Classification method and device based on superovulation scheme, electronic equipment and medium

Abstract

Embodiments of the present application disclose a classification method and device based on a super ovulation scheme, an electronic device and a medium. The method comprises: obtaining initial feature data of a target object, performing data preprocessing on the initial feature data to obtain feature data of the target object; wherein the initial feature data comprises basic physiological characteristics and hormone characteristics of the target object; performing dimension mapping and pixel mapping on the feature data to obtain a feature image corresponding to the feature data; inputting the feature image into a preset classification model to obtain a classification result output by the classification model; wherein the classification result comprises any one of a super-long scheme, a long scheme, an antagonist or a short scheme in the super ovulation scheme. Through the technical scheme disclosed in the embodiments of the present application, the accuracy of the classification of the super ovulation scheme is improved, and effective suggestions are provided for doctors to select a super ovulation scheme.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of software technology, and in particular to a classification method, apparatus, electronic device and medium based on superovulation protocols. Background Technology

[0002] Infertility is a disease that seriously affects reproductive health. In recent years, due to factors such as the increase in the average age of childbearing, changes in lifestyle, environmental and social pressures, the global prevalence of infertility has increased from 11.0% in 1997 to 15.0% in 2017, and is expected to reach 17.2% in 2023. Under the current circumstances, in vitro fertilization (IVF) has become the choice for many infertile couples.

[0003] Currently, personalized COS (Career of Support) plans for patients are mostly formulated by doctors based on their experience, which requires a high level of expertise. However, given the current state of medical resources in China, there is an extreme imbalance between the number of doctors and patients, and even fewer doctors with extensive experience. This instability in the choice of plans for different patients can affect the success rate of in vitro fertilization (IVF). Summary of the Invention

[0004] This invention provides a classification method, device, electronic device, and medium based on superovulation protocols, thereby improving the accuracy of superovulation protocol classification and providing effective advice for doctors in selecting superovulation protocols.

[0005] In a first aspect, embodiments of the present invention provide a classification method based on superovulation protocols, the method comprising:

[0006] The initial feature data of the target object is obtained, and the initial feature data is preprocessed to obtain the feature data of the target object; wherein, the initial feature data includes the basic physiological characteristics and hormonal characteristics of the target object;

[0007] Perform dimension mapping and pixel mapping on the feature data to obtain the feature image corresponding to the feature data;

[0008] The feature image is input into a preset classification model to obtain the classification result output by the classification model; wherein, the classification result includes any one of the ultra-long protocol, long protocol, antagonist protocol or short protocol in the superovulation protocol.

[0009] Optionally, the data preprocessing of the feature data includes:

[0010] Obtain the feature value corresponding to each feature name in the initial feature data. If the feature value corresponding to the current feature name is empty, obtain the alternative feature name of the current feature name and obtain the alternative feature value corresponding to the alternative feature name.

[0011] Optionally, the data preprocessing of the initial feature data further includes:

[0012] Obtain the feature value type of each of the aforementioned feature values. If the feature value type of the current feature value is not numeric, then perform feature value type conversion on the current feature value.

[0013] Based on a preset numerical range, the values ​​of each feature are normalized to obtain the feature data of the target object.

[0014] Optionally, the dimensional mapping of the feature data includes:

[0015] The data dimension of the feature data is obtained. If the data dimension does not match the preset dimension, the feature data is mapped to the data dimension to obtain the feature data of the target dimension.

[0016] Optionally, the pixel mapping of the feature data includes:

[0017] Obtain a preset range of pixel data, and perform pixel data mapping on each feature value in the feature data of the target dimension based on the pixel data range to obtain the feature image corresponding to the feature data.

[0018] Optionally, before inputting the feature image into a preset classification model, the method further includes:

[0019] The feature image is subjected to image enhancement processing to achieve adaptive compensation.

[0020] Optionally, the classification model includes at least one convolutional block, at least one residual connected layer, and at least one fully connected layer; wherein,

[0021] The convolutional block is used to perform local perception on the feature image input to the classification model to obtain local features, and to perform convolution operation on the local features to obtain the global features of the feature image;

[0022] The residual layer is connected to the convolutional block of the last layer to solve the problem of gradient vanishing during transmission of each convolutional layer in the convolutional block.

