Method for extracting uterine contraction information based on myoelectrical signals of uterus

Abstract

A method for extracting uterine contraction information based on uterine electromyogram signals, comprising: Step S1, collecting uterine electromyogram signals of a certain duration through electrodes on the abdomen of a pregnant woman and converting them into digital signals; Step S2, preprocessing the uterine electromyogram signals; Step S3, dividing the uterine electromyogram signals into a series of overlapping windows in a sliding window manner; Step S4, calculating the sample entropy SampEn and zero crossing rate ZC of each window; Step S5, normalizing the sample entropy and zero crossing rate values, and respectively performing weighted summation on the sample entropy and zero crossing rate of each window to obtain the uterine contraction intensity F corresponding to each window at each moment; Step S6, setting weight factors ω s , ω e to detect the start and end positions of uterine contractions existing in the signal, and calculate the occurrence frequency f c and average duration d c . The present invention can obtain more reliable uterine contraction information by means of non-invasively collected uterine electromyogram signals.

Description

Technical Field

[0001] The present invention belongs to the technical field of biological signal processing, and particularly relates to a method for extracting uterine contraction information based on electrohysterogram signals. Background Art

[0002] In order to promote the long-term balanced development of the population, in 2021, the country implemented the three-child fertility policy, established a positive fertility support policy system, improved the level of prenatal and postnatal care services, and at the same time ensured the health of pregnant women and children. Currently, preterm birth (giving birth prematurely, less than 37 weeks of pregnancy) is the most major pregnancy risk. Premature infants may have serious physical defects, which not only affect their life health and quality of life, but also affect the happiness and harmony of the family.

[0003] The direct cause of preterm birth is the strong contraction of the uterus. To avoid preterm birth, pregnant women in the second and third trimesters of pregnancy will go to the obstetrics department for medical treatment due to uterine contractions. Obstetricians usually monitor the intensity of uterine contractions to obtain information about whether the uterine pressure is high enough to expel the fetus or whether medical intervention is recommended. The commonly used detection device in clinical practice is the tocodynamometer (TOCO), which indirectly analyzes the presence of uterine contractions by measuring the pressure exerted by uterine contractions on the abdominal wall. However, TOCO is affected by factors such as the tightness of the strap and the thickness of the pregnant woman's fat, resulting in the easy omission of some uterine contractions and being not suitable for long-term detection.

[0004] The relationship between uterine contractions and the generation and propagation of action potentials in uterine muscle cells provides a new idea for uterine contraction recognition. These potential electrical activities can be measured by electrodes placed on the patient's abdomen, thereby obtaining electrohysterogram (EHG) signals. Physiological experiments show that the burst activity of electrical signals is closely related to the intensity, duration, and frequency of uterine contractions. However, most studies on uterine contraction detection based on electrohysterogram signals focus on the amplitude characteristics of the signals, such as calculating the root mean square of the signal to characterize the intensity change of uterine contractions. However, the signal-to-noise ratio of the actually collected EHG signals is usually relatively low, and it is not uncommon to observe a series of false spikes. The source of these spike-like noises is uncertain, such as slight displacement of the electrodes or interference from radio transmission. Due to the existence of false spikes in the collected signals, the root-mean-square-based method is prone to false detection. To improve the existing method, considering the occurrence mechanism of uterine contractions: when the uterus contracts, the intensity of the burst action potential continuously increases and the interval becomes smaller (the frequency increases); on the other hand, the complexity of the myoelectric signals generated when the uterus relaxes and contracts is also different. Therefore, designing a uterine contraction detection method by combining the characteristics of uterine contractions helps to improve the detection effect and further improve the performance of preterm birth diagnosis. Summary of the Invention

[0005] In order to overcome the shortcomings of the existing uterine electromyogram signal technology, the present invention proposes a method for extracting uterine contraction information by combining sample entropy and zero-crossing rate, and obtains relevant information such as the position and duration of uterine contractions.

[0006] The method for extracting uterine contraction information based on uterine electromyogram signals of the present invention realizes the following specific steps to solve its technical problems:

[0007] Step S1: Collect uterine electromyogram signals of a certain duration through electrodes on the abdomen of a pregnant woman and convert them into digital signals;

[0008] Step S2: Preprocess the uterine electromyogram signals;

[0009] Step S3: Divide the uterine electromyogram signals into a series of overlapping windows in a sliding window manner;

[0010] Step S4: Calculate the sample entropy SampEn and zero-crossing rate ZC of each window;

[0011] Step S5: Normalize the sample entropy and zero-crossing rate values, and perform weighted summation on the sample entropy and zero-crossing rate of each window respectively to obtain the uterine contraction intensity F at the corresponding moment of each window;

[0012] Step S6: Set the weight factors ω s , ω e to detect the start and end positions of uterine contractions existing in the signal, and calculate the occurrence frequency f c and the average duration d c .

