CN-121987242-A - Fetal heart rate determining method and system

CN121987242ACN 121987242 ACN121987242 ACN 121987242ACN-121987242-A

Abstract

The invention relates to the technical field of fetal heart signal processing, and discloses a fetal heart rate determining method and a fetal heart rate determining system, wherein the fetal heart rate determining method comprises the steps that an echo sequence is subjected to denoising normalization to obtain a target sequence, a segmentation window set is generated according to a preset window length and a step length, a peak parameter table is formed by extracting peak amplitude, peak width and neighborhood contrast under the caliber, three items are jointly participated in the process of grading in a separable way, so that low-contrast or wide peak is depressed, but the peak with weak amplitude but with constraint satisfied contrast and width is lifted, background fluctuation and the target peak are separated by a distance on grading to obtain the target peak set, thus weak echo peak separability is enhanced, the target peak set is used as a node to calculate a consistent index, the candidates with sufficient grading and consistent time positions are only connected, a path is selected to obtain a periodic track, time axis jump is restrained and tends to be smooth, a periodic vector arranged according to time sequence is finally generated, adjacent item difference converges and mutation is reduced, and period estimation continuity is guaranteed.

Inventors

ZHU XIAODAN

Assignees

金华市妇幼保健院(金华市妇幼保健计划生育服务中心)

Dates

Publication Date: 20260508
Application Date: 20260126

Claims (10)

