Preterm birth prediction method and device based on uterine vector electromyographic signals and processing equipment

Abstract

The application provides a preterm birth prediction method and device based on uterine vector electromyographic signals and a processing device. A novel electrode arrangement scheme is specially designed. The collected uterine vector electromyographic signals are different from traditional uterine surface electromyographic signals, can more comprehensively reflect the spatial characteristics of uterine contraction activities, contain more signal characteristics related to preterm birth, and thus can obtain higher prediction accuracy for preterm birth under the condition of fewer measurement electrodes, and has clinical application value.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of non-invasive uterine myoelectric signal detection, in particular to a preterm birth prediction method and device based on uterine vector myoelectric signals and a processing device. BACKGROUND

[0002] Early screening of preterm birth helps to identify risks early and reduce the incidence of complications of preterm birth. The current main method for predicting preterm birth is based on uterine myoelectric monitoring to predict preterm birth. The uterine myoelectric monitoring method actually places electrodes on the surface of the pregnant woman's abdomen to record the electrical activity conducted from the uterine muscle layer to the abdominal surface to obtain a uterine surface electromyogram.

[0003] The uterine surface electromyogram is a non-invasive monitoring method, which specifically collects myoelectric signals through the abdominal surface to reflect the electrical signal characteristics during the contraction of uterine muscle cells. Subsequently, the myoelectric signal-related indicators, including time-domain indicators, frequency-domain indicators, myoelectric conduction velocity, and nonlinear indicators, are analyzed to determine the true and false contractions of the uterus and the degree of contraction, thereby achieving preterm birth prediction.

[0004] However, the present inventors have found that the current preterm birth prediction method has the following problems:

[0005] 1) The signal feature extraction for preterm birth prediction is complex. The uterine surface electromyogram is easily disturbed by motion, fetal movement, cardiac electrical activity, and respiratory activity. The extraction of time-domain, frequency-domain, and myoelectric conduction velocity indicators from the collected myoelectric signals requires complex filtering and signal denoising, which may result in the loss of useful signal features and reduce the prediction accuracy to some extent.

[0006] 2) The uterine surface electromyogram cannot monitor comprehensive uterine electrophysiological activity. The uterine contraction movement is an anisotropic three-dimensional spatial activity. The uterine myoelectric signal can be regarded as the spatial sum of uterine muscle cell electrical activity. The electrodes of the surface electromyogram are arranged on the maternal abdomen, and the collected uterine surface electromyogram can only represent the local uterine contraction movement characteristics and cannot represent the comprehensive spatial activity characteristics.

[0007] 3) Some solutions propose using a multi-electrode (16-64) arrangement to collect uterine myoelectric signals in a larger area to obtain more abundant uterine electrophysiological activity characteristics. However, this electrode arrangement is too complex, and the labor and data collection costs are too high, which is not conducive to clinical application and promotion. SUMMARY

[0008] The application provides a preterm birth prediction method and device based on uterine vector electromyographic signals and a processing device.

[0009] In a first aspect, the application provides a preterm birth prediction device based on uterine vector electromyographic signals, and the device comprises:

[0010] Obtaining uterine vector electromyographic signals collected by a mother under a preset electrode arrangement scheme, wherein the preset electrode arrangement scheme involves six electrodes and is specifically configured as electrode pairs arranged orthogonally in pairs, the uterine vector electromyographic signals specifically include uterine electromyographic signals in three orthogonal directions corresponding to three electrode pairs, and the uterine vector electromyographic signals further include differential signals obtained by differentiating two signals of each electrode pair;

[0011] Inputting the uterine vector electromyographic signals into a preconfigured prediction model, wherein the prediction model is specifically obtained by training sample uterine vector electromyographic signals corresponding to the preset electrode arrangement scheme, and the prediction model is used to predict a corresponding preterm birth condition based on the uterine vector electromyographic signals input into the model;

[0012] Extracting a preterm birth condition prediction result output by the prediction model.

[0013] In a second aspect, the application provides a preterm birth prediction device based on uterine vector electromyographic signals, and the device comprises:

[0014] An obtaining unit is configured to obtain uterine vector electromyographic signals collected by a mother under a preset electrode arrangement scheme, wherein the preset electrode arrangement scheme involves six electrodes and is specifically configured as electrode pairs arranged orthogonally in pairs, the uterine vector electromyographic signals specifically include uterine electromyographic signals in three orthogonal directions corresponding to three electrode pairs, and the uterine vector electromyographic signals further include differential signals obtained by differentiating two signals of each electrode pair;

[0015] A prediction unit is configured to input the uterine vector electromyographic signals into a preconfigured prediction model, wherein the prediction model is specifically obtained by training sample uterine vector electromyographic signals corresponding to the preset electrode arrangement scheme, and the prediction model is used to predict a corresponding preterm birth condition based on the uterine vector electromyographic signals input into the model;

[0016] An extracting unit is configured to extract a preterm birth condition prediction result output by the prediction model.

[0017] In a third aspect, the present application provides a processing device, comprising a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program in the memory to implement the method in the first aspect or any possible implementation manner of the first aspect.

[0018] In a fourth aspect, the present application provides a computer readable storage medium, which stores a plurality of instructions, and the instructions are adapted to be loaded by a processor to implement the method in the first aspect or any possible implementation manner of the first aspect.

