energy management, environmental monitoring, etc. A remote metering system can include sensors, data transmission equipment, data processing units, and user interfaces. Sensors are used to collect various types of real-time data, such as body temperature, heart rate, or energy consumption. This data is sent to a central processing system via a wireless or wired network, where a software application analyzes and processes the data and finally displays the data to the user through a user interface. This technology makes data monitoring and management more efficient and accurate, especially in application scenarios that require remote real-time monitoring, such as remote medical monitoring or smart grid management.

[0003] Among them, the obstetrics and gynecology patient remote monitoring system is a medical system that uses remote metering technology, mainly used to monitor the health status of pregnant women and newborns. The system collects key vital signs data in real time through various medical-grade sensors installed on the patient, such as heart rate monitors and blood pressure monitors. These data are transmitted to the server of the medical service provider, and doctors and nurses can remotely view the patient's health status in real time through computers or mobile devices to detect potential health problems in a timely manner. In addition, the system can also be used to guide patients in self-care and improve the quality of medical services during pregnancy and postpartum. Its main purpose is to improve the accessibility and response speed of medical services, especially in remote areas or when access to traditional medical resources is limited.

[0004] In the existing technology, traditional remote monitoring systems usually rely on simple data collection and transmission, lacking in-depth analysis of data and timely pattern recognition. This limits the ability to respond quickly to changes in health conditions. When dealing with large or complex data sets, existing systems may not be able to effectively identify subtle abnormal patterns, resulting in the inability to prevent or identify potential health risks in a timely manner. For fetal heart rate monitoring of pregnant women, traditional methods may not be able to accurately analyze the periodic variation of fetal heart rate, thereby missing critical early warning information. This technical limitation not only reduces the timeliness and effectiveness of medical services, but may also increase unnecessary risks in emergency situations. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a remote monitoring system for obstetrics and gynecology patients.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a remote monitoring system for obstetrics and gynecology patients comprises:

[0007] The fetal heart rate data acquisition module acquires fetal heart rate signals, integrates signal timestamps and signal values, establishes a fetal heart rate time series data set, adjusts a dynamic scale factor of the fetal heart rate time series data set, and performs time series analysis to obtain time series analysis results;

[0008] The pregnancy data analysis module performs multi-scale entropy calculation based on the time series analysis result to obtain the entropy value of each scale, analyzes the fetal heart rate variability through entropy value comparison, obtains the entropy value comparison analysis result, performs periodic pattern recognition based on the entropy value comparison result, and generates a periodic recognition result;

[0009] The abnormal pattern detection module determines the abnormal pattern in the fetal heart rate data based on the periodic identification result, generates an abnormal pattern determination result, compares the abnormal pattern determination result with the historical health data, and generates a health comparison analysis result.

[0010] As a further solution of the present invention, the step of obtaining the time series analysis result is specifically:

[0011] Collect fetal heart rate signals, integrate signal timestamps and signal values, and use the formula:

[0012] S(t i )=(t i ,v i )

[0013] Calculate the signal data pair at each time point to generate a signal data pair set;

[0014] Among them, t i Represents the timestamp, v i Represents the fetal heart rate signal value, S(t i ) represents a set of signal data pairs;

[0015] The signal data pair set is converted into a fetal heart rate time series using the formula:

[0016]

[0017] Calculate the standardized fetal heart rate time series and generate the standardized fetal heart rate time series TS std ;

[0018] Among them, γ is the normalization factor, TS is the fetal heart rate time series;

[0019] The dynamic scaling factor of the standardized fetal heart rate time series is adjusted using the formula:

[0020]

[0021] Calculate and generate the adjusted fetal heart rate time series TSadj ;

[0022] Among them, β is the dynamic adjustment coefficient, max and min are TS std The maximum and minimum values ​​in ;

[0023] Based on the adjusted fetal heart rate time series, a time series analysis is performed using the formula:

[0024]

[0025] Calculate the time series analysis results and generate the time series analysis results R final ;

[0026] Among them, μ and σ are TS adj The mean and standard deviation of .