1. A method for determining a fetal heart rate, comprising: Denoising normalization processing is carried out on the echo sequence to obtain a target sequence, wherein the echo sequence is a time sequence which is acquired by a probe and is subjected to demodulation envelope; Executing segment generation processing on the target sequence based on a preset time axis to obtain a segment window set, wherein the segment window set is a time slice set established according to a preset window length and a step length; Performing peak parameter extraction processing on the target sequence according to the segmented window set to obtain a peak parameter table, and performing separable scoring processing on the basis of the peak parameter table to obtain a target peak set, wherein the peak parameter table is a column structure comprising peak amplitude, peak width and neighborhood contrast; Performing consistency computation processing on the segmentation window set according to the target peak set to obtain a consistency index, and performing path selection processing based on the consistency index to obtain a periodic track; and executing cycle generation processing based on the cycle track to obtain a cycle sequence, wherein the cycle sequence is a cycle vector arranged in time sequence.
2. A method of determining a fetal heart rate as claimed in claim 1 wherein the method of performing peak parameter extraction processing on the target sequence based on the set of segmentation windows to obtain a table of peak parameters comprises: Performing peak candidate generation processing on the target sequence according to the segmentation window set to obtain a peak candidate set, wherein the peak candidate set is a peak index set meeting local extremum constraint in each time slice; Performing amplitude calculation processing on the target sequence according to the peak candidate set to obtain a peak amplitude vector; performing width calculation processing on the target sequence according to the peak candidate set to obtain a peak width vector; performing neighborhood contrast processing on the target sequence according to the peak candidate set to obtain a contrast vector, wherein the contrast vector is a vector of a difference value between the peak top and a neighborhood baseline; and performing table assembly processing based on the peak amplitude vector, the peak width vector and the comparison vector to obtain a peak parameter table.
3. A method of determining a fetal heart rate as claimed in claim 2 wherein the method of performing a neighborhood contrast process on the target sequence based on the set of peak candidates to obtain a contrast vector comprises: performing neighborhood positioning processing according to the peak candidate set and the target sequence to obtain a neighborhood index, and performing sample extraction processing on the target sequence according to the neighborhood index to obtain a neighborhood table; Performing baseline estimation processing according to the neighborhood table to obtain a baseline vector, wherein the baseline vector is a median vector aligned with the neighborhood index; performing peak top calculation processing on the target sequence according to the peak candidate set to obtain a peak top vector, wherein the peak top vector is an amplitude vector aligned with the peak candidate set index; and performing difference value calculation processing based on the peak top vector and the base line vector to obtain a comparison vector.
4. A method of determining a fetal heart rate as claimed in claim 3 wherein the method of performing a separable scoring process based on a table of peak parameters to obtain a target peak set comprises: performing normalization calculation according to the peak parameter table to obtain a normalization table, wherein the normalization table is a column type table obtained by scaling peak amplitude vectors, peak width vectors and comparison vectors in the peak parameter table according to column ranges; performing robust estimation processing according to the normalization table to obtain a robust vector, wherein the robust vector is formed by calculating a partial difference of a corresponding column of the normalization table in each time slice; Performing continuous constraint processing based on the normalization table and the robust vector to obtain a continuous vector; Performing weight calculation processing according to the robust vector to obtain a weight vector, and performing grading generation processing based on the normalization table and the weight vector to obtain a grading vector; and executing threshold selection processing based on the scoring vector and the continuous vector to obtain a target peak set.
5. The method of claim 4, wherein the step of performing a score generation process based on the normalization table and the weight vector to obtain a score vector comprises: performing column alignment processing according to the normalization table and the weight vector to obtain an alignment table, wherein the alignment table is a table in which columns of the normalization table are in one-to-one correspondence with the weight vector according to column indexes; performing weighted synthesis processing according to the alignment table to obtain a synthesis vector, wherein the synthesis vector is a vector obtained by summing each column in the alignment table and the weight vector according to element products; performing scale standardization processing based on the synthesized vector and the normalization table to obtain a standard vector, wherein the standard vector is a vector obtained by linearly scaling the synthesized vector according to the value range of the synthesized vector; and performing boundary mapping processing based on the standard vector and the weight vector to obtain a scoring vector.
6. A method of determining a fetal heart rate as claimed in claim 5 wherein the method of performing a consistency computation on the segmentation window set based on the target peak set to obtain a consistency indicator comprises: performing adjacent pairing processing according to the target peak set and the segmented window set to obtain a pairing table, wherein the pairing table is a peak index pairing set established in the segmented window set according to adjacent time slices; Performing phase distance calculation according to the pairing table to obtain a phase distance vector, and performing displacement calculation according to the pairing table to obtain a displacement vector, wherein the phase distance vector is a vector formed by the arrival time differences of each pair of peak indexes in the pairing table on a time axis, and the displacement vector is a vector formed by the difference values of each pair of peak indexes in the pairing table on position coordinates; performing smoothing calculation according to the displacement vector to obtain a smoothing vector, and performing weight calculation based on the phase distance vector and the smoothing vector to obtain a confidence vector, wherein the smoothing vector is a vector obtained by performing sliding average value operation on the displacement vector; and performing weighted synthesis processing based on the phase distance vector and the confidence coefficient vector to obtain a consistent index.
7. The method of claim 6, wherein the step of performing a weighted synthesis process based on the phase distance vector and the confidence vector to obtain the coincidence indicator comprises: performing column alignment processing according to the period distance vector and the confidence coefficient vector to obtain an alignment table; Performing missing filling processing according to the alignment table to obtain a filling table, wherein the filling table is a two-dimensional table with a vacant position in the alignment table replaced by adjacent interpolation; performing weighted summation according to the filling table to obtain a synthetic vector; and performing scale mapping processing based on the synthesized vector and the phase distance vector to obtain a consistent index.
8. The method of claim 7, wherein the step of performing path selection processing based on the coincidence index to obtain the cycle trajectory comprises: Thresholding is carried out according to the consistency index to obtain candidate Boolean vectors; performing adjacency modeling according to the consistency index to obtain an adjacency relation table, wherein the adjacency relation table is an adjacent index pair set established according to the sequence index of the consistency index; Performing cost calculation processing based on the candidate Boolean vector and the adjacency list to obtain a cost vector; And executing minimum path processing according to the cost vector to obtain a periodic track, wherein the periodic track is an index sequence with minimum accumulated cost in an index domain of the consistent index.
9. A method of determining a fetal heart rate as claimed in claim 8 wherein the method of performing a minimum path process based on the cost vector to derive the periodic trajectory comprises: Performing index generation processing according to the cost vector to obtain a sequence index vector, and performing starting point setting processing according to the sequence index vector and the cost vector to obtain a starting point index, wherein the sequence index vector is a vector with the same length as the cost vector, and the starting point index is the first element of the sequence index vector; Performing adjacency construction processing according to the sequential index vectors to obtain an adjacency index table, and performing transfer mask processing according to the adjacency index table to obtain mask vectors; Performing accumulated update processing based on the cost vector and the adjacent index table to obtain an accumulated table, and performing parent index update processing according to the accumulated table and the adjacent index table to obtain a parent index vector, wherein the parent index vector is a vector aligned with the sequential index vector; and performing backtracking generation processing based on the accumulation table and the parent index vector to obtain a periodic track.
10. A fetal heart rate determination system for implementing a method of determining a fetal heart rate as claimed in any one of claims 1 to 9, comprising: the denoising normalization module is configured to perform denoising normalization processing on an echo sequence to obtain a target sequence, wherein the echo sequence is a time sequence acquired by a probe and subjected to demodulation envelope; The segmentation generation module is configured to execute segmentation generation processing on the target sequence based on a preset time axis to obtain a segmentation window set, wherein the segmentation window set is a time slice set established according to a preset window length and a step length; the peak parameter scoring module is configured to execute peak parameter extraction processing on the target sequence according to the segmentation window set to obtain a peak parameter table, execute separable scoring processing on the basis of the peak parameter table to obtain a target peak set, wherein the peak parameter table is a column structure comprising peak amplitude, peak width and neighborhood contrast; The uniform path searching module is configured to execute uniform degree calculation processing on the segmentation window set according to the target peak set to obtain a uniform index, and execute path selection processing based on the uniform index to obtain a periodic track; And the period generation module is configured to execute period generation processing based on the period track to obtain a period sequence, wherein the period sequence is a period vector arranged in time sequence.