[0019] From the above, the present application has the following beneficial effects:

[0020] For the target of premature birth prediction based on uterine vector electromyographic signals, the present application uses a specially designed novel electrode arrangement scheme, and the collected uterine vector electromyographic signals are different from traditional uterine surface electromyographic signals, can more comprehensively reflect the spatial characteristics of uterine contraction activity, contain more signal characteristics related to premature birth, so as to obtain higher prediction accuracy for premature birth under the condition of fewer measurement electrodes, and has clinical application value. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0022] Figure 1 FIG. 1 is a flowchart of a premature birth prediction method based on uterine vector electromyographic signals according to the present application;

[0023] Figure 2 FIG. 2 is a scene diagram of an electrode arrangement scheme according to the present application;

[0024] Figure 3 FIG. 3 is a structure diagram of a premature birth prediction device based on uterine vector electromyographic signals according to the present application;

[0025] Figure 4 FIG. 4 is a structure diagram of a processing device according to the present application. DETAILED DESCRIPTION

[0026] With reference to the drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts are within the scope of the present application.

[0027] The terms "first", "second", and the like in the description and claims of the present application and the above drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "comprise" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device comprising a series of steps or modules does not necessarily limit to those steps or modules clearly listed, but can include other steps or modules not clearly listed or inherent to the process, method, product or device. The naming or numbering of steps appearing in the present application does not mean that the steps in the method flow must be executed in the time / logical order indicated by the naming or numbering, and the named or numbered flow steps can change the execution order according to the technical purpose to be achieved, as long as the same or similar technical effects can be achieved.

[0028] The division of modules appearing in the present application is a logical division, and in actual application, there can be another division mode, for example, a plurality of modules can be combined or integrated in another system, or some features can be ignored or not executed, in addition, the coupling or direct coupling or communication connection between the displayed or discussed modules can be through some interface, the indirect coupling or communication connection between the modules can be electrical or other similar forms, which are not limited in the present application. Moreover, the modules or sub-modules described as separate components can or can not be physically separate, can or can not be physical modules, or can be distributed into a plurality of circuit modules, and some or all of the modules can be selected according to actual needs to achieve the purpose of the present application.

[0029] Before introducing the preterm birth prediction method based on uterine vector electromyographic signal provided by the present application, the background content involved in the present application is first introduced.

[0030] The premature birth prediction method based on uterine vector electromyographic signals, the device and the computer readable storage medium provided by the application can be applied to a processing device, and through a specially designed novel electrode arrangement scheme, the collected uterine vector electromyographic signals are different from traditional uterine surface electromyographic signals, can more comprehensively reflect the spatial characteristics of uterine contraction activity, contain more signal characteristics related to premature birth, and thus under the condition of fewer measurement electrodes, higher prediction accuracy of premature birth can be obtained, and the method has clinical application value.

[0031] The premature birth prediction method based on uterine vector electromyographic signals mentioned in the application can be executed by a premature birth prediction device based on uterine vector electromyographic signals, or a server, a physical host or a user equipment (UE) or other types of processing devices integrated with the premature birth prediction device based on uterine vector electromyographic signals. The premature birth prediction device based on uterine vector electromyographic signals can be realized in the form of hardware or software, the UE can be a terminal device such as a smart phone, a tablet computer, a notebook computer, a desktop computer or a personal digital assistant (PDA), and the processing device can be set in the form of a device cluster.

[0032] Next, the premature birth prediction method based on uterine vector electromyographic signals provided by the application will be introduced.

[0033] Firstly, referring to Figure 1 , Figure 1 Fig. 1 shows a flowchart of the premature birth prediction method based on uterine vector electromyographic signals provided by the application. The premature birth prediction method based on uterine vector electromyographic signals provided by the application can specifically include the following steps S101 to S103:

[0034] In step S101, uterine vector electromyographic signals collected by a mother under a preset electrode arrangement scheme are acquired, wherein the preset electrode arrangement scheme involves six electrodes and is specifically configured as electrode pairs arranged orthogonally in pairs, the uterine vector electromyographic signals specifically include uterine electromyographic signals in three directions that are orthogonal and correspond to three electrode pairs, and the uterine vector electromyographic signals further include differential signals obtained by differentiating two signals of each electrode pair;

[0035] It can be understood that the present application focuses on the lighter and higher performance of the premature prediction target, and a novel electrode arrangement scheme is specially designed, which specifically involves only 6 measuring electrodes, and the 6 electrodes arranged on the maternal skin specifically form the measurement condition of the electrode pair arranged orthogonally in pairs, so that the uterine myoelectric signals corresponding to the three directions of the three electrode pairs arranged orthogonally can be measured, and the spatial characteristics are further enhanced by combining the differential signals to form the uterine vector myoelectric signal with spatial attributes. The special uterine vector myoelectric signal can provide corresponding data input for the premature prediction logic based on the uterine vector myoelectric signal specially designed by the present application.

[0036] That is, under the design of the present application scheme, the uterine vector myoelectric signal theoretically only needs 6 measuring electrodes to complete the signal acquisition work, and the clinical operability is obviously enhanced.

[0037] Among them, the measuring electrode can usually be the electrode of the existing multi-channel myoelectric acquisition instrument, or in some cases, it can also be the electrode of the multi-channel myoelectric acquisition instrument specially made for the present application scheme.