[0023] The fully connected layer is used to output the classification result of the classification model.

[0024] Secondly, embodiments of the present invention also provide a classification device based on a superovulation protocol, the device comprising:

[0025] The feature data acquisition module is used to acquire initial feature data of the target object, perform data preprocessing on the initial feature data to obtain feature data of the target object; wherein, the initial feature data includes the basic physiological characteristics and hormonal characteristics of the target object;

[0026] The feature image acquisition module is used to perform dimension mapping and pixel mapping on the feature data to obtain the feature image corresponding to the feature data;

[0027] The classification result determination module is used to input the feature image into a preset classification model to obtain the classification result output by the classification model; wherein, the classification result includes any one of the ultra-long protocol, long protocol, antagonist protocol or short protocol in the superovulation protocol.

[0028] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising:

[0029] One or more processors;

[0030] Storage device for storing one or more programs.

[0031] When the one or more programs are executed by the one or more processors, the one or more processors implement the classification method based on superovulation protocols provided in any embodiment of the present invention.

[0032] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the classification method based on superovulation protocols provided in any embodiment of the present invention.

[0033] The technical solution provided in this embodiment specifically includes: acquiring initial feature data of the target object; performing data preprocessing on the initial feature data to obtain feature data of the target object; wherein the initial feature data includes the basic physiological and hormonal characteristics of the target object; acquiring multiple feature data of the target object to improve the accuracy of the classification results; performing dimensional mapping and pixel mapping on the feature data to obtain feature images corresponding to the feature data; mapping the feature data to feature images to obtain deeper feature information, thereby further improving the accuracy of the classification results; inputting the feature images into a preset classification model to obtain the classification results output by the classification model; wherein the classification results include any one of the ultra-long protocol, long protocol, antagonist protocol, or short protocol in the superovulation protocol; by acquiring multiple feature data of the target object, mapping the feature data to feature images to obtain deeper feature information, and obtaining the classification results corresponding to the classification data, the accuracy of the classification of the superovulation protocol is improved, providing effective suggestions for doctors to select a superovulation protocol. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of exemplary embodiments of the present invention, the accompanying drawings used in describing the embodiments are briefly introduced below. Obviously, the accompanying drawings described are only a portion of the drawings of the embodiments to be described in this invention, and not all of the drawings. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort.

[0035] Figure 1 This is a flowchart illustrating the classification method based on superovulation protocols provided in Embodiment 1 of the present invention;

[0036] Figure 2 This is a schematic diagram illustrating the result of data dimension transformation involved in Embodiment 1 of the present invention;

[0037] Figure 3 This is a schematic diagram of the classification model involved in Embodiment 1 of the present invention;

[0038] Figure 4 This is a flowchart illustrating the classification method based on superovulation protocols provided in Embodiment 2 of the present invention;

[0039] Figure 5 This is a schematic diagram of the classification device based on the superovulation protocol provided in Embodiment 3 of the present invention;

[0040] Figure 6 This is a schematic diagram of the structure of the electronic device provided in Embodiment 4 of the present invention. Detailed Implementation

[0041] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0042] Example 1

[0043] Figure 1 This is a flowchart illustrating a classification method based on superovulation protocols according to Embodiment 1 of the present invention. This embodiment is applicable to situations where the classification result of a superovulation protocol is determined based on user feature data. This method can be executed by a classification device based on a superovulation protocol, which can be implemented in software and / or hardware.

[0044] Before introducing the technical solutions of the embodiments of the present invention, the application scenarios of implementing the technical solutions of this embodiment will be described exemplarily. Of course, the following application scenarios are only optional; this embodiment can also be implemented in other application scenarios, and this embodiment does not limit the application scenarios of the implemented technical methods. Specifically, the application scenarios include: Infertility is a disease that seriously affects reproductive health. In recent years, due to factors such as the increase in the average age of childbearing, changes in lifestyle, environmental and social pressures, the global prevalence of infertility has increased rapidly. Currently, in vitro fertilization (IVF) has become the choice for many infertile couples.

