Fetal heart analysis method, device and equipment and storage medium

Abstract

The invention discloses a fetal heart analysis method, device and equipment and a storage medium. The method comprises the steps that fetal heart data and user basic information uploaded by a user are acquired; the fetal heart data and the basic information of the user are input into a preset fetal heart analysis model, a fetal heart analysis result is obtained, and the preset fetal heart analysis model comprises a position coding module, a bidirectional gating circulation unit, a channel attention mechanism, a multi-scale feature extraction module and a feature fusion module; and distributing the fetal heart data, the basic information of the user and the fetal heart analysis result to doctors of different grades according to the fetal heart analysis result. The fetal heart data and the basic information of the user are input into the preset fetal heart analysis model to obtain the fetal heart analysis result, and the preset fetal heart analysis model comprises the position coding module, the bidirectional gating circulation unit, the channel attention mechanism, the multi-scale feature extraction module and the feature fusion module. According to the invention, the accuracy of fetal heart analysis is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a fetal heart analysis method, device, equipment and storage medium. Background Art

[0002] Fetal monitoring is one of the important means to ensure the health of the fetus. By monitoring indicators such as the fetal heart rate and fetal movement, fetal abnormalities can be detected in time, and corresponding measures can be taken to ensure the healthy growth of the fetus. At present, fetal monitoring is mainly carried out in hospitals, requiring the use of professional monitoring equipment and medical personnel for monitoring. However, medical resources are tight, and the number of high-risk pregnant women has greatly increased. High-risk pregnant women need to monitor their fetal heart rate every day in the late pregnancy, so going to the hospital has become very inconvenient.

[0003] Therefore, remote fetal heart monitoring has become a good solution. Remote fetal heart monitoring can improve the quality of life of pregnant women. Pregnant women can monitor their fetal heart rate at home without having to go to the hospital frequently, thus reducing the burden and time cost of pregnant women. At the same time, remote fetal heart monitoring can also improve the sense of security of pregnant women, allowing them to spend their pregnancy more assured. Remote fetal heart monitoring can be achieved through smart devices. Pregnant women only need to wear a fetal heart monitor to monitor the heart rate and fetal movement of the fetus in real time. At the same time, the monitoring data can be transmitted back to the cloud via the Internet, and the cloud will be distributed to doctors, who will perform remote analysis and feedback to pregnant women. The benefits of remote fetal heart monitoring are not only convenient and fast, but also can effectively alleviate the problem of tight medical resources, balance the doctor resources in different regions and hospitals, and improve medical efficiency and quality. However, the existing fetal heart analysis models have many shortcomings, including insufficient data sets and poor model effects.

[0004] Therefore, there is an urgent need for a fetal heart analysis method that can improve the accuracy of fetal heart analysis and thereby solve the technical problems of shortage of hospital resources and low medical efficiency in the existing fetal heart monitoring process. Summary of the invention

[0005] The main purpose of the present invention is to provide a fetal heart analysis method, device, equipment and storage medium, aiming to solve the technical problem of how to improve the accuracy of fetal heart analysis.

[0006] To achieve the above object, the present invention provides a fetal heart rate analysis method, which comprises the following steps:

[0007] Obtain fetal heart data and basic user information uploaded by the user;

[0008] Inputting the fetal heart data and the basic information of the user into a preset fetal heart analysis model to obtain a fetal heart analysis result, wherein the preset fetal heart analysis model includes a position encoding module, a bidirectional gated recurrent unit, a channel attention mechanism, a multi-scale feature extraction module and a feature fusion module;

[0009] The fetal heart data, the user basic information and the fetal heart analysis result are distributed to doctors of different levels according to the fetal heart analysis result.

[0010] Optionally, before the step of obtaining the fetal heart data and basic user information uploaded by the user, the step further includes:

[0011] Acquire an original fetal heart data set, and perform data preprocessing on the original fetal heart data set to obtain a model training data set;

[0012] Constructing a one-dimensional convolutional neural network model, and using the one-dimensional convolutional neural network model as an initial fetal heart analysis model;

[0013] Improving the initial fetal heart analysis model to obtain a fetal heart analysis model to be trained;

[0014] The model training data set is input into the fetal heart analysis model to be trained for model training to obtain a preset fetal heart analysis model.

