CN-121987241-A - Fetal heart monitoring method and system based on morphological feature visualization

CN121987241ACN 121987241 ACN121987241 ACN 121987241ACN-121987241-A

Abstract

The invention discloses a fetal heart monitoring method and a fetal heart monitoring system based on morphological feature visualization, comprising the steps of extracting a fetal heart rate baseline and an acceleration and deceleration section of a fetal heart rate signal; extracting the contraction baseline and contraction state of the contraction signal, and realizing fetal heart monitoring according to the acceleration and deceleration section of the fetal heart rate signal and the contraction state of the contraction signal. The iterative filtering clipping baseline extraction algorithm adopted by the invention can effectively resist the peak, the loss and the interference of high-frequency noise in the signal through a multi-round threshold convergence strategy.

Inventors

DENG YANJUN
ZHANG YEFEI
WANG HAO
ZHAO ZHIDONG
Zhou Houying
FAN YIMING
LIU PENG
Zhao Lanpeng
CHEN LIXIA

Assignees

杭州电子科技大学平湖数字技术创新研究院有限公司
杭州电子科技大学

Dates

Publication Date: 20260508
Application Date: 20251226

Claims (10)

1. A fetal heart monitoring method based on morphological feature visualization is characterized by comprising the steps of extracting a fetal heart rate baseline and an acceleration and deceleration section of a fetal heart rate signal, extracting a contraction baseline and a contraction state of a contraction signal, and realizing fetal heart monitoring according to the acceleration and deceleration section of the fetal heart rate signal and the contraction state of the contraction signal; The specific implementation process of extracting the fetal heart rate baseline of the fetal heart rate signal and the acceleration and deceleration section comprises the following steps: step S1, acquiring a fetal heart rate signal, and extracting a fetal heart rate baseline by adopting a multi-round iterative filtering pruning strategy; step S2, identifying an acceleration section in a fetal heart rate curve based on fetal heart rate acceleration; Step S3, identifying a deceleration section in the fetal heart rate curve based on the fetal heart rate deceleration; the specific implementation process for extracting the contraction baseline and the contraction state of the contraction signal comprises the following steps: s4, acquiring a uterine contraction signal, and acquiring a uterine contraction baseline by adopting a multi-round iterative filtering pruning strategy; and S5, identifying the uterine contraction state according to the uterine contraction base line.
2. The method according to claim 1, wherein step S1 is specifically: S101, converting a fetal heart rate signal into a heartbeat peak interval signal, and then counting data distribution peak points within 300-666 ms of the heartbeat peak interval signal ; Step S102, backward filtering pretreatment: marking a baseline initial value of the heartbeat peak value interval signal as For the heart beat peak value interval signal From the slave Reverse traversal to 1, N represents the length of the heartbeat peak interval signal, if the ith heartbeat peak interval signal Satisfy its and data distribution peak point If the initial value is smaller than or equal to the first threshold value, updating the initial value of the base line, otherwise, not updating; step S103, forward filtering, namely performing backward filtering on the heartbeat peak value interval signal after preprocessing From 1 to Forward traversal if the ith heartbeat peak interval signal Satisfy its and data distribution peak point A first threshold value or less, a peak interval signal for each heartbeat Updating, otherwise updating to the previous uterine contraction signal; Step S104, backward filtering, namely performing forward filtering on the heartbeat peak value interval signal From the slave To 1 forward traversal, for a heartbeat peak interval signal Updating again; step S105, clipping strategy, namely converting the heartbeat peak value interval signal after the backward filtering in step S104 into a fetal heart rate format and taking the fetal heart rate format as a temporary base line Detecting the original fetal heart rate signal If there is an ith fetal heart rate signal within the segment U represents the upper threshold of the base line, and updates the fetal heart rate value of the segment to be . Finally, detecting If present within the fragment L represents the baseline lower threshold, and updates the fetal heart rate value of the segment to ;; Step S106, sequentially selecting the gradually reduced U and L, and repeating the steps S101 to S105 until the ideal U and the ideal L are reached, thereby obtaining the accurate fetal heart rate baseline BL.
3. The method according to claim 2, wherein in step S102, the update formula of the baseline initial value is , Are both weights, the sum of which is 1 and both of which are greater than 0.
4. The method of claim 2, wherein in step S103, the update formula of each heartbeat peak value interval signal is , Are both weights, the sum of which is 1 and both of which are greater than 0.
5. The method of claim 2, wherein in step S104, the re-update formula of each heartbeat peak interval signal is , Are both weights, the sum of which is 1 and both of which are greater than 0.