Description

Fetal heart rate determining method and system Technical Field The invention relates to the technical field of fetal heart signal processing, in particular to a fetal heart rate determining method and a fetal heart rate determining system. Background The household fetal heart monitoring equipment generally acquires the fetal heart echo envelope by a handheld Doppler probe, forms a cardiac cycle and fetal heart rate curve under the constraint of end-side computing force and low time delay, adopts the idea of detecting the envelope peak in order to realize cycle estimation in the prior art, or extracting rhythm information by means of transform domain characterization, correlation measurement and the like, assisting smoothing and screening on time sequence to improve usability, the method can give continuous estimation results when the signal-to-noise ratio is higher, the probe posture is more stable and the rhythm change is gentle, and the basic requirements of partial family monitoring are met. However, the household handheld working condition has significant unstable factors, the echo energy is reduced and the peak shape is slowed down due to the changes of the incidence angle and the fetal position, so that a weak echo condition is formed, and the micro-movement of the probe and the abdominal wall also causes the adjacent time segments to generate observable time sequence deviation, so that the periodic continuity on the time axis is difficult to stably maintain due to the fact that the local judgment and the general smoothness are only relied on, and accordingly, the technical problem to be solved is how to ensure the problem of periodic estimation continuity under the conditions of enhancing the weak echo peak value separability and enhancing the periodic consistency of the adjacent segments. In view of the above, the present invention provides a method and a system for determining a fetal heart rate to solve the above-mentioned problems. Disclosure of Invention In order to overcome the above-mentioned drawbacks of the prior art, the present invention provides a method and system for determining a fetal heart rate. In order to achieve the above purpose, the present invention provides the following technical solutions: In a first aspect, there is provided a fetal heart rate determination method comprising: Denoising normalization processing is carried out on the echo sequence to obtain a target sequence, wherein the echo sequence is a time sequence which is acquired by a probe and is subjected to demodulation envelope; Executing segment generation processing on the target sequence based on a preset time axis to obtain a segment window set, wherein the segment window set is a time slice set established according to a preset window length and a step length; Performing peak parameter extraction processing on the target sequence according to the segmented window set to obtain a peak parameter table, and performing separable scoring processing on the basis of the peak parameter table to obtain a target peak set, wherein the peak parameter table is a column structure comprising peak amplitude, peak width and neighborhood contrast; Performing consistency computation processing on the segmentation window set according to the target peak set to obtain a consistency index, and performing path selection processing based on the consistency index to obtain a periodic track; and executing cycle generation processing based on the cycle track to obtain a cycle sequence, wherein the cycle sequence is a cycle vector arranged in time sequence. In some embodiments, a method for performing peak parameter extraction processing on a target sequence according to a segmentation window set to obtain a peak parameter table includes: Performing peak candidate generation processing on the target sequence according to the segmentation window set to obtain a peak candidate set, wherein the peak candidate set is a peak index set meeting local extremum constraint in each time slice; Performing amplitude calculation processing on the target sequence according to the peak candidate set to obtain a peak amplitude vector; performing width calculation processing on the target sequence according to the peak candidate set to obtain a peak width vector; performing neighborhood contrast processing on the target sequence according to the peak candidate set to obtain a contrast vector, wherein the contrast vector is a vector of a difference value between the peak top and a neighborhood baseline; and performing table assembly processing based on the peak amplitude vector, the peak width vector and the comparison vector to obtain a peak parameter table. In some embodiments, a method for performing neighborhood contrast processing on a target sequence according to a peak candidate set to obtain a contrast vector includes: performing neighborhood positioning processing according to the peak candidate set and the target sequence to obtain a neighborhoo