[0027] As a further solution of the present invention, the step of obtaining the entropy value comparison analysis result is specifically:

[0028] Based on the time series analysis results, multi-scale entropy calculation is performed using the entropy calculation formula:

[0029]

[0030] Generate a set of entropy values ​​for each scale;

[0031] Among them, E(s) represents the entropy value of scale s, p i (s) represents the probability of the state under scale s, λ is the regularization parameter, and k is the total number of states;

[0032] The fluctuation range of the entropy value set is calculated to analyze the fetal heart rate variability using the formula:

[0033]

[0034] Generate fetal heart rate variability measures;

[0035] Where V represents the measure of fetal heart rate variability, represents the average value of the entropy value set, and m represents the number of entropy values;

[0036] Based on the fetal heart rate variability metric, the abnormal changes in the entropy value set are compared and analyzed using the formula:

[0037]

[0038] Generate entropy value comparison analysis result D;

[0039] Among them, max(E(s)) and min(E(s)) represent the maximum and minimum values ​​in E(s).

[0040] As a further solution of the present invention, the step of obtaining the periodic identification result is specifically:

[0041] Based on the entropy comparison and analysis results, the threshold of period identification is determined using the standard deviation and the average value, using the formula:

[0042]

[0043] Generate threshold result θ standard ;

[0044] Among them, μ D , σ D are the mean and standard deviation of D, α, β, and γ are adjustment parameters, and D min is the minimum value in D;

[0045] Apply the threshold result to filter out the scales in the entropy value set that are higher than the threshold, using the formula:

[0046]

[0047] Identify and select entropy scales above the threshold to generate a high entropy scale set H high ;

[0048] Among them, κ, λ, ξ are used to adjust the weight coefficients of threshold calculation;

[0049] Based on the high entropy scale set, statistical analysis is performed to identify periodic patterns using the formula:

[0050] P=mode({s∣s∈H and E(s)>μ E +δ·σ E})

[0051] The calculation scale is used as the dominant periodic mode to generate the periodic identification result P;

[0052] Among them, μ E , σ E are the mean and standard deviation of the entropy value, and δ is a weight coefficient used to increase the calculation accuracy.

[0053] As a further solution of the present invention, the step of obtaining the abnormal mode judgment result is specifically:

[0054] Based on the periodicity identification results, fetal heart rate data is collected and an abnormal pattern recognition algorithm is applied, using the formula:

[0055]

[0056] Calculate the anomaly index of each data point and generate anomaly index array M;

[0057] Among them, p i represents the weight in the periodic identification result, x i represents the fetal heart rate data point, w represents the adjustment weight, and n is the total number of data points;

[0058] Using the anomaly index array, determine the anomaly threshold θ using the formula:

[0059] E={x||M(x)-μ|>θ}

[0060] Filter the data points whose abnormal index exceeds the threshold and generate an abnormal data point set;

[0061] Among them, E represents the set of abnormal data points, θ represents the abnormal threshold, and μ represents the average value of the abnormal index array M;

[0062] Using the set of abnormal data points, the formula is adopted:

[0063] R = classify(E)

[0064] Generate abnormal mode judgment results;

[0065] Among them, R represents the abnormal pattern judgment result, and classify(E) is the classification function.

[0066] As a further solution of the present invention, the steps of obtaining the health comparison analysis results are specifically as follows:

[0067] To obtain the abnormal mode judgment result and historical health data, the formula is used:

[0068]

[0069] Calculate the cosine fit of the abnormal pattern judgment result and the historical health data to generate an intersection data set S;

[0070] Among them, R i The element representing the abnormal mode judgment result, H i represents the elements of historical health data, and n represents the size of the data set;

[0071] Using the intersection dataset, analysis was performed to identify health trends using the formula:

[0072]

[0073] Generate trend analysis results T;

[0074] Among them, S i is the element of the intersection dataset S, α and β represent adjustment parameters, and n is the total number of data points;

[0075] Using the trend analysis results, compared with all historical data, the formula is:

[0076]

[0077] Generate health comparison analysis result C;

[0078] Among them, T i is the element of trend analysis result, H i Elements of historical health data.