[0038] In addition, it should be understood that corresponding to the application of the measuring electrode, a grounding electrode or reference electrode is usually configured, which can be configured at a specific position such as the surface of the maternal thigh.

[0039] Further, for the acquisition and processing of the uterine vector myoelectric signal, it can be real-time acquisition and processing or extraction of existing signals in specific operation, which is possible in actual situation.

[0040] Among them, as mentioned above, in addition to the myoelectric signals measured by the 6 measuring electrodes as the subsequent data input, the present application further processes the 6 basic myoelectric signals to obtain new dimensional myoelectric signal features to provide more abundant data input and data reference.

[0041] Specifically, the overall uterine vector myoelectric signal also includes the differential signal obtained by differentiating the two signals of each electrode pair.

[0042] Among them, the difference is a mature mathematical operation.

[0043] That is to say, after obtaining the uterine myoelectric signals in the three orthogonal directions corresponding to the three basic electrode pairs, the two signals of each electrode pair can be differentiated to obtain the differential signals of the two signal differences of each electrode pair, and the uterine vector myoelectric signals composed of the uterine myoelectric signals in the three orthogonal directions corresponding to the three basic electrode pairs form a whole, so that compared with the original six-channel uterine myoelectric signals, nine-channel uterine myoelectric signals are formed.

[0044] At the same time, as a specific embodiment, referring to Figure 2 A scene schematic diagram of the electrode arrangement scheme of the present application is shown, and for the three electrode pairs involved in the preset electrode arrangement scheme of the present application, specifically:

[0045] 1) The first electrode pair is composed of the first electrode (i.e. electrode 1 or point 1) and the fourth electrode (i.e. electrode 4 or point 4), the first electrode is arranged above the middle of the maternal abdominal wall, and the fourth electrode is arranged below the middle of the maternal waist and back;

[0046] The preset distance between the first electrode and the maternal waist horizontal line (or waist reference line, horizontal line of navel) (i.e. the specific distance of the first electrode arranged above the middle of the maternal abdominal wall) and the preset distance between the fourth electrode and the maternal waist horizontal line (i.e. the specific distance of the fourth electrode arranged below the middle of the maternal waist and back) are not necessarily equal, which is related to the asymmetry of the abdominal wall and the waist and back, and also related to the body size difference between different individuals.

[0047] Further, the following limiting conditions can exist:

[0048] The first electrode is located above the waist line, generally not more than the lowermost rib plane of the chest, and the second electrode is located below the waist line of the back, generally not more than the end of the coccyx.

[0049] 2) The second electrode pair is composed of the second electrode (i.e. electrode 2 or point 2) and the third electrode (i.e. electrode 3 or point 3), the second electrode is arranged above the middle of the maternal waist and back, and the third electrode is arranged below the middle of the maternal abdominal wall, and the connecting line of the first electrode pair and the connecting line of the second electrode pair form a perpendicular relationship;

[0050] Among them, the preset distance between the second electrode and the maternal waist horizontal line (i.e. the specific distance of the second electrode arranged above the middle of the waist and back), and the preset distance between the third electrode and the maternal waist horizontal line (i.e. the specific distance of the third electrode arranged below the middle of the maternal abdominal wall), both have the feature that they are not necessarily equal as mentioned above.

[0051] And, as can be clearly seen from the figure, the line connecting point 1 and point 4, and the line connecting point 2 and point 3, have a perpendicular relationship, that is, a 90° angle relationship.

[0052] 3) The third electrode pair is composed of the fifth electrode (i.e., electrode 5 or point 5) and the sixth electrode (i.e., electrode 6 or point 6), and the fifth electrode and the sixth electrode are along the maternal waist horizontal line and are respectively located on the left side of the abdominal wall and the right side of the abdominal wall. The line connecting the third electrode pair forms an orthogonal relationship with the line connecting the first electrode pair and the line connecting the second electrode pair.

[0053] Among them, on the maternal waist horizontal line, the preset distance between the fifth electrode and the navel is equal to the preset distance between the sixth electrode and the navel.

[0054] And, as can be clearly seen from the figure, the line connecting point 1 and point 4, and the line connecting point 2 and point 3, are both perpendicular to the line connecting point 5 and point 6, or in other words, the plane in which points 1, 2, 3, and 4 are located (the plane in which the navel is located and which is perpendicular to the horizontal plane) is perpendicular to the line connecting point 5 and point 6.

[0055] In addition, in addition to the differential signal obtained by differentiating the two signals corresponding to the above electrode pairs, the application can also perform new differentiation processing on two signals not belonging to the same electrode pair to obtain new signal characteristics in a new direction (or dimension) to continue to enhance the corresponding vector signal content and quality.

[0056] For this purpose, as an exemplary embodiment, the uterine vector electromyogram signal can also include an auxiliary differential signal obtained by differentiating two signals not belonging to the same electrode pair (or belonging to different electrode pairs).

[0057] As an example, the two signals measured by the first electrode (belonging to the first electrode pair) and the second electrode (belonging to the second electrode pair) can be differentiated to obtain a new differential signal.

[0058] It can be understood that under the new differential signal configuration mechanism here, more dimensional signal characteristics can be flexibly formed to promote subsequent more delicate and more accurate premature birth prediction processing.