6. The method according to claim 1, wherein step S2 is specifically: step S201, searching for a fetal heart rate signal point X with a difference value between an original fetal heart rate signal and an accurate fetal heart rate baseline BL being larger than a threshold value, searching for a maximum peak value within a time t1 after the point, and recording the time of the maximum peak value ; Step S202, looking up ratio of original fetal heart rate signal in t2 time period from maximum peak time A large minimum fetal heart rate signal point is marked as an acceleration starting point; representing a fetal heart rate acceleration segment lifting threshold; Step S203, looking up the ratio of the original fetal heart rate signals in the period from the maximum peak time to the back t3 A large minimum fetal heart rate signal point is marked as an acceleration end point; And step S204, if the time difference between the acceleration end point and the acceleration start point exceeds a threshold value, recording the time as one effective fetal heart rate acceleration, otherwise, searching the next fetal heart rate signal point X in the step S201 again, and repeating the steps S201-S204.
7. The method according to claim 1, wherein step S3 is specifically: Step S301, searching for a fetal heart rate signal point Y with a difference value between the original fetal heart rate signal and the accurate fetal heart rate baseline BL lower than a threshold value, searching for a minimum valley value within a time t1 after the point, and recording the moment of the minimum valley value ; Step S302, looking up ratio of fetal heart rate signal points in a period t2 from the minimum valley time to the front A small maximum fetal heart rate signal point is marked as a deceleration starting point; representing a fetal heart rate deceleration segment reduction threshold; Step S303, searching ratio of fetal heart rate signal points in a period from the minimum valley time to the back t3 A small maximum fetal heart rate signal point is marked as a deceleration end point; and step S304, if the time difference between the deceleration end point and the deceleration start point exceeds a threshold value, recording the section as one effective fetal heart rate deceleration, otherwise, searching the next fetal heart rate signal point Y in the step S301 again, and repeating the steps S301-S304.
8. The method according to claim 1, wherein step S4 is specifically: S401, counting the numerical distribution of the uterine contraction signal UC, and searching for peak points of uterine contraction data distribution within the range of 0 to the maximum value ;; S402, preprocessing by backward filtering, and recording the length of the uterine contraction signal as Initial value of uterine contraction baseline is marked as For uterine contraction signal From the slave To 1, reverse traversal is performed if the ith uterine contraction signal point Satisfy the peak point of the distribution of the uterine contraction data If the initial value is smaller than or equal to the second threshold value, updating the initial value of the uterine contraction baseline, otherwise, not updating; The update formula of the initial value of the uterine contraction baseline is , Are both weights, the sum of the weights is 1, and the weights are both greater than 0; s403, forward filtering, namely performing backward filtering pretreatment on the uterine contraction signal From 1 to Forward traversal, updated using the following formula . If the ith uterine contraction signal point Satisfy the peak point of the distribution of the uterine contraction data If the signal is smaller than or equal to the second threshold value, the signal points are determined to be in the uterine contraction Updating, otherwise updating to the previous uterine contraction signal point; The update formula of the uterine contraction signal point is that ; Step S404, backward filtering, namely performing forward filtering pretreatment on the uterine contraction signal From the slave Going to 1 forward traversal, updating each uterine contraction signal point again; The re-update formula of the uterine contraction signal point is as follows: ; Step S405, pruning strategy, namely marking the uterine contraction signal after the backward filtering in step S404 as a temporary baseline Detecting the original uterine contraction signal If there is a segment within the segment , Indicating the upper limit threshold of the uterine contraction baseline, and updating all the uterine contraction signal points in the segment to be Then detecting If present within the fragment , Representing the lower limit threshold of the uterine contraction baseline, and updating all the uterine contraction signal points in the segment to be ; Step S406, sequentially selecting gradually decreasing And Repeating S401-S405 to obtain accurate uterine contraction baseline 。
9. The method according to claim 1, wherein step S5 is specifically: step S501, searching an original uterine contraction signal and an accurate uterine contraction baseline The difference value between the uterine contraction signal points Z is larger than a threshold value, the maximum peak value is searched within t4 minutes after the point, and the moment of recording the maximum peak value is ; Step S502, from the peak point Original uterine contraction signal search ratio within t5 minutes forward The large minimum uterine contraction signal point is marked as the starting point of uterine contraction; A lift threshold representing effective uterine contraction; Step S503, from the peak point Raw uterine contraction signal search ratio within t6 minutes backward The large minimum uterine contraction signal point is marked as the endpoint of uterine contraction; Step S504, if the time difference between the end point and the starting point exceeds the threshold value, recording the section as one effective uterine contraction, otherwise, searching one uterine contraction signal point Z in step S501 again, and repeating the steps S501-S504.
10. A fetal heart monitoring system implementing the method of any one of claims 1-9, comprising: the fetal heart rate signal processing module is used for extracting a fetal heart rate baseline and an acceleration and deceleration section of the fetal heart rate signal; the uterine contraction signal processing module is responsible for extracting a uterine contraction baseline and a uterine contraction state of a uterine contraction signal.