[0079] Compared with the prior art, the advantages and positive effects of the present invention are:

[0080] In the present invention, by establishing a time series data set and adjusting the dynamic scale factor, data accuracy and processing efficiency are effectively enhanced. By performing multi-scale entropy calculations and periodic pattern recognition, the system improves the depth of data analysis and strengthens the ability to identify periodic patterns, which is crucial for early identification of pregnancy problems. The detection of abnormal patterns and the comparison of historical health data further improve the sensitivity and preventiveness of monitoring, which helps to detect and deal with abnormal situations in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0081] Figure 1 is a system flow chart of the present invention;

[0082] Figure 2 A flow chart of the steps for obtaining the time series analysis results of the present invention;

[0083] Figure 3 Flow chart of the steps for obtaining entropy value comparison analysis results of the present invention;

[0084] Figure 4 A flow chart of the steps for obtaining periodic identification results of the present invention;

[0085] Figure 5 A flowchart of the steps for obtaining abnormal mode judgment results of the present invention;

[0086] Figure 6 The figure is a flow chart of the steps for obtaining the health comparison analysis results of the present invention. DETAILED DESCRIPTION

[0087] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0088] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0089] Embodiment 1

[0090] See also Figure 1 , a remote monitoring system for obstetrics and gynecology patients comprises:

[0091] The fetal heart rate data acquisition module acquires fetal heart rate signals, integrates signal timestamps and signal values, establishes a fetal heart rate time series data set, adjusts the dynamic scale factor of the fetal heart rate time series data set, and performs time series analysis to obtain time series analysis results;

[0092] The pregnancy data analysis module performs multi-scale entropy calculation based on the time series analysis results to obtain the entropy value of each scale, analyzes the fetal heart rate variability through entropy value comparison, obtains the entropy value comparison analysis results, and performs periodic pattern recognition based on the entropy value comparison results to generate periodic recognition results;

[0093] The abnormal pattern detection module determines the abnormal pattern in the fetal heart rate data based on the periodic identification results, generates abnormal pattern judgment results, compares the abnormal pattern judgment results with historical health data, and generates health comparison analysis results.

[0094] The results of time series analysis include scale factor, anomaly index, and trend index. The results of entropy comparison analysis include scale entropy, variation index, and stability index. The results of periodic identification include waveform recognition, coefficient of variation, and stability score. The results of abnormal pattern judgment include pattern classification, signal characteristics, and continuous period. The results of health comparison analysis include trend comparison, matching index, and risk rating.

[0095] See also Figure 2 , the specific steps for obtaining the time series analysis results are:

[0096] Collect fetal heart rate signals, integrate signal timestamps and signal values, and use the formula:

[0097] S(t i )=(t i ,v i )

[0098] Calculate the signal data pair at each time point to generate a signal data pair set;

[0099] Among them, t i Represents the timestamp, v i Represents the fetal heart rate signal value, S(t i ) represents a set of signal data pairs;

[0100] Convert the signal data pair set into a fetal heart rate time series using the formula:

[0101]

[0102] Calculate the standardized fetal heart rate time series and generate the standardized fetal heart rate time series TS std ;

[0103] Among them, γ is the normalization factor, TS is the fetal heart rate time series;

[0104] Adjust the dynamic scaling factor of the normalized FHR time series using the formula:

[0105]

[0106] Calculate and generate the adjusted fetal heart rate time series TS adj ;

[0107] Among them, β is the dynamic adjustment coefficient, max and min are TS std The maximum and minimum values ​​in ;

[0108] Based on the adjusted FHR time series, time series analysis was performed using the formula:

[0109]

[0110] Calculate the time series analysis results and generate the time series analysis results R final ;

[0111] Among them, μ and σ are TS adj The mean and standard deviation of .

[0112] Formula for step 1:

[0113] S(t i )=(t i ,v i )

[0114] This formula is constructed at time point t i and the corresponding fetal heart rate v i In actual operation, t i represents the specific time point of collection, and v iIt is the fetal heart rate value measured at this time point.