[0013] As a further technical solution of the present invention, step S2 includes the following steps:

[0014] S21: Subtract the measurement signal of one electrode from the measurement signal of another electrode to form a bipolar signal to eliminate shared electrical activities;

[0015] S22: Filter the signal using a band-pass filter;

[0016] S23: Downsample the signal to reduce the computational amount.

[0017] Further, step S3 specifically includes: Since the duration of uterine contractions is about 60 seconds, the sliding window length is set to 60 seconds, and the EHG signal is divided into a series of window segments at intervals of 1 second.

[0018] Further, step S4 specifically includes: For the uterine electromyogram window signal (x1, x2,..., x N ) of length N, calculate the zero-crossing rate and sample entropy; the calculation method of the zero-crossing rate is as follows:

[0019]

[0020] The calculation method of sample entropy is as follows:

[0021]

[0022] Where the integer m represents the length of the comparison vector, the real number r represents the measure value of similarity, and B m (r) represents the number of vector pattern matches when the vector length is m and the similarity is r.

[0023] Furthermore, step S5 specifically includes: for each signal window, calculate their corresponding uterine contraction intensity:

[0024] F = λ·ZC + ω·SampEn (3)

[0025] Where λ and ω are the weight coefficients of the zero-crossing rate and sample entropy respectively, and λ + ω = 1.

[0026] Furthermore, both λ and ω are set to 0.5.

[0027] Furthermore, step S6 specifically includes: through the weight factor ω s , ω e to obtain the uterine contraction start point threshold T s = ω s × max(F) and the end point threshold T e = ω e × max(F);

[0028] For the electromyogram signal of the uterus with a recording duration of t s , the detected number of uterine contractions N c , calculate the parameter characterizing the uterine contraction occurrence frequency:

[0029]

[0030] Assume that the duration of each detected uterine contraction is respectively , calculate the average duration of uterine contractions:

[0031]

[0032] Furthermore, both ω s , ω e are set to 0.2.

[0033] The advantages of the present invention are as follows: By means of the non-invasive acquisition of uterine myoelectric signals, relevant information on uterine contractions is obtained. Different from the common calculation of the root mean square, the zero-crossing rate and sample entropy are less affected by spike noise. The increase in the frequency of action potentials in the uterine myoelectric signal is quantified by the zero-crossing rate, while the change in signal complexity caused by uterine contractions can be represented by the sample entropy, so as to obtain more reliable uterine contraction information and provide a reference for doctors to analyze whether a pregnant woman has premature labor. Brief Description of the Drawings

[0034] Figure 1 is a flowchart of the method of the present invention.

[0035] Figure 2 is an example of calculating the uterine contraction intensity of the present invention. The left side is an example of a normal signal, and the right side is an example after adding spike noise. Detailed Description of the Invention

[0036] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs. The terms used in the specification of this application herein are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0038] Next, the technical solutions of the present invention will be described in conjunction with the accompanying drawings.

[0039] The method for extracting uterine contraction information based on uterine myoelectric signals of the present invention realizes the following specific steps to solve its technical problems:

[0040] Step S1: Collect uterine myoelectric signals of a certain duration through electrodes on the abdomen of a pregnant woman and convert them into digital signals;

[0041] Step S2: Preprocess the uterine myoelectric signals;

[0042] Step S3: Divide the uterine myoelectric signals into a series of overlapping windows in a sliding window manner;

[0043] Step S4: Calculate the sample entropy SampEn and zero-crossing rate ZC of each window;

[0044] Step S5: Normalize the sample entropy sum and the zero-crossing rate value, and perform weighted summation on the sample entropy sum and the zero-crossing rate of each window respectively to obtain the uterine contraction intensity F corresponding to each window at each moment;

[0045] Step S6: Set the weight factors ω s , ω e for detecting the start and end of uterine contractions, and calculate the occurrence frequency f c and the average duration d c ;

[0046] The said step S2 includes the following steps:

[0047] S21: Subtract the measurement signal of one electrode from that of another electrode to form a bipolar signal so as to eliminate the shared electrical activity;

[0048] S22: Since the main frequency range of the EHG signal is from 0.1 to 3 Hz, a band-pass filter with a frequency band range of 0.1 to 3 Hz is used to filter the signal;

[0049] S23: Downsample the data to reduce the computational amount;

[0050] The above step S3 specifically includes: Since the duration of uterine contractions is about 60 seconds, the sliding window length is set to 60 seconds, and the EHG signal is divided into a series of window segments at intervals of 1 second as the step size.