[0015] Optionally, the step of improving the initial fetal heart analysis model to obtain a fetal heart analysis model to be trained includes:

[0016] Improving the initial fetal heart analysis model into a deep residual model to obtain a first fetal heart analysis model;

[0017] Introducing a bidirectional gated cyclic unit into the first fetal heart analysis model to obtain a second fetal heart analysis model;

[0018] Introducing a multi-scale feature extraction module and a feature fusion module into the second fetal heart analysis model to obtain a third fetal heart analysis model;

[0019] Introducing a position coding module into the third fetal heart analysis model to obtain a fourth fetal heart analysis model;

[0020] A channel attention mechanism is introduced into the fourth fetal heart analysis model to obtain a fetal heart analysis model to be trained.

[0021] Optionally, the step of obtaining an original fetal heart data set and performing data preprocessing on the original fetal heart data set to obtain a model training data set includes:

[0022] Acquire an original fetal heart data set, and classify the original fetal heart data set to obtain a classification result;

[0023] Annotating the original fetal heart data set based on the classification result to obtain a labeled fetal heart data set;

[0024] Cleaning the labeled fetal heart data set according to a fetal heart data screening standard to obtain a cleaned fetal heart data set;

[0025] The cleaned fetal heart data set is subjected to moving average smoothing processing to obtain a model training data set.

[0026] Optionally, the step of performing moving average smoothing processing on the cleaned fetal heart data set to obtain a model training data set includes:

[0027] Performing moving average smoothing processing on the cleaned fetal heart data set according to a preset time window to obtain a processed fetal heart data set;

[0028] Performing data enhancement processing on the processed fetal heart data set to obtain an enhanced fetal heart data set;

[0029] The enhanced fetal heart data set is adjusted according to the number of each fetal heart data category in the enhanced fetal heart data set to obtain a model training data set.

[0030] Optionally, the step of distributing the fetal heart data, the user basic information and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result includes:

[0031] If the fetal heart analysis result is normal, the fetal heart data, the user basic information and the fetal heart analysis result are sent to a doctor for fetal heart analysis review;

[0032] If the fetal heart analysis result is abnormal, the fetal heart data, the user basic information and the fetal heart analysis result are distributed to two doctors at the level of associate chief physician or above for fetal heart analysis review.

[0033] Optionally, after the step of inputting the model training data set into the fetal heart analysis model to be trained for model training to obtain a preset fetal heart analysis model, the step further includes:

[0034] The preset fetal heart analysis model is evaluated using model evaluation indicators, and the model evaluation indicators include: accuracy, precision, recall rate and F1 score.

[0035] In addition, to achieve the above-mentioned purpose, the present invention also provides a fetal heart rate analysis device, the device comprising:

[0036] Data acquisition module, used to obtain fetal heart data and basic user information uploaded by the user;

[0037] A data analysis module, used to input the fetal heart data and the user's basic information into a preset fetal heart analysis model to obtain a fetal heart analysis result, wherein the preset fetal heart analysis model includes a position encoding module, a bidirectional gated recurrent unit, a channel attention mechanism, a multi-scale feature extraction module, and a feature fusion module;

[0038] A data distribution module is used to distribute the fetal heart data, the user basic information and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result.

[0039] In addition, to achieve the above-mentioned purpose, the present invention also proposes a fetal heart analysis device, which includes: a memory, a processor, and a fetal heart analysis program stored in the memory and executable on the processor, wherein the fetal heart analysis program is configured to implement the steps of the fetal heart analysis method described above.

[0040] In addition, to achieve the above-mentioned purpose, the present invention also proposes a storage medium, on which a fetal heart analysis program is stored, and when the fetal heart analysis program is executed by a processor, the steps of the fetal heart analysis method described above are implemented.

[0041] The present invention discloses the method of obtaining fetal heart data and basic user information uploaded by a user; inputting the fetal heart data and the basic user information into a preset fetal heart analysis model to obtain a fetal heart analysis result, wherein the preset fetal heart analysis model includes a position encoding module, a bidirectional gated cyclic unit, a channel attention mechanism, a multi-scale feature extraction module, and a feature fusion module; and distributing the fetal heart data, the basic user information, and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result. Since the fetal heart data and the basic user information are input into a preset fetal heart analysis model to obtain a fetal heart analysis result, and the preset fetal heart analysis model includes a position encoding module, a bidirectional gated cyclic unit, a channel attention mechanism, a multi-scale feature extraction module, and a feature fusion module in the present invention, compared with the prior art, the present invention improves the accuracy of fetal heart analysis, thereby solving the technical problems of shortage of hospital resources and low medical efficiency in the existing fetal heart monitoring process. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flowchart of the first embodiment of the fetal heart rate analysis method of the present invention;

[0043] Figure 2 A schematic diagram of a model architecture of a preset fetal heart analysis model in the fetal heart analysis method of the present invention;