[0045] However, judging from the current medical resources in China, there is an extreme imbalance between the number of doctors and patients, and even fewer doctors with rich experience. As a result, there is instability in the choice of treatment plan for different patients, which will affect the success rate of in vitro fertilization.

[0046] To address the aforementioned technical problems, the technical solution of this invention considers determining the specific superovulation protocol for a user based on the user's basic vital signs data and the vital signs data in the hormone test results. This information is provided for doctors' reference and helps with subsequent diagnosis, especially for doctors lacking clinical experience. This improves the accuracy of superovulation protocol classification and provides effective advice for doctors in selecting superovulation protocols.

[0047] Specifically, the technical solution provided in this embodiment includes: acquiring initial feature data of the target object; performing data preprocessing on the initial feature data to obtain feature data of the target object; wherein the initial feature data includes the basic physiological and hormonal characteristics of the target object; acquiring multiple feature data of the target object to improve the accuracy of the classification results; performing dimensional mapping and pixel mapping on the feature data to obtain feature images corresponding to the feature data; mapping the feature data to feature images to obtain deeper feature information, thereby further improving the accuracy of the classification results; inputting the feature images into a preset classification model to obtain the classification results output by the classification model; wherein the classification results include any one of the ultra-long protocol, long protocol, antagonist protocol, or short protocol in the superovulation protocol; by acquiring multiple feature data of the target object, mapping the feature data to feature images to obtain deeper feature information, and obtaining the classification results corresponding to the classification data, the accuracy of the superovulation protocol classification is improved, providing effective suggestions for doctors to select superovulation protocols.

[0048] like Figure 1 As shown, the method specifically includes the following steps:

[0049] S110. Obtain the initial feature data of the target object, perform data preprocessing on the initial feature data, and obtain the feature data of the target object.

[0050] In this embodiment of the invention, the target object can be a user who has undergone hormone testing and wants to try in vitro fertilization (IVF). The initial characteristic data of the target object includes initial basic characteristics and initial hormonal characteristics; specifically, hormonal characteristics include basal hormones and daily hormone levels. To determine a more suitable classification result for the target object, in the technical solution of this embodiment, the target object is classified based on its age and whether it is the first time the classification has been determined, and the initial characteristic data of different categories of target objects are also different.

[0051] For example, based on the target individuals' cycle and age, they are divided into a standard group and a non-standard group. Those in their first cycle and aged 35 or younger are classified as the standard group; the rest are in the non-standard group. Specifically, the non-standard group is further divided into first cycle and multiple cycles (not the first cycle). Because diagnosing multi-cycle protocols requires referencing treatment results from previous cycles, and because hormone levels differ for users of different ages, the characteristic data needed to determine the appropriate superovulation protocol for each group also differs. Correspondingly, the classification models used also differ; specifically, three classification models are used: a standard group model, a non-standard group one-cycle model, and a non-standard group multi-cycle model.

[0052] Furthermore, after determining the group class to which the target object belongs, the corresponding initial feature data for the target object is determined. In this embodiment, the target object belonging to the standard group is used as an example for illustrative explanation. For example, feature selection is first based on chi-square calculation of a single category, removing features with low scores, and finally selecting 28 initial feature data, such as: basic information, basic hormones, daily hormones, etc. Specifically, the initial feature data is shown in Table 1 below:

[0053] Table 1, Initial Feature Data of the Standard Group

[0054]

[0055] Specifically, the aforementioned 28 initial feature data are the results of feature selection based on the standard group and are used in the standard group model; for the non-standard group in one period, based on sample statistics, 30 features are selected, and for the non-standard group in multiple periods, 31 features are selected. Of course, the number and features selected for each group are only optional embodiments. The technical solution of this embodiment can also select other quantities and features as feature data according to the actual situation, and this embodiment does not limit this.

[0056] Furthermore, after obtaining the initial feature data of the target object, the data needs to be preprocessed to facilitate subsequent data mapping. In this embodiment, data preprocessing may include data processing methods such as null value handling, outlier handling, and normalization.

[0057] Optionally, for handling null values ​​in the initial feature data, the processing method of the technical solution in this embodiment may include: obtaining the feature value corresponding to each feature name in the initial feature data; if the feature value corresponding to the current feature name is null, then obtaining the alternative feature name of the current feature name and obtaining the alternative feature value corresponding to the alternative feature name.