Description

Fetal heart monitoring method and system based on morphological feature visualization Technical Field The invention relates to the field of medical signal processing and medical monitoring, in particular to a fetal heart monitoring method and system based on morphological feature visualization, which can be used for carrying out high-precision acquisition, feature extraction, intelligent identification and visual display on fetal heart rate and uterine contraction signals and is used for assisting obstetrician in fetal intrauterine health assessment. Background Fetal heart and uterine contractility monitoring (Cardiotocography, CTG) is an important means of assessing fetal intrauterine conditions in prenatal monitoring by monitoring fetal heart rate (FETAL HEART RATE, FHR) and uterine contractility pressure (Uterine Contraction, UC) simultaneously to assess fetal health. Traditional CTG equipment mainly relies on doctors to carry out visual interpretation and manual measurement on recorded waveform drawings, and the method has obvious defects: The subjectivity is strong, different doctors can possibly draw different conclusions on the same CTG waveform, and the diagnosis consistency is poor. In long-time monitoring, the manual identification is easy to cause feature omission or misjudgment due to visual fatigue, and particularly, the identification of variant deceleration and late deceleration is easy to carry out. Some automated analysis algorithms exist in the prior art, such as using moving average, median filtering, or rule-based thresholding to estimate fetal heart rate baselines and identify characteristic events. However, these conventional methods have poor robustness when the signal is disturbed by severe fetal movements, maternal movements or power frequency noise, and the baseline estimation is inaccurate, which in turn leads to an increased error rate of subsequent feature recognition. In addition, most of the existing algorithms are only realized on a general-purpose computer through software, so that the calculation complexity is high, and the algorithms are difficult to integrate into an embedded terminal with low power consumption and real-time requirements. Most of the existing portable fetal heart monitoring devices on the market only have a simple signal display function and lack deep morphological feature analysis and auxiliary diagnosis capabilities. Therefore, an automatic morphological feature analysis and visualization scheme which can be integrated in an embedded terminal and has high accuracy and robustness is urgently needed in the field so as to solve the problems of strong subjectivity, low efficiency, error easiness and the like in the prior art. Disclosure of Invention The invention aims to overcome the defects of the prior art, and provides an embedded fetal heart monitoring system and method, which can collect fetal heart rate and uterine contraction pressure signals with high quality, automatically and accurately identify key clinical characteristics of fetal heart rate baselines, acceleration, deceleration, uterine contraction baselines and uterine contraction states by adopting an improved morphological characteristic extraction algorithm, and effectively improve objectivity, consistency and efficiency of diagnosis. In order to achieve the above purpose, the present invention adopts the following technical scheme: The invention provides a morphological feature extraction method of fetal heart and uterine contraction, which comprises the steps of extracting a fetal heart rate baseline and an acceleration and deceleration section of a fetal heart rate signal, extracting a uterine contraction baseline and a uterine contraction state of a uterine contraction signal, and realizing fetal heart monitoring according to the acceleration and deceleration section of the fetal heart rate signal and the uterine contraction state of the uterine contraction signal; The specific implementation process of extracting the fetal heart rate baseline of the fetal heart rate signal and the acceleration and deceleration section comprises the following steps: step S1, acquiring a fetal heart rate signal, and extracting a fetal heart rate baseline by adopting a multi-round iterative filtering pruning strategy; step S2, identifying an acceleration section in a fetal heart rate curve based on fetal heart rate acceleration; Step S3, identifying a deceleration section in the fetal heart rate curve based on the fetal heart rate deceleration; the specific implementation process for extracting the contraction baseline and the contraction state of the contraction signal comprises the following steps: s4, acquiring a uterine contraction signal, and acquiring a uterine contraction baseline by adopting a multi-round iterative filtering pruning strategy; and S5, identifying the uterine contraction state according to the uterine contraction base line. Preferably, step S1 is specifically: S101, converting a fet