[0044] Figure 3 It is a flowchart of a second embodiment of the fetal heart rate analysis method of the present invention;

[0045] Figure 4Schematic diagram of a single residual block structure constituting a depth residual model in the fetal heart analysis method of the present invention;

[0046] Figure 5 Schematic diagram of the structure of two feature extraction modules of different scales in the fetal heart analysis method of the present invention;

[0047] Figure 6 It is a schematic diagram of the channel attention structure in the fetal heart analysis method of the present invention;

[0048] Figure 7 It is a flowchart of a third embodiment of a fetal heart rate analysis method of the present invention;

[0049] Figure 8 is a structural block diagram of a first embodiment of a fetal heart rate analysis device of the present invention;

[0050] Fig. 9 It is a structural schematic diagram of a fetal heart analysis device in a hardware operating environment involved in an embodiment of the present invention.

[0051] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0052] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0053] The embodiment of the present invention provides a fetal heart rate analysis method, referring to Figure 1 , Figure 1 FIG. 1 is a flow chart of a first embodiment of a fetal heart rate analysis method according to the present invention.

[0054] In this embodiment, the fetal heart rate analysis method includes steps S10 to S30:

[0055] Step S10: Obtain the fetal heart data and basic user information uploaded by the user.

[0056] It should be noted that the execution subject of this embodiment can be a computer server device with data processing, network communication and program running functions used in the fetal heart analysis scenario, such as a personal computer, a laptop computer, a mobile phone, a smart watch, etc., or an electronic device capable of realizing the above functions, a fetal heart analysis device, etc. The following takes a system including a fetal heart analysis device (hereinafter referred to as the system) as an example to illustrate this embodiment and the following embodiments.

[0057] It should be understood that fetal heart data is an important indicator reflecting the status of the fetus in the womb, mainly including fetal heart rate and its changes. User basic information may include the user's age, month of pregnancy, contact information and home address.

[0058] It should be added that the above system can provide users with a convenient and accurate remote fetal heart analysis platform through the close integration of the front-end and the back-end. The front-end can use the Vue3+Element-Plus framework to ensure the simplicity of the user interface and the completeness of the functions. The back-end can use the Django 2 framework, combined with the preset fetal heart analysis model, to achieve real-time analysis and result feedback of fetal heart monitoring data. The platform is managed by roles, and different user roles have different transaction processing processes, so that users of different roles can efficiently upload, review and communicate data on the system, ensuring that pregnant women can get timely feedback on the health status of the fetus and reduce the burden of high-risk pregnant women frequently going back and forth to the hospital.

[0059] It is understandable that the system can provide login and registration functions for three identities: pregnant women (patient), doctors (doctor), and administrators (admin). When pregnant women log in to the system, they upload fetal heart data and basic user information through the system interface; pregnant women can also view all their uploaded records and fetal heart analysis results. The uploaded records can also include the upload date, uploaded data, doctor's review results, etc.; in addition, pregnant women can also communicate with the assigned doctors to resolve doubts. When the doctor logs in to the system, the doctor can view the assigned list of data to be reviewed (the list of data to be reviewed includes fetal heart data and basic user information). The list of data to be reviewed can be sorted by normal / abnormal fetal heart rate and submission time; after the doctor reviews the data, he fills in the feedback and submits it. In addition, the system also supports visualization of fetal heart data; if the review result is abnormal, the doctor can promptly remind the patient to go to the doctor offline; if there is an abnormality and it is impossible to contact the pregnant woman in time, the doctor can directly communicate with the administrator and let the administrator follow up. When the administrator logs into the system, the administrator can view all fetal heart analysis data. If the doctor's review result is abnormal, the administrator can directly contact the pregnant woman by phone to remind her to go to the hospital for treatment in time. In addition, for data judged as abnormal by the preset fetal heart analysis model, the administrator can directly remind the doctor to review it as soon as possible.

[0060] Step S20: Input the fetal heart data and the user basic information into a preset fetal heart analysis model to obtain a fetal heart analysis result. The preset fetal heart analysis model includes a position encoding module, a bidirectional gated recurrent unit, a channel attention mechanism, a multi-scale feature extraction module and a feature fusion module.

[0061] It should be noted that in the research of the model, the preset fetal heart analysis model was designed by referring to relevant literature and based on the standard method used by doctors to judge the health status of the fetus using fetal heart rate (FHR). Doctors mainly focus on the following aspects when evaluating the health status of the fetus:

[0062] (1) Overall baseline fetal heart rate: Is it within the normal range, i.e. 110-160 beats per minute (bpm)?