[0059] Step S102, input the uterine vector electromyogram signal into a pre-configured prediction model, wherein the prediction model is specifically trained by a labeled sample uterine vector electromyogram signal corresponding to a preset electrode arrangement scheme, and the prediction model is used to predict a corresponding premature birth based on the model input uterine vector electromyogram signal;

[0060] It can be understood that, as introduced before, under the preset electrode arrangement scheme specially designed by the present application, the corresponding prediction model of the present application is pre-trained, which can also be called a preterm birth prediction model, and the training sample, i.e. the labeled sample uterine vector myoelectric signal, is also obtained by configuring the preset electrode scheme.

[0061] Among them, for the sample uterine vector myoelectric signal, it is usually obtained by configuring the corresponding measurement electrode under the preset electrode scheme, and in some cases, it can also be extracted from some ready-made signals, that is, it does not exclude the case that the sample uterine myoelectric signal corresponding to the individual measurement electrode itself is collected in advance according to some signal collection requirements, or in some cases, it can also be a signal processed on the basis of the obtained uterine vector myoelectric signal to further improve the sample size, or in some cases, it can also be a uterine vector myoelectric signal directly constructed, which is convenient for obtaining the corresponding sample uterine vector myoelectric signal for special groups or special symptoms.

[0062] And for the labeling processing of the sample uterine vector myoelectric signal, i.e. the labeling processing of the theoretical model prediction result-preterm birth condition prediction result, it can usually be manually labeled by artificial, or it can also be completed by the corresponding automatic labeling tool, and the automatic labeling tool needs to be pre-configured with the corresponding automatic labeling logic.

[0063] In this way, after obtaining the sample uterine vector myoelectric signal with complete labeling, the initial model can be trained under the configuration of the corresponding model training scheme and the specific loss function, and the loss function is quantified by means of the model prediction result and the sample label. Under the back propagation algorithm, the weight parameters involved in the related model structure are constantly optimized and recalculated according to the loss function calculation result, so that when the model training requirements such as training time, training times, prediction accuracy are met, the training can be completed, and the prediction model can be put into actual use.

[0064] Therefore, it can also be obtained that the present application can also involve data processing in the pre-model training link, and for this, as an exemplary embodiment, the method of the present application can also include:

[0065] Obtaining a sample uterine vector myoelectric signal corresponding to a preset electrode arrangement scheme;

[0066] Configuring a corresponding label for the sample uterine vector myoelectric signal;

[0067] Training a prediction model by the labeled sample uterine vector myoelectric signal corresponding to the preset electrode arrangement scheme.

[0068] In addition, it can be understood that, for the differential signal in the uterine vector myoelectric signal, in some cases, the prediction model can also be used to make a difference based on the input of 6 basic uterine myoelectric signals.

[0069] At this time, after obtaining the uterine vector myoelectric signal of the user currently required to predict the premature birth, it can be input into the prediction model to expand the corresponding premature birth prediction processing.

[0070] In addition, for the model structure involved in the prediction model, the present application also gives a specific landing supporting scheme.

[0071] Correspondingly, as an exemplary embodiment, the prediction model of the present application can specifically include a signal encoding module and a signal decoding module, and the signal encoding module and the signal decoding module are configured by a BiLSTM model with separately set parameters.

[0072] That is, the prediction model of the present application is based on the coding and decoding structure design of the BiLSTM model, wherein the BiLSTM model is a bidirectional long short-term memory (Bi-directional Long Short-Term Memory, BiLSTM), which is a relatively mature existing model, and its main advantages are:

[0073] 1. Simultaneously consider past and future information: BiLSTM runs two LSTM layers simultaneously, one processes data in forward order and the other in reverse order, so that the output of each time step contains information before and after the current time step, thus better capturing dynamic characteristics and context relationships in sequence data;

[0074] 2. Improve information utilization: through forward and backward propagation, BiLSTM can more comprehensively utilize forward and backward feature information of sequences, significantly improving modeling ability for sequence data, and this bidirectional structure enables the model to better understand semantics and context relationships in sequence data;

[0075] 3. Suitable for complex sequence data: BiLSTM is particularly suitable for data with time-varying and complex characteristics, and can effectively process nonlinear and non-stationary signals to capture dynamic characteristics in sequence data.

[0076] More specifically, the signal encoding module in the working process can specifically include the following processing contents:

[0077] 1) Feature extraction of signals in each channel of the uterine vector myoelectric signal input to the model, outputting corresponding initial one-dimensional time sequence features;

[0078] Corresponding to the previous embodiment, the uterine vector myoelectric signal input into the model can be specifically 6-channel basic uterine myoelectric signals, or can be 6-channel basic uterine myoelectric signals and 3-channel differential signals.

[0079] Here corresponds to the feature conversion link on the input side of the model, which converts the initial signal into the corresponding initial one-dimensional time sequence feature. In a popular way, it is also called signal coding processing, which converts the initial signal into the corresponding coded output.

[0080] 2) Vertically splice each initial one-dimensional time sequence feature to form a two-dimensional feature matrix;

[0081] It can be understood that after obtaining the initial one-dimensional time sequence feature of each channel in the previous, vertical splicing or channel splicing can be performed to form a two-dimensional feature matrix with a specification of channel number x encoder input / output size (such as 6xS, 9xS).