[0058] For example, let's take the case where the characteristic value corresponding to the hormone on that day is empty: If the number of days gn (the number of days of the treatment plan) is 1 and the hormone on that day is empty, then the hormone value of the basal hormone is used to replace the hormone value of the hormone on that day; similarly, if the number of days gn (the number of days of the treatment plan) is not 1 and the hormone on that day is empty, then the hormone value of the day of consultation or ultrasound is used to replace the hormone value of the hormone on that day; optionally, for other characteristic values ​​that are empty, preset values ​​can also be used to replace them; of course, specific processing methods can be adopted for empty values ​​according to the actual situation. This embodiment does not impose specific limitations on the empty value processing method.

[0059] Optionally, for the preprocessing of the initial feature data, the processing method of the technical solution in this embodiment may further include: obtaining the feature value type of each feature value; if the feature value type of the current feature value is not numerical, then performing feature value type conversion on the current feature value; and normalizing the value of each feature value based on a preset numerical range to obtain the feature data of the target object.

[0060] Specifically, in the initial feature data, most of the feature values ​​corresponding to each feature name are numerical feature types, with some being binary or discrete feature types. To facilitate subsequent feature data mapping, the technical solution in this embodiment uniformly converts the feature types of each feature value to numerical types. However, the value ranges of each feature after type conversion vary considerably; therefore, normalization is performed on all features. Specifically, the following expression can be used for normalization:

[0061] ;

[0062] Where x represents the current feature value, Min represents the minimum value of the feature, and Max represents the maximum value of the feature; the range of the new feature value is [0,1).

[0063] S120. Perform dimension mapping and pixel mapping on the feature data to obtain the feature image corresponding to the feature data.

[0064] In this embodiment of the invention, the feature data obtained based on the technical solution in the above embodiment has a short vector length and is a one-dimensional column vector. In the subsequent classification process using the model, the classification effect obtained by the one-dimensional feature vector is not ideal. Therefore, in order to improve the classification effect of the model, the technical solution of this embodiment transforms the current one-dimensional feature vector into a two-dimensional vector, and obtains the classification result output by the two-dimensional convolutional neural network classification model based on the two-dimensional vector, thereby improving the accuracy of the classification result.

[0065] Optionally, the method for mapping the feature data to a data dimension includes: obtaining the data dimension of the feature data; if the data dimension does not match a preset dimension, then mapping the feature data to a data dimension to obtain feature data of a target dimension.

[0066] Specific examples Figure 2 As shown, if the feature data of the current target object is a 1-dimensional feature vector and the number of features is 7, then a 7*7 two-dimensional array is first constructed. The first row of the array is filled with the original 1-dimensional feature vector; the second row copies the feature vector from the previous row, pops the first feature from the beginning of the vector and appends it to the end of the vector, adding the vector to the second row of the array; the operation is similar for the third to seventh rows, thus completing the mapping from one-dimensional feature data to two-dimensional feature data, resulting in 7*7 two-dimensional feature data.

[0067] Furthermore, in order to fully extract the feature information from the feature data, the technical solution of this embodiment maps the feature data into a feature image, and obtains the classification result of the classification model based on the feature image, thereby improving the accuracy of the classification result.

[0068] Optionally, the method for performing pixel mapping on feature data includes: obtaining a preset pixel data range, performing pixel data mapping on each feature value in the feature data of the target dimension based on the pixel data range, and obtaining the feature image corresponding to the feature data.

[0069] Specifically, each feature value corresponding to each feature name in the feature data is obtained, the proportion of any feature value in each feature value is determined, and the pixel data corresponding to the feature value in the pixel data range of 0-255 is determined based on the proportion; further, the pixel data corresponding to each feature value is determined based on the above method, and the corresponding feature image is generated based on each pixel data, so as to facilitate the subsequent determination of the classification result corresponding to the target object based on the feature image.