[0063] (2) Fetal heart rate baseline variation: Is it moderate? Is its small fluctuation range between 6-25 bpm?

[0064] (3) Fetal heart rate acceleration: quality and frequency of acceleration.

[0065] (4) Fetal heart rate deceleration: whether it exists and its characteristics.

[0066] Based on the above considerations, the model was designed accordingly. For example, Figure 2 , Figure 2 The schematic diagram of the model architecture of the preset fetal heart analysis model in the fetal heart analysis method of the present invention. In the figure, Input represents input data; Scale-1featureextractor and Scale-2feature extractor represent two feature extractors of different scales, i.e., multi-scale feature extraction modules; Positional Encoding represents position encoding module; BiGRU represents bidirectional gated recurrent unit; Layer Norm represents layer normalization; Linear represents fully connected layer; Fusion represents feature fusion module; Output represents output data. In addition, a channel attention mechanism is introduced in the preset fetal heart analysis model to enhance the model's ability to capture key information.

[0067] It should be noted that, considering that the fetal heart data is a time series, its front-to-back temporal position relationship may have an important impact on the model performance. Therefore, the position encoding module and the bidirectional gated recurrent unit (BiGRU) are introduced to process this temporal information.

[0068] It should be noted that Scale-1feature extractor and Scale-2feature extractor are feature extractors of different scales, namely the multi-scale feature extraction module, which is designed to capture the overall baseline with strong integrity (for example, 20-minute scale), medium-scale acceleration and deceleration (for example, 30-second scale), and then to smaller-scale baseline variations (for example, 1-second scale) to extract features of different scales.

[0069] Step S30: Distribute the fetal heart data, the user basic information and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result.

[0070] It should be understood that the above fetal heart analysis results include both normal and abnormal situations.

[0071] In a specific implementation, if the fetal heart analysis result is normal, the fetal heart data, the user's basic information and the fetal heart analysis result will be sent to a doctor for fetal heart analysis review, and priority will be given to doctors who are of concern to pregnant women; if the fetal heart analysis result is abnormal, the fetal heart data, the user's basic information and the fetal heart analysis result will be distributed to two doctors at the level of associate chief physician or above for fetal heart analysis review, and priority will be given to doctors who are of concern to pregnant women.

[0072] This embodiment discloses obtaining the fetal heart data and basic information uploaded by the user; inputting the fetal heart data and the basic information of the user into a preset fetal heart analysis model to obtain the fetal heart analysis result, wherein the preset fetal heart analysis model includes a position encoding module, a bidirectional gated cyclic unit, a channel attention mechanism, a multi-scale feature extraction module, and a feature fusion module; and distributing the fetal heart data, the basic information of the user, and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result. Since in this embodiment, the fetal heart data and the basic information of the user are input into a preset fetal heart analysis model to obtain the fetal heart analysis result, and the preset fetal heart analysis model includes a position encoding module, a bidirectional gated cyclic unit, a channel attention mechanism, a multi-scale feature extraction module, and a feature fusion module, compared with the prior art, this embodiment improves the accuracy of fetal heart analysis, thereby solving the technical problems of shortage of hospital resources and low medical efficiency in the existing fetal heart monitoring process.

[0073] refer to Figure 3 , Figure 3 FIG. 4 is a flow chart of a second embodiment of a fetal heart rate analysis method according to the present invention.

[0074] Based on the first embodiment described above, in this embodiment, before step S10, steps S01 to S04 are also included:

[0075] Step S01: obtaining an original fetal heart data set, and performing data preprocessing on the original fetal heart data set to obtain a model training data set.

[0076] It should be understood that the above data preprocessing may include labeling the original fetal heart data set and then performing data cleaning on the labeled original fetal heart data set.

[0077] Step S02: construct a one-dimensional convolutional neural network model, and use the one-dimensional convolutional neural network model as an initial fetal heart analysis model.

[0078] Step S03: improving the initial fetal heart analysis model to obtain a fetal heart analysis model to be trained.

[0079] In a specific implementation, in order to improve the accuracy of fetal heart analysis, step S03 includes steps S031 to S035:

[0080] Step S031: improving the initial fetal heart analysis model into a depth residual model to obtain a first fetal heart analysis model.

[0081] Step S032: introducing a bidirectional gated cyclic unit into the first fetal heart analysis model to obtain a second fetal heart analysis model.

[0082] Step S033: introducing a multi-scale feature extraction module and a feature fusion module into the second fetal heart analysis model to obtain a third fetal heart analysis model.