[0082] 3) Adopting a multi-layer convolutional neural network to perform feature transformation and dimension conversion, and outputting a target one-dimensional time sequence feature.

[0083] After splicing the two-dimensional feature matrix, convolution processing can be continued to realize feature transformation and dimension conversion, so as to output a one-dimensional feature matrix with a specification of 1x encoder output size, i.e. a target one-dimensional time sequence feature.

[0084] Specifically, as an example, the signal coding process of the signal coding module, i.e. the encoder, is configured as:

[0085] The number of hidden layer units of the BiLSTM network is set to 128, and the number of layers is set to 2; the input size corresponding to the processing of each signal alone is set to 1; the output size of the BiLSTM is set to S, which is consistent with the sequence length of the signal segment.

[0086] The BiLSTM network model is used to independently encode each channel in the input signal segment; the coding outputs of each channel are spliced by channel to form a 9xS two-dimensional feature matrix; then a 1x1 convolutional neural network is used to convolve the two-dimensional feature matrix by channel dimension, and a 1xS one-dimensional feature matrix is output.

[0087] On the other hand, the signal decoding module in the working process can specifically include the following processing contents:

[0088] The target one-dimensional time sequence feature is expanded to predict the premature birth situation, wherein the premature birth situation or pregnancy outcome specifically includes three types, corresponding to which:

[0089] The first type is premature birth (delivery before 37 weeks of pregnancy);

[0090] The second type is full-term and normal delivery;

[0091] The third type is full-term, but induced labor or cesarean section.

[0092] It can be understood that the signal decoding module directly decodes the target one-dimensional time sequence features transmitted by the preceding signal encoding module to fuse as the final model output result, that is, the corresponding label of the specific premature birth situation type.

[0093] Among them, the label of the first type can be specifically configured as 0, the label of the second type can be specifically configured as 1, and the label of the third type can be specifically configured as 2.

[0094] Specifically, as an example, the number of hidden layer units of the signal decoding module, that is, the decoder, is set to 64, the number of layers is set to 1, and the output size is set to 2. The output vector after decoding is first calculated by the Softmax function, and then the maximum value is set to 1 and the others are set to zero to form a binary output vector corresponding to the above three labels.

[0095] It can be understood that based on the above-mentioned coding and decoding structure design, or based on the above-mentioned coding-decoding signal processing method, without involving or without the need for a complex filtering process, the powerful feature extraction capability of the neural network can be directly used to extract subtle features and multi-dimensional conversion of the uterine myoelectric signal and / or the differential signal. The subtle feature extraction can include more nonlinear signal features, and the multi-dimensional conversion is conducive to fully identifying the time domain and frequency domain characteristics, and these measures can effectively improve the accuracy of the current method for predicting premature birth.

[0096] In addition, for the acquisition of sample uterine vector myoelectric signals, the application also gives a specific supporting scheme.

[0097] Specifically, as an example embodiment, acquiring sample uterine vector myoelectric signals corresponding to a preset electrode arrangement scheme can include:

[0098] 1) Acquire M initial sample uterine vector myoelectric signals of cases with complete pregnancy outcomes corresponding to a preset electrode arrangement scheme, wherein the initial sample uterine vector myoelectric signal is not less than 30 minutes, the signal sampling rate is not less than 360Hz, and M is greater than 200;

[0099] Among them, for the case, or the initial sample uterine vector myoelectric signal, it can be specifically limited to not more than 26 weeks of pregnancy, corresponding to the actual situation of existing premature birth prediction needs.

[0100] And the specific sample content can be as set above, involving three types of premature birth situations or pregnancy outcomes, and indicated by labels 0, 1, and 2.

[0101] 2) slice the initial sample uterine vector electromyographic signal according to time period, form N groups of signal segments, wherein N is greater than 600, each signal segment contains 9 sequence data with length S, S is the standard sequence length of model input (or signal coding module input), S is greater than 360 sampling points;

[0102] It can be understood that the specific slicing process here is beneficial to standardized and more delicate sample configuration effect in actual application.

[0103] 3) combine the N groups of signal segments of M cases respectively to form MxN groups of signal segments, and then randomly shuffle to form corresponding batch sequence data, wherein the batch size is 16-128 groups (signal segments).

[0104] It can be understood that the random generation of sequence data with the same batch size or different batch size can be used as the final highly diversified training sample, which is sent into the signal coding module according to the batch size to promote the specific model training process.

[0105] In addition, as an exemplary embodiment, the cross entropy (Cross Entropy) function can be used as an existing loss function.

[0106] In this case, the annotation data involved above can be formed into a corresponding vector using one-hot encoding for the cross entropy function to calculate the loss error value between the model output and the label true value.

[0107] Step S103, extracting the premature birth prediction result of the prediction model output.

[0108] It is easy to understand that when the prediction model completes the premature birth prediction process and outputs the corresponding premature birth prediction result, the premature birth prediction result can be extracted, and the premature birth prediction task for the current user is completed.

[0109] Among them, the premature birth prediction result indicates the specific adapted premature birth situation type, and further can give the corresponding probability.

[0110] At the same time, corresponding to the data application requirement, further data application links can also be involved, for example, based on the premature birth prediction result, local storage, remote storage, result display, result printing, result pushing or result analysis (disease diagnosis, etc.) can be carried out.