[0051] The above step S4 specifically includes: For the myoelectrical window signal (x1, x2,..., x N ) of the uterus with a length of N, calculate the zero-crossing rate and the sample entropy; the calculation method of the zero-crossing rate is as follows:

[0052]

[0053] The calculation method of the sample entropy is as follows:

[0054]

[0055] where the integer m represents the length of the comparison vector, the real number r represents the measure value of similarity, and B m (r) represents the number of vector pattern matches in the case of a vector length of m and a similarity of r.

[0056] The above step S5 specifically includes: For each signal window, calculate their corresponding uterine contraction intensity:

[0057] F = λ·ZC + ω·SampEn (8)

[0058] Where λ and ω are the weight coefficients of the zero-crossing rate and sample entropy respectively, and λ + ω = 1. In this example, both λ and ω are set to 0.5.

[0059] The above step S6 specifically includes: through the weight factor ω s , ω e to obtain the contraction start point threshold T s = ω s × max(F) and the end point threshold T e = ω e × max(F), ω s , ω e are both set to 0.2 in this example.

[0060] For the electromyogram signal of the uterus with a recording duration of t s , the number of contractions N c detected is calculated to represent the contraction occurrence frequency parameter:

[0061]

[0062] Assume that the duration of each detected contraction is respectively Calculate the average contraction duration:

[0063]

[0064] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0065] The above-described embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be understood as a limitation to the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several deformations and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method for extracting uterine contraction information based on uterine electromyogram signals, comprising the following steps: Step S1: Collect uterine electromyogram signals of a certain duration through electrodes on the abdomen of a pregnant woman and convert them into digital signals; Step S2: Preprocess the uterine electromyogram signals; Step S3: Divide the uterine electromyogram signals into a series of overlapping windows in a sliding window manner; Step S4: Calculate the sample entropy SampEn and zero crossing rate ZC of each window; specifically including: for the uterine EMG window signals x1, x2,..., x of length N N , calculate the zero crossing rate and sample entropy; the calculation method of the zero crossing rate is as follows: The calculation method of sample entropy is as follows: where the integer m represents the length of the comparison vector, the real number r represents the measure of similarity, and B m (r) represents the number of vector pattern matches when the vector length is m and the similarity is r; Step S5: Normalize the sample entropy and zero-crossing rate values, and perform weighted summation on the sample entropy and zero-crossing rate of each window respectively to obtain the uterine contraction intensity F at the corresponding moment of each window; specifically including: for each signal window, calculate their corresponding uterine contraction intensities: F = λ·ZC + ω·SampEn (3) where λ and ω are the weight coefficients of the zero-crossing rate and sample entropy respectively, and λ + ω = 1; Step S6: Set the weight factors ω for the start and end of uterine contractions s , ω e to detect the start and end of uterine contractions, and calculate the occurrence frequency f c and the average duration d c ; specifically including: obtaining the uterine contraction start point threshold T s , ω e to get the uterine contraction start point threshold T s = ω s × max(F) and the end point threshold T e = ω e × max(F); For the myometrial electrical signals with a recording duration of t s the number of detected uterine contractions is N c , calculate the parameter characterizing the frequency of uterine contractions: Assume that the duration of each detected uterine contraction is respectively Calculate the average duration of uterine contractions:

2. The method for extracting uterine contraction information based on uterine myoelectric signals according to claim 1, wherein: Step S2 includes the following steps: S21: Subtract the measurement signal of one electrode from that of another electrode to form a bipolar signal to eliminate shared electrical activities; S22: Since the main frequency range of the EHG signal is between 0.1 and 3 Hz, a band-pass filter with a frequency band range of 0.1 to 3 Hz is used to filter the signal; S23: Downsample the data to reduce the computational amount.

3. The method for extracting uterine contraction information based on uterine electromyogram signals according to claim 1, wherein: Step S3 specifically includes: Since the duration of uterine contractions is about 60 seconds, the sliding window length is set to 60 seconds, and the EHG signal is divided into a series of window segments at an interval of 1 second as the step size.

4. The method for extracting uterine contraction information based on myoelectrical signals of uterus according to claim 1, wherein: Both λ and ω are set to 0.

5.

5. The method for extracting uterine contraction information based on uterine myoelectric signals according to claim 1, characterized in that: The described ω s , ω e are both set to 0.2.

Images