[0083] Step S034: introducing a position coding module into the third fetal heart analysis model to obtain a fourth fetal heart analysis model.

[0084] Step S035: Introduce a channel attention mechanism into the fourth fetal heart analysis model to obtain a fetal heart analysis model to be trained.

[0085] It should be noted that in order to increase the depth of the model and provide better classification results when processing longer sequence data, the initial fetal heart analysis model (i.e., one-dimensional convolutional neural network model) is improved to a deep residual model. Figure 4 , Figure 4 The schematic diagram of the structure of a single residual block constituting the deep residual model in the fetal heart analysis method of the present invention is shown in the figure. In the figure, ResBlock represents the residual block; x represents the input; Conv1D represents the 1D convolution; Batch Norm represents the batch normalization; ReLU represents the activation function; Projection represents the projection; Shortcut represents directly passing the projected value to the output end without passing through the intermediate processing part, which is used to establish the residual connection.

[0086] In order to capture the overall baseline with strong integrity (e.g., 20-minute scale), the acceleration and deceleration at medium scale (e.g., 30-second scale), and the baseline variation at smaller scale (e.g., 1-second scale), that is, to extract features of different scales, a multi-scale feature extraction module and a feature fusion module are introduced. For example, refer to Figure 5 , Figure 5 The schematic diagram is a structural diagram of two feature extraction modules of different scales in the fetal heart analysis method of the present invention. In the figure, Scale-1 and Scale-2 are feature extraction structures of two different scales; ResBlock represents the residual block; AvgPool represents the average pooling layer; Channel Attention represents the channel attention.

[0087] Furthermore, considering that fetal heart data is a time series, its antecedent temporal position relationship may have an important impact on the model performance. Therefore, a position encoding module and a bidirectional gated recurrent unit (BiGRU) are introduced to process this temporal information.

[0088] It should be noted that after multiple convolutions, the number of channels of the model can reach 512. In order to manage this redundant information and enhance the model's ability to capture key information, a channel attention mechanism is introduced. For example, Figure 6 , Figure 6 Schematic diagram of the channel attention structure in the fetal heart analysis method of the present invention. In the figure, Channel Attention represents channel attention; Input feature represents input feature; MaxPool represents maximum pooling layer; AvgPool represents average pooling layer; SharedMLP represents multi-layer shared perceptron.

[0089] Step S04: inputting the model training data set into the fetal heart analysis model to be trained to perform model training and obtain a preset fetal heart analysis model.

[0090] In a specific implementation, the model training data set can be input into the fetal heart analysis model to be trained for model training to obtain a training result; then the model parameters of the fetal heart analysis model to be trained are optimized according to the training result to obtain a preset fetal heart analysis model.

[0091] It is understandable that after the step of inputting the model training data set into the fetal heart analysis model to be trained for model training and obtaining the preset fetal heart analysis model, it also includes: using model evaluation indicators to evaluate the preset fetal heart analysis model, and the model evaluation indicators include: accuracy, precision, recall rate and F1 score.

[0092] In the specific implementation, the performance of each model can be compared, including accuracy, precision, recall and F1 score. For example, Table 1 shows the performance comparison of each model:

[0093] Table 1 Performance comparison of each model

[0094] Model Accuracy Accuracy Recall F1 score Initial fetal heart rate analysis model 86.34 86.51 85.12 85.81 First fetal heart rate analysis model 88.99 90.21 88.98 89.59 Second fetal heart rate analysis model 91.60 91.49 90.61 91.05 The third fetal heart analysis model 92.35 92.19 90.42 91.30 Fourth fetal heart analysis model 93.05 93.52 92.44 92.98 Preset fetal heart analysis model 95.67 94.58 96.93 95.74

[0095] The data in the performance comparison table of each model show the gradual improvement in performance of each model, proving the effectiveness of gradually improving the model architecture in this embodiment.

[0096] This embodiment discloses obtaining an original fetal heart data set, and performing data preprocessing on the original fetal heart data set to obtain a model training data set; constructing a one-dimensional convolutional neural network model, and using the one-dimensional convolutional neural network model as an initial fetal heart analysis model; improving the initial fetal heart analysis model to a deep residual model to obtain a first fetal heart analysis model; introducing a bidirectional gated recurrent unit into the first fetal heart analysis model to obtain a second fetal heart analysis model; introducing a multi-scale feature extraction module and a feature fusion module into the second fetal heart analysis model to obtain a third fetal heart analysis model; introducing a position encoding module into the third fetal heart analysis model to obtain a fourth fetal heart analysis model; introducing a channel attention mechanism into the fourth fetal heart analysis model to obtain a fetal heart analysis model to be trained; inputting the model training data set into the fetal heart analysis model to be trained for model training to obtain a preset fetal heart analysis model. Compared with the prior art, this embodiment effectively improves the performance of model analysis and further improves the accuracy of fetal heart analysis.