[0070] It should be noted that the feature image obtained based on the above technical solution in this embodiment is a three-channel image; specifically, the same 1D vector is processed three times to form three 2D vectors, where the 1D vector obtained in the second processing is the reverse of the original vector. The use of a three-channel feature image is to obtain more layers of feature information, thereby making the classification result more accurate.

[0071] S130. Input the feature image into the preset classification model to obtain the classification result output by the classification model.

[0072] In this embodiment of the invention, the classification result includes any one of the ultra-long protocol, long protocol, antagonist protocol, or short protocol in the superovulation protocol. Of course, the above classification method is only an exemplary classification method introduced in this embodiment. The classification result of the superovulation protocol in this embodiment may also include other protocols such as the ultra-short protocol and the microstimulation protocol, depending on actual needs. These will not be listed one by one in this embodiment, and the classification method of the superovulation protocol is not limited in this embodiment.

[0073] In this embodiment, due to the small number and length of feature data, a smaller CNN network is used in the classification model to improve classification efficiency. Of course, other classification networks can also be used; the CNN network described above is merely an example. Specifically, the classification network in this embodiment, based on the VGGNet network, increases the network width of each layer by concatenating multi-sized convolutional kernels, thus enabling richer feature extraction capabilities. Furthermore, by increasing the residual connections in each convolutional block, the network depth is increased while reducing propagation loss.

[0074] It should be noted that the classification model in this embodiment includes at least one convolutional block, at least one residual layer, and at least one fully connected layer. The convolutional block is used to perform local perception on the feature image input to the classification model to obtain local features, and then performs a convolution operation on these local features to obtain the global features of the feature image. The residual layer is connected to the last convolutional block and is used to solve the problem of gradient vanishing during transmission in each convolutional layer of the convolutional block. This is equivalent to adding an identity term 1 to each derivative, so that even if the original derivative becomes very small, the error can still be effectively backpropagated. The fully connected layer is used to output the classification result of the classification model.

[0075] Specific examples Figure 3As shown, the classification model used in this embodiment (i.e., the mc_vgg network) consists of four convolutional blocks, each connected using residual connections. Each convolutional block includes at least one MCBatchNorm component. Specifically, the mc_vgg network contains 13 convolutional layers (excluding Max Pooling, ReLU, and DropOut layers); the number of channels in each convolutional layer is [64, 128, 256, 512]; each convolutional layer uses a multi-kernel batch normalization block component (MCBatchNorm) with the same structure, as shown in the lower right corner of Figure 3: each component consists of three convolutional kernels sequentially concatenated; specifically, the kernel sizes are 1x1, 3x3, and 5x5; furthermore, a pre-batch normalization process (BatchNorm) is applied before the output. The model ends with three fully connected (FC) layers, with the first two FC layers followed by Dropout layers.

[0076] Furthermore, the input vector (128,3,28,28) of the mc_vgg network represents a batch size of 128, 3 channels, and a graph size of 28x28; the input uses 3 channels.

[0077] The technical solution provided in this embodiment specifically includes: acquiring initial feature data of the target object; performing data preprocessing on the initial feature data to obtain feature data of the target object; wherein the initial feature data includes the basic physiological and hormonal characteristics of the target object; acquiring multiple feature data of the target object to improve the accuracy of the classification results; performing dimensional mapping and pixel mapping on the feature data to obtain feature images corresponding to the feature data; mapping the feature data to feature images to obtain deeper feature information, thereby further improving the accuracy of the classification results; inputting the feature images into a preset classification model to obtain the classification results output by the classification model; wherein the classification results include any one of the ultra-long protocol, long protocol, antagonist protocol, or short protocol in the superovulation protocol; by acquiring multiple feature data of the target object, mapping the feature data to feature images to obtain deeper feature information, and obtaining the classification results corresponding to the classification data, the accuracy of the classification of the superovulation protocol is improved, providing effective suggestions for doctors to select a superovulation protocol.

[0078] Example 2

[0079] Figure 4 This is a flowchart of a classification method based on a superovulation protocol provided in Embodiment 2 of the present invention. Based on the above embodiments, this embodiment adds a step of "image enhancement processing of the feature image" before the step "inputting the feature image into a preset classification model". Explanations of terms that are the same as or corresponding to those in the above embodiments are not repeated here. See also Figure 4The classification method based on superovulation protocols provided in this embodiment includes:

[0080] S210. Obtain the initial feature data of the target object, perform data preprocessing on the initial feature data, and obtain the feature data of the target object.