[0115] Assume that the following fetal heart rate data is obtained from a monitoring device:

[0116] Time point t1 = 1s, fetal heart rate v1 = 120bpm

[0117] Time point t2 = 2s, fetal heart rate v2 = 122bpm

[0118] Use the formula:

[0119] S(1s)=(1s,120)

[0120] S(2s)=(2s,122)

[0121] The result is a set of pairs of each time point and the corresponding fetal heart rate, which is the basic data for time series analysis.

[0122] Formula for step 2:

[0123]

[0124] This formula is used to convert the signal data pair obtained in step 1 into a standardized time series. The γ in the formula is a standardization factor used to control the range of values ​​in the series to make it suitable for further analysis.

[0125] Assumptions:

[0126] γ=10

[0127] n = 2 (i.e. there are two data points)

[0128] The calculation process is:

[0129]

[0130] This is presented in a simplified form; in reality, time and value should be treated separately.

[0131] Formula for step 3:

[0132]

[0133] This formula is used to adjust the scale of the standardized time series to make it more suitable for further time series analysis. β is the dynamic adjustment coefficient.

[0134] Assumptions:

[0135] max(TS std )=12.2

[0136] min(TS std )=12.2

[0137] β=2

[0138] Calculation process:

[0139]

[0140] This result suggests that it is necessary to readjust β or TS std The calculation method of , because the maximum and minimum values ​​are the same, the adjustment result is zero.

[0141] Formula for step 4:

[0142]

[0143] This formula is the final analysis of the time series, and the R final Indicates the degree of discreteness of the time series.

[0144] Assumptions:

[0145] TS adj =[0,0]

[0146] μ=0

[0147] σ=1

[0148] Calculation process:

[0149]

[0150] This result R final = 0 means there is no discreteness in the series, which is usually because the data are not correctly scaled.

[0151] Network search and parameter setting

[0152] The parameters such as γ and β should be set based on the actual range of the fetal heart rate data. The fetal heart rate is usually between 120 and 160 bpm, so these parameters should be set so that the sequence data is properly normalized and adjusted before analysis.

[0153] See also Figure 3 , the specific steps for obtaining the entropy value comparison analysis results are:

[0154] Based on the time series analysis results, multi-scale entropy calculation is performed using the entropy calculation formula:

[0155]

[0156] Generate a set of entropy values ​​for each scale;

[0157] Among them, E(s) represents the entropy value of scale s, p i(s) represents the probability of the state under scale s, λ is the regularization parameter, and k is the total number of states;

[0158] Calculate the fluctuation range of the entropy value set and analyze the fetal heart rate variability using the formula:

[0159]

[0160] Generate fetal heart rate variability measures;

[0161] Where V represents the measure of fetal heart rate variability, represents the average value of the entropy value set, and m represents the number of entropy values;

[0162] Based on the fetal heart rate variability measure, the abnormal changes in the entropy value set are compared and analyzed using the formula:

[0163]

[0164] Generate entropy value comparison analysis result D;

[0165] Among them, max(E(s)) and min(E(s)) represent the maximum and minimum values ​​in E(s).

[0166] Detailed explanation and derivation of the formula for step 1

[0167] formula:

[0168]

[0169] Parameter details:

[0170] p i (s): The probability of the i-th state at scale s. Assume there are three states with probabilities of 0.1, 0.2, and 0.7 respectively.

[0171] λ: Regularization parameter, used to increase the stability of the calculation, usually takes a value less than 1, and is set to 0.5 here.

[0172] k: The total number of states, here assumed to be 3.

[0173] Calculation process:

[0174] Calculate the information entropy of each state probability:

[0175] -0.1log(0.1)-0.2log(0.2)-0.7log(0.7)

[0176] -0.1×(-1)-0.2×(-0.7)-0.7×(-0.36)

[0177] 0.1+0.14+0.252=0.492

[0178] Compute the regularization term:

[0179] 0.5×(0.1 2 +0.2 2 +0.7 2 )

[0180] 0.5×(0.01+0.04+0.49)

[0181] 0.5×0.54=0.27

[0182] Final result: 0.492+0.27=0.762

[0183] This result of 0.762 represents the information entropy value calculated under the influence of regularization, which reflects the combination of uncertainty and stability of a given probability distribution.