[0097] refer to Figure 7 , Figure 7 FIG. 4 is a flow chart of a third embodiment of a fetal heart rate analysis method according to the present invention.

[0098] Based on the above embodiments, in this embodiment, step S01 includes steps S011 to S014:

[0099] Step S011: obtaining an original fetal heart data set, and classifying the original fetal heart data set to obtain a classification result.

[0100] Step S012: labeling the original fetal heart data set based on the classification result to obtain a labeled fetal heart data set.

[0101] Step S013: Cleaning the labeled fetal heart data set according to the fetal heart data screening standard to obtain a cleaned fetal heart data set.

[0102] Step S014: performing moving average smoothing processing on the cleaned fetal heart data set to obtain a model training data set.

[0103] It should be understood that classifying the original fetal heart data set is to determine the positive (positive example) samples and negative (negative example) samples in the original fetal heart data set.

[0104] It should be explained that the above-mentioned fetal heart data screening criteria include: deleting all records with disconnection time exceeding 2 minutes; eliminating records with a total monitoring time of less than 20 minutes, except for disconnection and obvious data abnormalities (fetal heart rate lower than 50bpm or higher than 220bpm, usually caused by poor contact); and excluding data that doctors cannot give an accurate judgment on.

[0105] It should be noted that in order to further improve the data quality, step S014 also includes: performing moving average smoothing processing on the cleaned fetal heart data set according to a preset time window to obtain a processed fetal heart data set; performing data enhancement processing on the processed fetal heart data set to obtain an enhanced fetal heart data set; adjusting the enhanced fetal heart data set according to the number of each fetal heart data category in the enhanced fetal heart data set to obtain a model training data set.

[0106] It is understandable that the above-mentioned preset time window may be a custom setting by a system administrator, and this embodiment does not impose any limitation on this.

[0107] It should be noted that the above adjustment of the enhanced fetal heart data set according to the number of each fetal heart data category in the enhanced fetal heart data set is to make the number of each fetal heart data category in the fetal heart data set equal.

[0108] In the specific implementation, it is assumed that the original data set contains a total of 20484 records. After labeling, the number of positive (positive examples) samples is 2790, and the number of negative (negative examples) samples is 17694. After data cleaning of the labeled fetal heart data set, the number of positive and negative samples remaining in the cleaned fetal heart data set is reduced to 2342 and 13556, respectively. In order to further improve the data quality, the cleaned fetal heart data set is subjected to moving average smoothing, and the window size used is 10 (2.5 seconds). Then, each record is cut into a segment of length 4800 (20 minutes), of which the overlapping part is 1200, so as to perform data enhancement. The cropped data maintains the same label as the original data. After completing this step, the number of positive and negative samples is adjusted to 5400 and 19654, respectively. In order to maintain the balance between samples, 5400 data are randomly selected from the negative samples. Finally, a dataset consisting of a total of 10,800 data (i.e., the model training dataset), which contained an equal number of positive and negative samples, was used for subsequent analysis.

[0109] This embodiment discloses obtaining an original fetal heart data set, and classifying the original fetal heart data set to obtain a classification result; annotating the original fetal heart data set based on the classification result to obtain an annotated fetal heart data set; performing data cleaning on the annotated fetal heart data set according to a fetal heart data screening standard to obtain a cleaned fetal heart data set; performing moving average smoothing on the cleaned fetal heart data set according to a preset time window to obtain a processed fetal heart data set; performing data enhancement processing on the processed fetal heart data set to obtain an enhanced fetal heart data set; adjusting the enhanced fetal heart data set according to the number of each fetal heart data category in the enhanced fetal heart data set to obtain a model training data set. Compared with the prior art, this embodiment improves data quality and ensures the validity of the fetal heart analysis results generated by the preset fetal heart analysis model.

[0110] In addition, an embodiment of the present invention further provides a storage medium, on which a fetal heart analysis program is stored. When the fetal heart analysis program is executed by a processor, the steps of the fetal heart analysis method described above are implemented.

[0111] Reference Figure 8 , Figure 8 FIG. 1 is a structural block diagram of the first embodiment of the fetal heart rate analysis device of the present invention.

[0112] like Figure 8 As shown, the fetal heart analysis device proposed in the embodiment of the present invention includes: a data acquisition module 801, a data analysis module 802 and a data distribution module 803.