[0111] It can be understood that the specific data application processing can be adaptively adjusted according to the pre-configuration and the real-time configuration of the data application strategy, and is more flexible in actual cases.

[0112] Finally, in general, for the preterm birth prediction based on the uterine vector myoelectric signal, the application can obtain higher prediction accuracy for preterm birth under the condition of fewer measurement electrodes by using a novel electrode arrangement scheme specially designed, and the uterine vector myoelectric signal collected is different from the traditional uterine surface myoelectric signal, which can more comprehensively reflect the spatial characteristics of uterine contraction activity and contain more signal characteristics related to preterm birth, thereby having clinical application value.

[0113] The above is an introduction to the preterm birth prediction method based on the uterine vector myoelectric signal provided by the application. In order to better implement the preterm birth prediction method based on the uterine vector myoelectric signal provided by the application, the application also provides a preterm birth prediction device based on the uterine vector myoelectric signal from the functional module.

[0114] Reference Figure 3 , Figure 3 is a structural schematic diagram of the preterm birth prediction device based on the uterine vector myoelectric signal, in the application, the preterm birth prediction device based on the uterine vector myoelectric signal 300 can specifically include the following structures:

[0115] The acquisition unit 301 is configured to acquire the uterine vector myoelectric signal collected by the mother under a preset electrode arrangement scheme, wherein the preset electrode arrangement scheme involves 6 electrodes and is specifically configured as electrode pairs arranged orthogonally in pairs, the uterine vector myoelectric signal specifically includes uterine myoelectric signals in three orthogonal directions corresponding to three electrode pairs, and the uterine vector myoelectric signal further includes differential signals obtained by differentiating two signals of each electrode pair;

[0116] The prediction unit 302 is configured to input the uterine vector myoelectric signal into a pre-configured prediction model, wherein the prediction model is specifically trained by a sample uterine vector myoelectric signal corresponding to the preset electrode arrangement scheme, and the prediction model is used to predict the corresponding preterm birth condition based on the uterine vector myoelectric signal input into the model;

[0117] The extraction unit 303 is configured to extract the preterm birth condition prediction result output by the prediction model.

[0118] As an exemplary embodiment, for the three electrode pairs, there are:

[0119] The first electrode pair is composed of the first electrode and the fourth electrode, and the first electrode is arranged above the middle of the abdominal wall of the mother, and the fourth electrode is arranged below the middle of the mother's waist and back;

[0120] The second electrode pair is composed of a second electrode and a third electrode, the second electrode is arranged above the median of the back of the mother's waist, and the third electrode is arranged below the median of the abdominal wall of the mother, and the connecting line of the first electrode pair and the connecting line of the second electrode pair form a vertical relationship;

[0121] The third electrode pair is composed of a fifth electrode and a sixth electrode, the fifth electrode and the sixth electrode are along the horizontal line of the mother's waist and are respectively located on the left side of the abdominal wall and the right side of the abdominal wall, and the connecting line of the third electrode pair forms an orthogonal relationship with the connecting line of the first electrode pair and the connecting line of the second electrode pair respectively.

[0122] As another exemplary embodiment, the uterine vector electromyogram signal further includes an auxiliary differential signal obtained by differentiating two signals not belonging to the same electrode pair. As another exemplary embodiment, the prediction model includes a signal encoding module and a signal decoding module, and the signal encoding module and the signal decoding module are configured by a BiLSTM model with separately set parameters;

[0123] The signal encoding module includes the following processing contents in the working process:

[0124] The signal of each channel of the uterine vector electromyogram signal input to the model is extracted, and the corresponding initial one-dimensional time sequence feature is output;

[0125] Each initial one-dimensional time sequence feature is vertically spliced to form a two-dimensional feature matrix;

[0126] A multi-layer convolutional neural network is used for feature transformation and dimension conversion, and a target one-dimensional time sequence feature is output;

[0127] The signal decoding module includes the following processing contents in the working process:

[0128] The target one-dimensional time sequence feature is subjected to prediction processing of the premature birth situation, wherein the premature birth situation specifically includes three types, and correspondingly, there are:

[0129] The first type is premature birth;

[0130] The second type is full-term and normal delivery;

[0131] The third type is full-term, but induced labor or cesarean section.

[0132] As another exemplary embodiment, the device further includes a training unit 304, configured to:

[0133] Obtain a sample uterine vector electromyogram signal corresponding to a preset electrode arrangement scheme;

[0134] Configure a corresponding label for the sample uterine vector electromyogram signal;

[0135] The prediction model is trained using labeled uterine vector electromyography signals corresponding to a preset electrode arrangement.

[0136] As another exemplary embodiment, training unit 304 is specifically used for:

[0137] Acquire the initial sample uterine vector electromyography (EMG) signals of M cases with complete pregnancy outcomes corresponding to the preset electrode arrangement scheme. The initial sample uterine vector EMG signal pairs are not less than 30 minutes, the signal sampling rate is not less than 360 Hz, and M is greater than 200.

[0138] The initial sample uterine vector electromyography signal is sliced ​​into segments of 1 to 3 seconds in length to form N groups of signal segments, where N is greater than 600. Each signal segment contains 9 sequence data of length S, where S is the standard sequence length of the model input and S is greater than 360 sampling points.