[0184] Detailed explanation and derivation of the formula for step 2

[0185] formula:

[0186]

[0187] Parameter details:

[0188] E(s): Entropy value of each scale. Assuming that the entropy values ​​corresponding to scales 1, 2, and 3 are 0.8, 0.762, and 0.7 respectively.

[0189] m: The number of entropy values, here 3.

[0190] Calculation process:

[0191] Calculate the average entropy value μ Es

[0192]

[0193] Calculate the variance:

[0194]

[0195] Final result:

[0196] This result of 0.0412 indicates the degree of fluctuation of the entropy value set, which explains the variability of fetal heart rate at different scales.

[0197] Detailed explanation and derivation of the formula for step 3

[0198] formula:

[0199]

[0200] Parameter details:

[0201] max(E s ) and min(E s ): Maximum and minimum entropy values, here are 0.8 and 0.7 respectively.

[0202] V: Result of step 2: 0.0412.

[0203] Calculation process:

[0204] Calculate the difference between the maximum entropy and the minimum entropy:

[0205] 0.8-0.7=0.1

[0206] Calculation and comparison results:

[0207]

[0208] This result of 2.427 represents the ratio of the maximum variation of the fetal heart rate entropy value to its fluctuation degree, indicating the significance of the fetal heart rate variation.

[0209] See also Figure 4 , the specific steps for obtaining the periodic identification results are:

[0210] Based on the entropy comparison analysis results, the threshold of cycle identification is determined using the standard deviation and the average value, using the formula:

[0211]

[0212] Generate threshold result θ standard ;

[0213] Among them, μ D , σ D are the mean and standard deviation of D, α, β, and γ are adjustment parameters, and D min is the minimum value in D;

[0214] Apply the threshold result to filter out the scales with entropy values ​​above the threshold, using the formula:

[0215]

[0216] Identify and select entropy scales above the threshold to generate a high entropy scale set H high ;

[0217] Among them, κ, λ, and ξ are used to adjust the weight coefficients of threshold calculation;

[0218] Based on the high entropy scale set, statistical analysis is performed to identify periodic patterns using the formula:

[0219] P=mode({s∣s∈H and E(s)>μ E +δ·σ E})

[0220] The calculation scale is used as the dominant periodic mode to generate the periodic identification result P;

[0221] Among them, μ E , σ E are the mean and standard deviation of the entropy value, and δ is a weight coefficient used to increase the calculation accuracy.

[0222] Threshold calculation formula

[0223] formula:

[0224]

[0225] Parameter explanation:

[0226] μ D : The average value of the dataset D.

[0227] σ D : The standard deviation of the data set D.

[0228] α, β, γ: fixed adjustment coefficients, usually set according to data characteristics.

[0229] D min : The minimum value in the data set D.

[0230] Assume data:

[0231] μ D =100

[0232] σ D =15

[0233] α=5

[0234] β=2

[0235] γ=1

[0236] D min =80

[0237] Calculation process:

[0238] Calculate 1.96·σ D =1.96·15=29.4

[0239] calculate Calculate θ = 100 + 29.4 + 0.061 ≈ 129.461

[0240] Result interpretation:

[0241] θ≈129.461 is the set screening threshold, which is used to select abnormally high values ​​in the data set D.

[0242] Filter scales above threshold

[0243] formula:

[0244]

[0245] Parameter explanation:

[0246] E(s): quantitative index for scale s.

[0247] θ standard =θ: The threshold calculated in the previous step.

[0248] κ, λ, ξ: Here they are assumed to be specific constant values.

[0249] Assume data:

[0250] E(s)=150

[0251] κ=3

[0252] λ=2

[0253] ξ=0.05

[0254] Calculation process:

[0255] calculate

[0256] calculate

[0257] Result interpretation:

[0258] If E(s)>40.908, s is added to the set H. In this example, 150 is significantly greater than 40.908, so s is added to H.