[0113] The data acquisition module 801 is used to acquire the fetal heart data and user basic information uploaded by the user.

[0114] The data analysis module 802 is used to input the fetal heart data and the user's basic information into a preset fetal heart analysis model to obtain a fetal heart analysis result. The preset fetal heart analysis model includes a position encoding module, a bidirectional gated recurrent unit, a channel attention mechanism, a multi-scale feature extraction module and a feature fusion module.

[0115] The data distribution module 803 is used to distribute the fetal heart data, the user basic information and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result.

[0116] The data distribution module 803 is also used to send the fetal heart data, the user's basic information and the fetal heart analysis result to a doctor for fetal heart analysis review if the fetal heart analysis result is normal; if the fetal heart analysis result is abnormal, the fetal heart data, the user's basic information and the fetal heart analysis result are distributed to two doctors at the level of associate chief physician or above for fetal heart analysis review.

[0117] The embodiment of the present device discloses obtaining the fetal heart data and basic information of the user uploaded by the user; inputting the fetal heart data and the basic information of the user into a preset fetal heart analysis model to obtain the fetal heart analysis result, wherein the preset fetal heart analysis model includes a position encoding module, a bidirectional gated cyclic unit, a channel attention mechanism, a multi-scale feature extraction module and a feature fusion module; and distributing the fetal heart data, the basic information of the user and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result. Since the fetal heart data and the basic information of the user are input into the preset fetal heart analysis model to obtain the fetal heart analysis result, and the preset fetal heart analysis model includes a position encoding module, a bidirectional gated cyclic unit, a channel attention mechanism, a multi-scale feature extraction module and a feature fusion module in the embodiment of the present device, compared with the prior art, the embodiment of the present device improves the accuracy of fetal heart analysis, thereby solving the technical problems of shortage of hospital resources and low medical efficiency in the existing fetal heart monitoring process.

[0118] Based on the above-mentioned first embodiment of the fetal heart rate analysis device of the present invention, a second embodiment of the fetal heart rate analysis device of the present invention is proposed.

[0119] In this embodiment, the data acquisition module 801 is also used to obtain an original fetal heart data set, and perform data preprocessing on the original fetal heart data set to obtain a model training data set; construct a one-dimensional convolutional neural network model, and use the one-dimensional convolutional neural network model as an initial fetal heart analysis model; improve the initial fetal heart analysis model to obtain a fetal heart analysis model to be trained; input the model training data set into the fetal heart analysis model to be trained for model training, and obtain a preset fetal heart analysis model.

[0120] The data acquisition module 801 is also used to improve the initial fetal heart analysis model into a deep residual model to obtain a first fetal heart analysis model; introduce a bidirectional gated recurrent unit into the first fetal heart analysis model to obtain a second fetal heart analysis model; introduce a multi-scale feature extraction module and a feature fusion module into the second fetal heart analysis model to obtain a third fetal heart analysis model; introduce a position encoding module into the third fetal heart analysis model to obtain a fourth fetal heart analysis model; introduce a channel attention mechanism into the fourth fetal heart analysis model to obtain a fetal heart analysis model to be trained.

[0121] The data acquisition module 801 is further used to perform model evaluation on the preset fetal heart analysis model using model evaluation indicators, and the model evaluation indicators include: accuracy, precision, recall rate and F1 score.

[0122] Other embodiments or specific implementations of the fetal heart rate analysis device of the present invention may refer to the above-mentioned method embodiments, which will not be described in detail here.

[0123] The present application provides a fetal heart analysis device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the fetal heart analysis method in the above-mentioned embodiment one.

[0124] Reference below Fig. 9 , which shows a schematic diagram of the structure of a fetal heart analysis device suitable for implementing the embodiment of the present application. The fetal heart analysis device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig. 9 The fetal heart analysis device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0125] like Fig. 9 As shown, the fetal heart analysis device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the fetal heart analysis device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the fetal heart analysis device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a fetal heart analysis device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0126] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0127] The fetal heart analysis device provided by the present application adopts the fetal heart analysis method in the above embodiment, which can solve the technical problem of how to improve the accuracy of fetal heart analysis. Compared with the prior art, the beneficial effects of the fetal heart analysis device provided by the present application are the same as the beneficial effects of the fetal heart analysis method provided by the above embodiment, and the other technical features in the fetal heart analysis device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0128] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0129] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0130] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0131] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0132] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0133] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A fetal heart rate analysis method, characterized in that: The method comprises: Obtain fetal heart data and basic user information uploaded by the user; Inputting the fetal heart data and the basic information of the user into a preset fetal heart analysis model to obtain a fetal heart analysis result, wherein the preset fetal heart analysis model includes a position encoding module, a bidirectional gated recurrent unit, a channel attention mechanism, a multi-scale feature extraction module and a feature fusion module; The fetal heart data, the user basic information and the fetal heart analysis result are distributed to doctors of different levels according to the fetal heart analysis result.