[0139] The N signal segments from each of the M cases were merged to form M×N signal segments, which were then randomly shuffled to form corresponding batches of sequence data, with batch sizes ranging from 16 to 128 groups.

[0140] As another exemplary embodiment, the loss function used in the training process is specifically the cross-entropy function, and the labeled data is one-hot encoded to form corresponding vectors for use in the calculation of the cross-entropy function.

[0141] This application also provides a processing device from a hardware architecture perspective, see [link / reference]. Figure 4 , Figure 4 This diagram illustrates a structural schematic of the processing device of this application. Specifically, the processing device may include a processor 401, a memory 402, and an input / output device 403. The processor 401 executes the computer program stored in the memory 402 to implement, for example... Figure 1 The steps of the preterm birth prediction method based on uterine vector electromyography signals in the corresponding embodiment; or, when the processor 401 executes the computer program stored in the memory 402, it implements as follows: Figure 3 Corresponding to the functions of each unit in the embodiment, the memory 402 is used to store the functions executed by the processor 401 as described above. Figure 1 The computer program required for the preterm birth prediction method based on uterine vector electromyography signals in the corresponding embodiment.

[0142] For example, a computer program may be divided into one or more modules / units, one or more of which are stored in memory 402 and executed by processor 401 to complete this application. One or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in a computer device.

[0143] The processing device can include, but is not limited to, the processor 401, the memory 402, the input / output device 403. Those skilled in the art can understand that the schematic diagram is only an example of the processing device, and does not constitute a limitation on the processing device, and can include more or less components than the schematic diagram, or combine certain components, or different components, for example, the processing device can also include a network access device, a bus, etc., and the processor 401, the memory 402, the input / output device 403 are connected through the bus.

[0144] The processor 401 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor is the control center of the processing device, and connects various parts of the entire device through various interfaces and lines.

[0145] The memory 402 can be used to store computer programs and / or modules, and the processor 401 realizes various functions of the computer device by running or executing the computer programs and / or modules stored in the memory 402, and calling the data stored in the memory 402. The memory 402 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required by a function, etc.; the data storage area can store data created according to the use of the processing device, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, for example, a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0146] When the processor 401 is used to execute the computer programs stored in the memory 402, the following functions can be realized:

[0147] acquire a uterine vector electromyography signal collected by a mother under a preset electrode arrangement scheme, wherein the preset electrode arrangement scheme involves 6 electrodes and is specifically configured as electrode pairs arranged orthogonally in pairs, the uterine vector electromyography signal specifically includes uterine electromyography signals in three orthogonal directions corresponding to the three electrode pairs, and the uterine vector electromyography signal further includes a differential signal obtained by differentiating two signals of each electrode pair;

[0148] input the uterine vector electromyography signal into a pre-configured prediction model, wherein the prediction model is specifically trained by a sample uterine vector electromyography signal corresponding to the preset electrode arrangement scheme, and the prediction model is used to predict a corresponding premature birth condition based on the uterine vector electromyography signal input into the model;

[0149] extract a premature birth condition prediction result output by the prediction model.

[0150] Those skilled in the art can clearly understand the specific working process of the premature birth prediction device based on the uterine vector electromyography signal, the processing device and the corresponding units thereof described above for the convenience and brevity of description, which can be referred to as Figure 1 The description of the premature birth prediction method based on the uterine vector electromyography signal in the corresponding embodiments will not be repeated here.

[0151] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by related hardware controlled by the instructions, which can be stored in a computer readable storage medium and loaded and executed by a processor.

[0152] To this end, the present application provides a computer readable storage medium, which stores a plurality of instructions capable of being loaded by a processor to execute the present application as Figure 1 The steps of the premature birth prediction method based on the uterine vector electromyography signal in the corresponding embodiments can be specifically referred to as Figure 1 The description of the premature birth prediction method based on the uterine vector electromyography signal in the corresponding embodiments will not be repeated here.

[0153] The computer readable storage medium can include a read only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0154] Due to the instructions stored in the computer readable storage medium, the present application as Figure 1 The steps of the premature birth prediction method based on the uterine vector electromyography signal in the corresponding embodiments, therefore, the present application as Figure 1The beneficial effects that can be achieved by the preterm birth prediction method based on uterine vector electromyographic signals in the corresponding embodiments are described in the foregoing description, which will not be repeated here.

[0155] The preterm birth prediction method based on uterine vector electromyographic signals, the device, the processing equipment and the computer readable storage medium provided by the present application are described in detail above, the principles and implementation modes of the present application are described in this paper, and the above embodiment description is only used to help understand the core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed, and the above description should not be understood as the limitation of the present application.