[0259] Statistical analysis identifies cyclical patterns

[0260] formula:

[0261] P=mode({s∣s∈H and E(s)>μ E +δ·σ E})

[0262] Parameter explanation:

[0263] μ E , σ E : The mean and standard deviation of E.

[0264] δ: An adjustment factor used to enhance calculation accuracy.

[0265] Assumptions:

[0266] μ E =40

[0267] σ E =10

[0268] δ=2

[0269] Calculation process:

[0270] Calculate μ E +δ·σ E =40+2·10=60

[0271] Result interpretation:

[0272] The E(s) of s needs to be greater than 60 to be considered the dominant scale of a periodic pattern. In this example, E(s) = 150 is clearly greater than 60, so s is the dominant periodic pattern.

[0273] See also Figure 5 , the specific steps for obtaining the abnormal mode judgment result are:

[0274] Based on the periodicity identification results, the fetal heart rate data was collected and the abnormal pattern recognition algorithm was applied, using the formula:

[0275]

[0276] Calculate the anomaly index of each data point and generate anomaly index array M;

[0277] Among them, p i represents the weight in the periodic identification result, x i represents the fetal heart rate data point, w represents the adjustment weight, and n is the total number of data points;

[0278] Using the anomaly index array, determine the anomaly threshold θ using the formula:

[0279] E={x||M(x)-μ|>θ}

[0280] Filter the data points whose abnormal index exceeds the threshold and generate an abnormal data point set;

[0281] Among them, E represents the set of abnormal data points, θ represents the abnormal threshold, and μ represents the average value of the abnormal index array M;

[0282] Using the abnormal data point set, the formula is:

[0283] R = classify(E)

[0284] Generate abnormal pattern judgment results, classify(E) is the classification function;

[0285] Among them, R represents the abnormal mode judgment result.

[0286] Detailed explanation and example of step 1

[0287] formula:

[0288]

[0289] Parameter Description:

[0290] x i : Fetal heart rate data point, assumed to be heart rate reading, in bpm (beats per minute).

[0291] p i : The weight in the periodic identification result represents the importance of each data point in the model.

[0292] w: Adjust weights to account for data variability.

[0293] n: Total number of data points, representing the number of fetal heart rate data points considered in the analysis.

[0294] Calculation process and examples:

[0295] Assume that there are four fetal heart rate data points: x = [120, 130, 140, 150] bpm, the corresponding periodicity identification weights are p = [0.2, 0.3, 0.4, 0.1], the weight adjustment factor w = 0.5, and the total number of data points n = 4.

[0296] Calculation process:

[0297] Apply the formula to calculate the anomaly index for each data point:

[0298]

[0299] Compute the mean of the array M of anomaly indices:

[0300]

[0301] Detailed explanation and example of step 2

[0302] formula:

[0303] E={x||M(x)-μ|>θ}

[0304] Parameter Description:

[0305] E: set of abnormal data points.

[0306] θ: anomaly threshold, assumed to be 5000, selected based on medical range and statistical data. μ: mean value of anomaly index array M.

[0307] M(x): Array of anomaly indices obtained from step 1.

[0308] Calculation process and examples:

[0309] Set the abnormal threshold θ = 5000

[0310] Filter data points whose anomaly index exceeds the threshold:

[0311] because

[0312] |M(120)-4521.61|=|2890.95-4521.61|≈1630.66<5000, 120 bpm is not abnormal.

[0313] |M(130)-4521.61|≈559.79<5000,130bpm is not abnormal.

[0314] |M(140)-4521.61|≈3330.22<5000,140bpm is not abnormal.

[0315] |M(150)-4521.61|≈2259.36<5000,150bpm is not abnormal.

[0316] Since the anomaly index of all data points does not exceed the threshold, there is no anomaly data point set E.

[0317] Detailed explanation and example of step 3

[0318] Since there is no abnormal data point set E in step 2, step 3 cannot be performed. In practical applications, if there is abnormal data, the classification function classify(E) will be used to generate the abnormal pattern judgment result R.