2. The fetal heart rate analysis method according to claim 1, characterized in that: Before the step of obtaining the fetal heart data and basic user information uploaded by the user, the method further includes: Acquire an original fetal heart data set, and perform data preprocessing on the original fetal heart data set to obtain a model training data set; Constructing a one-dimensional convolutional neural network model, and using the one-dimensional convolutional neural network model as an initial fetal heart analysis model; Improving the initial fetal heart analysis model to obtain a fetal heart analysis model to be trained; The model training data set is input into the fetal heart analysis model to be trained for model training to obtain a preset fetal heart analysis model.

3. The fetal heart rate analysis method according to claim 2, characterized in that: The step of improving the initial fetal heart analysis model to obtain the fetal heart analysis model to be trained includes: Improving the initial fetal heart analysis model into a deep residual model to obtain a first fetal heart analysis model; Introducing a bidirectional gated cyclic unit into the first fetal heart analysis model to obtain a second fetal heart analysis model; Introducing a multi-scale feature extraction module and a feature fusion module into the second fetal heart analysis model to obtain a third fetal heart analysis model; Introducing a position coding module into the third fetal heart analysis model to obtain a fourth fetal heart analysis model; A channel attention mechanism is introduced into the fourth fetal heart analysis model to obtain a fetal heart analysis model to be trained.

4. The fetal heart rate analysis method according to claim 2, characterized in that: The step of obtaining an original fetal heart data set and performing data preprocessing on the original fetal heart data set to obtain a model training data set includes: Acquire an original fetal heart data set, and classify the original fetal heart data set to obtain a classification result; Annotating the original fetal heart data set based on the classification result to obtain a labeled fetal heart data set; Cleaning the labeled fetal heart data set according to a fetal heart data screening standard to obtain a cleaned fetal heart data set; The cleaned fetal heart data set is subjected to moving average smoothing processing to obtain a model training data set.

5. The fetal heart rate analysis method according to claim 4, characterized in that: The step of performing moving average smoothing processing on the cleaned fetal heart data set to obtain a model training data set includes: Performing moving average smoothing processing on the cleaned fetal heart data set according to a preset time window to obtain a processed fetal heart data set; Performing data enhancement processing on the processed fetal heart data set to obtain an enhanced fetal heart data set; The enhanced fetal heart data set is adjusted according to the number of each fetal heart data category in the enhanced fetal heart data set to obtain a model training data set.

6. The fetal heart rate analysis method according to claim 1, characterized in that: The step of distributing the fetal heart data, the user basic information and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result comprises: If the fetal heart analysis result is normal, the fetal heart data, the user basic information and the fetal heart analysis result are sent to a doctor for fetal heart analysis review; If the fetal heart analysis result is abnormal, the fetal heart data, the user basic information and the fetal heart analysis result are distributed to two doctors at the level of associate chief physician or above for fetal heart analysis review.

7. The fetal heart rate analysis method according to claim 2, characterized in that: After the step of inputting the model training data set into the fetal heart analysis model to be trained for model training to obtain a preset fetal heart analysis model, the method further includes: The preset fetal heart analysis model is evaluated using model evaluation indicators, and the model evaluation indicators include: accuracy, precision, recall rate and F1 score.

8. A fetal heart rate analysis device, characterized in that: The device comprises: Data acquisition module, used to obtain fetal heart data and basic user information uploaded by the user; A data analysis module, used to input the fetal heart data and the user's basic information into a preset fetal heart analysis model to obtain a fetal heart analysis result, wherein the preset fetal heart analysis model includes a position encoding module, a bidirectional gated recurrent unit, a channel attention mechanism, a multi-scale feature extraction module, and a feature fusion module; A data distribution module is used to distribute the fetal heart data, the user basic information and the fetal heart analysis result to doctors of different levels according to the fetal heart analysis result.

9. A fetal heart analysis device, characterized in that: The device comprises: a memory, a processor, and a fetal heart analysis program stored in the memory and executable on the processor, wherein the fetal heart analysis program is configured to implement the steps of the fetal heart analysis method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a fetal heart analysis program, which, when executed by a processor, implements the steps of the fetal heart analysis method according to any one of claims 1 to 7.

Images