[0081] S220. Perform dimension mapping and pixel mapping on the feature data to obtain the feature image corresponding to the feature data.

[0082] S230. Perform image enhancement processing on the feature image.

[0083] In this embodiment of the invention, data augmentation mainly compensates for the insufficient quantity or uneven distribution of feature data, and it is difficult to find suitable methods for data augmentation of the initial feature data. Specifically, the image augmentation methods for feature images in this embodiment include, but are not limited to, brightness adjustment, saturation adjustment, and chroma adjustment.

[0084] S240. Input the feature image into the preset classification model to obtain the classification result output by the classification model.

[0085] The technical solution provided in this embodiment obtains multiple feature data of the target object, maps the feature data into feature images, obtains deeper feature information, and further performs data augmentation processing on the feature images to make up for the problems of insufficient sample quantity or uneven distribution, and can improve the classification accuracy of subsequent base models, thereby improving the classification accuracy of superovulation protocols and providing effective suggestions for doctors to select superovulation protocols.

[0086] The following are embodiments of the classification device based on the superovulation protocol provided in this invention. This device and the classification method based on the superovulation protocol in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the classification device based on the superovulation protocol, please refer to the embodiments of the classification method based on the superovulation protocol described above.

[0087] Example 3

[0088] Figure 5 This is a schematic diagram of the classification device based on superovulation protocols provided in Embodiment 3 of the present invention. This embodiment is applicable to situations where the classification results of superovulation protocols are determined based on user feature data. See also... Figure 5 The specific structure of this classification device based on the superovulation protocol includes: a feature data acquisition module 310, a feature image acquisition module 320, and a classification result determination module 330; wherein,

[0089] The feature data acquisition module 310 is used to acquire initial feature data of the target object, perform data preprocessing on the initial feature data, and obtain feature data of the target object; wherein, the initial feature data includes the basic physiological characteristics and hormonal characteristics of the target object;

[0090] The feature image acquisition module 320 is used to perform dimension mapping and pixel mapping on the feature data to obtain the feature image corresponding to the feature data.

[0091] The classification result determination module 330 is used to input the feature image into a preset classification model to obtain the classification result output by the classification model; wherein, the classification result includes any one of the ultra-long protocol, long protocol, antagonist protocol or short protocol in the superovulation protocol.

[0092] The technical solution provided in this embodiment specifically includes: acquiring initial feature data of the target object; performing data preprocessing on the initial feature data to obtain feature data of the target object; wherein the initial feature data includes the basic physiological and hormonal characteristics of the target object; acquiring multiple feature data of the target object to improve the accuracy of the classification results; performing dimensional mapping and pixel mapping on the feature data to obtain feature images corresponding to the feature data; mapping the feature data to feature images to obtain deeper feature information, thereby further improving the accuracy of the classification results; inputting the feature images into a preset classification model to obtain the classification results output by the classification model; wherein the classification results include any one of the ultra-long protocol, long protocol, antagonist protocol, or short protocol in the superovulation protocol; by acquiring multiple feature data of the target object, mapping the feature data to feature images to obtain deeper feature information, and obtaining the classification results corresponding to the classification data, the accuracy of the classification of the superovulation protocol is improved, providing effective suggestions for doctors to select a superovulation protocol.

[0093] Based on the above embodiments, the feature data acquisition module 310 includes:

[0094] The null value processing unit is used to obtain the feature value corresponding to each feature name in the initial feature data. If the feature value corresponding to the current feature name is null, the unit obtains the alternative feature name of the current feature name and the alternative feature value corresponding to the alternative feature name.

[0095] Based on the above embodiments, the feature data acquisition module 310 includes:

[0096] The feature value type conversion unit is used to obtain the feature value type of each feature value. If the feature value type of the current feature value is not a numeric type, then the feature value type of the current feature value is converted.

[0097] The normalization processing unit is used to normalize the values ​​of each feature value based on a preset numerical range to obtain the feature data of the target object.