[0319] See also Figure 6 , the specific steps for obtaining the health comparison analysis results are:

[0320] To obtain abnormal mode judgment results and historical health data, use the formula:

[0321]

[0322] Calculate the cosine fit of the abnormal pattern judgment result and the historical health data to generate an intersection data set S;

[0323] Among them, R i The element representing the abnormal mode judgment result, H i represents the elements of historical health data, and n represents the size of the data set;

[0324] Using the intersection dataset, analysis was performed to identify health trends using the formula:

[0325]

[0326] Generate trend analysis results T;

[0327] Among them, S i is the element of the intersection dataset S, α and β represent adjustment parameters, and n is the total number of data points;

[0328] Use the trend analysis results to compare all historical data and use the formula:

[0329]

[0330] Generate health comparison analysis result C;

[0331] Among them, T i is the element of trend analysis result, H i Elements of historical health data.

[0332] Step 1: Intersection dataset generation

[0333] Assume the following data:

[0334] Abnormal mode judgment result R = [1,0,1,0,1]

[0335] Historical health data H = [1,1,0,0,1]

[0336] formula:

[0337]

[0338] Parameter explanation:

[0339] R i It is an element of the abnormal mode judgment result.

[0340] H i It is an element of historical health data.

[0341] n is the size of the dataset.

[0342] Calculation process:

[0343] calculate

[0344] calculate

[0345] calculate

[0346] Substituting into the formula we get

[0347] Result interpretation:

[0348] This result It represents the cosine fit between the abnormal pattern judgment result and the historical health data. A value close to 1 indicates a high fit.

[0349] Step 2: Generate trend analysis results

[0350] Assume that the intersection dataset obtained from step 1 is

[0351] formula:

[0352]

[0353] Parameter settings:

[0354] α=0.5

[0355] β=1.5

[0356] n=5 (assuming dataset size)

[0357] Calculation process:

[0358] Treat S as each S i Value

[0359] calculate

[0360] calculate

[0361] Substituting into the formula we get

[0362] Result interpretation:

[0363] This result It represents the result of trend analysis after weighted processing. A value higher than 1 indicates that a significant trend or pattern has been identified.

[0364] Step 3: Generate health comparison analysis results

[0365] Assuming historical data H = [1,1,0,0,1], trend analysis results

[0366] formula:

[0367]

[0368] Calculation process:

[0369] Assumptions For all i

[0370] calculate

[0371]

[0372] Sum

[0373] Substituting into the formula we get

[0374] Result interpretation:

[0375] This result Represents the results of health comparison analysis, indicating the mean absolute error between trend analysis results and historical data. The smaller the value, the closer the analysis results are to historical data.