Claims

1. A method for predicting preterm birth based on uterine vector electromyography signals, characterized in that, The method includes: The method acquires uterine vector electromyography (EMG) signals collected from the mother under a preset electrode arrangement scheme, wherein the preset electrode arrangement scheme involves 6 electrodes, specifically configured as pairs of electrodes arranged orthogonally in pairs. The uterine vector EMG signals specifically include uterine EMG signals corresponding to 3 of the electrode pairs in three orthogonal directions. The uterine vector EMG signals also include differential signals obtained by differentiating the two signals of each electrode pair. The uterine vector electromyography signal is input into a pre-configured prediction model, wherein the prediction model is specifically trained by a labeled sample uterine vector electromyography signal corresponding to the preset electrode arrangement scheme, and the prediction model is used to predict the corresponding preterm birth situation based on the uterine vector electromyography signal input to the model. Extract the prediction results of preterm birth from the prediction model; For the three electrode pairs, we have: The first electrode pair consists of a first electrode and a fourth electrode. The first electrode is positioned slightly above the center of the mother's abdominal wall, and the fourth electrode is positioned slightly below the center of the mother's lower back. The second electrode pair consists of a second electrode and a third electrode. The second electrode is positioned slightly above the center of the mother's back, and the third electrode is positioned slightly below the center of the mother's abdominal wall. The line connecting the first electrode pair and the line connecting the second electrode pair are perpendicular to each other. The third electrode pair consists of the fifth and sixth electrodes, which are located along the horizontal line of the mother's waist and on the left and right sides of the abdominal wall, respectively. The line connecting the third electrode pair is orthogonal to the line connecting the first electrode pair and the line connecting the second electrode pair.

2. The method according to claim 1, characterized in that, The uterine vector electromyography signal also includes an auxiliary differential signal obtained by differentiating two signals that do not belong to the same electrode pair.

3. The method according to claim 1, characterized in that, The prediction model includes a signal encoding module and a signal decoding module, which are configured by a BiLSTM model with individually set parameters. The signal encoding module performs the following processing steps during operation: Feature extraction is performed on the signal of each channel of the uterine vector electromyography signal input to the model, and the corresponding initial one-dimensional temporal features are output. Each of the initial one-dimensional temporal features is vertically concatenated to form a two-dimensional feature matrix; A multi-layer convolutional neural network is used for feature transformation and dimension conversion to output a one-dimensional temporal feature of the target. The signal decoding module includes the following processing steps during operation: The preterm birth situation is predicted by performing a one-dimensional temporal feature analysis on the target. Specifically, the preterm birth situation includes three types, namely: Type 1, premature birth; The second type is full-term and normal delivery; The third type is full-term labor, but induced labor or cesarean section.

4. The method according to claim 1, characterized in that, The method further includes: Obtain the vector electromyographic signal of the sample uterus corresponding to the preset electrode arrangement scheme; Configure corresponding annotations for the uterine vector electromyography signals of the sample; The prediction model is trained using labeled sample uterine vector electromyography signals corresponding to the preset electrode arrangement.

5. The method according to claim 4, characterized in that, The step of acquiring the sample uterine vector electromyography signal corresponding to the preset electrode arrangement includes: Acquire initial sample uterine vector electromyography (EMG) signals from M cases with complete pregnancy outcomes corresponding to the preset electrode arrangement scheme, wherein the initial sample uterine vector EMG signal pairs are for no less than 30 minutes, the signal sampling rate is no less than 360 Hz, and M is greater than 200; The initial sample uterine vector electromyography signal is sliced ​​into segments of 1 to 3 seconds in length to form N groups of signal segments, where N is greater than 600. Each signal segment contains 9 sequence data of length S, where S is the standard sequence length of the model input and S is greater than 360 sampling points. The N signal segments from each of the M cases are merged to form M×N signal segments, which are then randomly shuffled to form corresponding batches of sequence data, with batch sizes ranging from 16 to 128 groups.

6. The method according to claim 4, characterized in that, The loss function used during training is the cross-entropy function, and the labeled data is encoded using one-hot encoding to form corresponding vectors for use in the calculation of the cross-entropy function.

7. A device for predicting preterm birth based on uterine vector electromyography signals, characterized in that, The device includes: The acquisition unit is used to acquire uterine vector electromyography (EMG) signals collected by the mother under a preset electrode arrangement scheme. The preset electrode arrangement scheme involves 6 electrodes, specifically configured as pairs of electrodes arranged orthogonally. The uterine vector EMG signals specifically include uterine EMG signals in three orthogonal directions corresponding to the three electrode pairs. The uterine vector EMG signals also include differential signals obtained by differentiating the two signals of each electrode pair. The prediction unit is used to input the uterine vector electromyography signal into a pre-configured prediction model. Specifically, the prediction model is trained by annotated sample uterine vector electromyography signals corresponding to the preset electrode arrangement scheme. The prediction model is used to predict the corresponding preterm birth situation based on the uterine vector electromyography signal input into the model. An extraction unit is used to extract the prediction results of preterm birth output by the prediction model; For the three electrode pairs, we have: The first electrode pair consists of a first electrode and a fourth electrode. The first electrode is positioned slightly above the center of the mother's abdominal wall, and the fourth electrode is positioned slightly below the center of the mother's lower back. The second electrode pair consists of a second electrode and a third electrode. The second electrode is positioned slightly above the center of the mother's back, and the third electrode is positioned slightly below the center of the mother's abdominal wall. The line connecting the first electrode pair and the line connecting the second electrode pair are perpendicular to each other. The third electrode pair consists of the fifth and sixth electrodes, which are located along the horizontal line of the mother's waist and on the left and right sides of the abdominal wall, respectively. The line connecting the third electrode pair is orthogonal to the line connecting the first electrode pair and the line connecting the second electrode pair.

8. A processing apparatus, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the method as described in any one of claims 1 to 6 when it invokes the computer program in the memory.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted for loading by a processor to perform the method of any one of claims 1 to 6.

Images