[0098] Based on the above embodiments, the feature image acquisition module 320 includes:

[0099] The data dimension mapping unit is used to obtain the data dimension of the feature data. If the data dimension does not match the preset dimension, the feature data is mapped to obtain the feature data of the target dimension.

[0100] Based on the above embodiments, the feature image acquisition module 320 includes:

[0101] A pixel data mapping unit is used to obtain a preset pixel data range, and perform pixel data mapping on each feature value in the feature data of the target dimension based on the pixel data range to obtain the feature image corresponding to the feature data.

[0102] Based on the above embodiments, the device further includes:

[0103] An enhancement processing unit is used to perform image enhancement processing on the feature image before inputting it into a preset classification model, so as to achieve adaptive compensation of the feature image.

[0104] Based on the above embodiments, the classification model includes at least one convolutional block, at least one residual connected layer, and at least one fully connected layer; wherein,

[0105] The convolutional block is used to perform local perception on the feature image input to the classification model to obtain local features, and to perform convolution operation on the local features to obtain the global features of the feature image;

[0106] The residual layer is connected to the convolutional block of the last layer to solve the problem of gradient vanishing during transmission of each convolutional layer in the convolutional block.

[0107] The fully connected layer is used to output the classification result of the classification model.

[0108] The classification device based on the superovulation protocol provided in this embodiment of the invention can execute the classification method based on the superovulation protocol provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method execution.

[0109] It is worth noting that in the above embodiments of the classification device based on the superovulation protocol, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0110] Example 4

[0111] Figure 6 This is a schematic diagram of the structure of an electronic device provided in Embodiment 4 of the present invention. Figure 6 A block diagram is shown of an exemplary electronic device 12 suitable for implementing embodiments of the present invention. Figure 6 The electronic device 12 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of the present invention.

[0112] like Figure 6 As shown, the electronic device 12 is represented in the form of a general-purpose computing electronic device. The components of the electronic device 12 may include, but are not limited to: one or more processors or processing units 16, system memory 28, and bus 18 connecting different system components (including system memory 28 and processing unit 16).

[0113] Bus 18 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0114] Electronic device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by electronic device 12, including volatile and non-volatile media, removable and non-removable media.

[0115] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. Electronic device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be used to read and write non-removable, non-volatile magnetic media (… Figure 6 Not shown; usually referred to as a "hard drive"). Although Figure 6As not shown, disk drives for reading and writing to removable non-volatile disks (e.g., "floppy disks") and optical disc drives for reading and writing to removable non-volatile optical discs (e.g., CD-ROMs, DVD-ROMs, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0116] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 42 typically perform the functions and / or methods described in the embodiments of the present invention.

[0117] Electronic device 12 can also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), and with one or more devices that enable a user to interact with electronic device 12, and / or with any device that enables electronic device 12 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed via input / output (I / O) interface 22. Furthermore, electronic device 12 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 20. Figure 6 As shown, network adapter 20 communicates with other modules of electronic device 12 via bus 18. It should be understood that, although... Figure 6 As not shown, other hardware and / or software modules may be used in conjunction with electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0118] The processing unit 16 executes various functional applications and acquires sample data by running programs stored in the system memory 28. For example, it implements the steps of a classification method based on a superovulation protocol provided in this embodiment. The classification method based on a superovulation protocol includes:

[0119] The initial feature data of the target object is obtained, and the initial feature data is preprocessed to obtain the feature data of the target object; wherein, the initial feature data includes the basic physiological characteristics and hormonal characteristics of the target object;

[0120] Perform dimension mapping and pixel mapping on the feature data to obtain the feature image corresponding to the feature data;

[0121] The feature image is input into a preset classification model to obtain the classification result output by the classification model; wherein, the classification result includes any one of the ultra-long protocol, long protocol, antagonist protocol or short protocol in the superovulation protocol.

[0122] Of course, those skilled in the art will understand that the processor can also implement the technical solution of the sample data acquisition method provided in any embodiment of the present invention.