[0376] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A remote monitoring system for obstetrics and gynecology patients, characterized in that: The system comprises: The fetal heart rate data acquisition module acquires fetal heart rate signals, integrates signal timestamps and signal values, establishes a fetal heart rate time series data set, adjusts a dynamic scale factor of the fetal heart rate time series data set, and performs time series analysis to obtain time series analysis results; The steps for obtaining the time series analysis results are specifically as follows: Collect the fetal heart rate signal, integrate the signal timestamp and signal value, and use the formula: S(t i )=(t i ,v i ) Calculate the signal data pair at each time point to generate a signal data pair set; Among them, t i Represents the timestamp, v i Represents the fetal heart rate signal value, S(t i ) represents a set of signal data pairs; The signal data pair set is converted into a fetal heart rate time series using the formula: Calculate the standardized fetal heart rate time series and generate the standardized fetal heart rate time series TS std ; Among them, γ is the normalization factor, TS is the fetal heart rate time series; The dynamic scaling factor of the standardized fetal heart rate time series is adjusted using the formula: Calculate and generate the adjusted fetal heart rate time series TS adj ; Among them, β is the dynamic adjustment coefficient, max and min are TS std The maximum and minimum values ​​in ; Based on the adjusted fetal heart rate time series, a time series analysis is performed using the formula: Calculate the time series analysis results and generate the time series analysis results R final ; Among them, μ and σ are TS adj The mean and standard deviation of The pregnancy data analysis module performs multi-scale entropy calculation based on the time series analysis results to obtain the entropy value of each scale, analyzes the fetal heart rate variability through entropy value comparison, obtains the entropy value comparison analysis results, performs periodic pattern recognition based on the entropy value comparison analysis results, and generates a periodicity recognition result; The steps for obtaining the entropy value comparison analysis result are specifically as follows: Based on the time series analysis results, multi-scale entropy calculation is performed using the entropy calculation formula: Generate a set of entropy values ​​for each scale; Among them, E(s) represents the entropy value of scale s, p i (s) represents the probability of the state under scale s, λ is the regularization parameter, and k is the total number of states; The fluctuation range of the entropy value set is calculated to analyze the fetal heart rate variability using the formula: Generate fetal heart rate variability measures; Where V represents the measure of fetal heart rate variability, represents the average value of the entropy value set, and m represents the number of entropy values; Based on the fetal heart rate variability metric, the abnormal changes in the entropy value set are compared and analyzed using the formula: Generate entropy value comparison analysis result D; Among them, max(E(s)) and min(E(s)) represent the maximum and minimum values ​​in E(s); The abnormal pattern detection module determines the abnormal pattern in the fetal heart rate data based on the periodic identification result, generates an abnormal pattern determination result, compares the abnormal pattern determination result with the historical health data, and generates a health comparison analysis result; The steps for obtaining the abnormal mode judgment result are specifically as follows: Based on the periodicity identification results, fetal heart rate data is collected and an abnormal pattern recognition algorithm is applied, using the formula: Calculate the anomaly index of each data point and generate anomaly index array M; Among them, p i represents the weight in the periodic identification result, x i represents the fetal heart rate data point, w represents the adjustment weight, and n is the total number of data points; Using the anomaly index array, determine the anomaly threshold θ using the formula: E={x||M(x)-μ|>θ} Filter the data points whose abnormal index exceeds the threshold and generate an abnormal data point set; Among them, E represents the set of abnormal data points, θ represents the abnormal threshold, and μ represents the average value of the abnormal index array M; Using the set of abnormal data points, the formula is adopted: R=classify (E) Generate abnormal mode judgment results; Among them, R represents the abnormal pattern judgment result, and classify (E) is the classification function.

2. The remote monitoring system for obstetrics and gynecology patients according to claim 1, characterized in that: The steps for obtaining the periodic identification result are specifically as follows: Based on the entropy comparison and analysis results, the threshold of period identification is determined using the standard deviation and the average value, using the formula: Generate threshold result θ standard ; Among them, μ D , σ D are the mean and standard deviation of D, α, β, and γ are adjustment parameters, and D min is the minimum value in D; Apply the threshold result to filter out the scales in the entropy value set that are higher than the threshold, using the formula: Identify and select entropy scales above the threshold to generate a high entropy scale set H high ; Among them, κ, λ, ξ are used to adjust the weight coefficients of threshold calculation; Based on the high entropy scale set, statistical analysis is performed to identify periodic patterns using the formula: P=mode({s|s∈H and E(s)>μ E +d·s E }) The calculation scale is used as the dominant periodic mode to generate the periodic identification result P; Among them, μ E , σ E are the mean and standard deviation of the entropy value, and δ is a weight coefficient used to increase the calculation accuracy.

3. The remote monitoring system for obstetrics and gynecology patients according to claim 1, characterized in that: The steps for obtaining the health comparison analysis results are specifically as follows: To obtain the abnormal mode judgment result and historical health data, the formula is used: Calculate the cosine fit of the abnormal pattern judgment result and the historical health data to generate an intersection data set S; Among them, R i The element representing the abnormal mode judgment result, H i represents the elements of historical health data, and n represents the size of the data set; Using the intersection dataset, analysis was performed to identify health trends using the formula: Generate trend analysis results T; Among them, S i is the element of the intersection dataset S, α and β represent adjustment parameters, and n is the total number of data points; Using the trend analysis results, compared with all historical data, the formula is: Generate health comparison analysis result C; Among them, T i is the element of trend analysis result, H i Elements of historical health data.

Images