[0123] Example 5

[0124] This fifth embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements, for example, the steps of a classification method based on a superovulation protocol provided in this embodiment. The classification method based on a superovulation protocol includes:

[0125] The initial feature data of the target object is obtained, and the initial feature data is preprocessed to obtain the feature data of the target object; wherein, the initial feature data includes the basic physiological characteristics and hormonal characteristics of the target object;

[0126] Perform dimension mapping and pixel mapping on the feature data to obtain the feature image corresponding to the feature data;

[0127] The feature image is input into a preset classification model to obtain the classification result output by the classification model; wherein, the classification result includes any one of the ultra-long protocol, long protocol, antagonist protocol or short protocol in the superovulation protocol.

[0128] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0129] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0130] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0131] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0132] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computing device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0133] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A classification method based on superovulation protocols, characterized in that, include: Acquire initial feature data of the target object, and perform data preprocessing on the initial feature data to obtain feature data of the target object; wherein, the initial feature data includes the basic physiological characteristics and hormonal characteristics of the target object; the data preprocessing on the initial feature data includes: Obtain the feature value corresponding to each feature name in the initial feature data. If the feature value corresponding to the current feature name is empty, obtain the alternative feature name of the current feature name and obtain the alternative feature value corresponding to the alternative feature name. The data preprocessing of the initial feature data further includes: Obtain the feature value type of each of the aforementioned feature values. If the feature value type of the current feature value is not numeric, then perform feature value type conversion on the current feature value. Based on a preset numerical range, the numerical values ​​of each feature value are normalized to obtain the feature data of the target object. The feature data is subjected to dimension mapping and pixel mapping to obtain the feature image corresponding to the feature data; The feature image is input into a preset classification model to obtain the classification result output by the classification model; wherein, the classification result includes any one of the ultra-long protocol, long protocol, antagonist protocol or short protocol in the superovulation protocol.

2. The method according to claim 1, characterized in that, The dimension mapping of the feature data includes: The data dimension of the feature data is obtained. If the data dimension does not match the preset dimension, the feature data is mapped to the data dimension to obtain the feature data of the target dimension.

3. The method according to claim 2, characterized in that, The pixel mapping of the feature data includes: Obtain a preset range of pixel data, and perform pixel data mapping on each feature value in the feature data of the target dimension based on the pixel data range to obtain the feature image corresponding to the feature data.

4. The method according to claim 1, characterized in that, Before inputting the feature image into the preset classification model, the following steps are also included: The feature image is subjected to image enhancement processing to achieve adaptive compensation.

5. The method according to claim 1, characterized in that, The classification model includes at least one convolutional block, at least one residual connected layer, and at least one fully connected layer; wherein... The convolutional block is used to perform local perception on the feature image input to the classification model to obtain local features, and to perform convolution operation on the local features to obtain the global features of the feature image; The residual layer is connected to the convolutional block of the last layer to solve the problem of gradient vanishing during transmission of each convolutional layer in the convolutional block. The fully connected layer is used to output the classification result of the classification model.

6. A classification device based on a superovulation protocol, characterized in that, include: The feature data acquisition module is used to acquire initial feature data of the target object, perform data preprocessing on the initial feature data to obtain feature data of the target object; wherein, the initial feature data includes the basic physiological characteristics and hormonal characteristics of the target object; The feature data acquisition module includes: The null value processing unit is used to obtain the feature value corresponding to each feature name in the initial feature data. If the feature value corresponding to the current feature name is null, the unit obtains the alternative feature name of the current feature name and the alternative feature value corresponding to the alternative feature name. The feature data acquisition module further includes: The feature value type conversion unit is used to obtain the feature value type of each feature value. If the feature value type of the current feature value is not a numeric type, then the feature value type of the current feature value is converted. The normalization processing unit is used to normalize the values ​​of each feature value based on a preset numerical range to obtain the feature data of the target object. The feature image acquisition module is used to perform dimension mapping and pixel mapping on the feature data to obtain the feature image corresponding to the feature data; The classification result determination module is used to input the feature image into a preset classification model to obtain the classification result output by the classification model; wherein, the classification result includes any one of the ultra-long protocol, long protocol, antagonist protocol or short protocol in the superovulation protocol.

7. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the classification method based on the superovulation protocol as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the classification method based on the superovulation protocol as described in any one of claims 1-5.

Images