A method and system for screening in vitro fertilization high developmental potential embryos

Abstract

The application discloses a method and system for screening high development potential embryos of in vitro fertilization. The method for screening high development potential embryos of in vitro fertilization comprises the following steps: segmenting an embryo from a multi-focus embryo image obtained; segmenting a TE image from the multi-focus embryo image after embryo segmentation, and unfolding the segmented TE image; inputting the multi-focus embryo image after embryo segmentation, the unfolded TE image after segmentation, biochemical characteristics of a patient couple and maternal uterine state characteristics into a trained prediction model, and outputting an embryo development potential score; and selecting an embryo with the highest score as a highest development potential embryo. The application comprehensively considers the multi-focus embryo image after embryo segmentation, the unfolded TE image after segmentation, the biochemical characteristics of the patient couple and the maternal uterine state characteristics, quantifies the development potential of each embryo in multiple embryos of the same patient from multiple aspects, and can quickly and accurately screen the highest development potential embryo.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of embryo screening, and particularly to a screening method and system for high developmental potential embryos of in vitro fertilization. BACKGROUND

[0002] Infertility is a global public health issue affecting over 48.5 million couples worldwide. Since the birth of the first in vitro fertilization (IVF) baby in 1978, over 8 million children have been treated with IVF. Canada performs 35,000 IVF treatment cycles per year, and over 1.5 million treatment cycles globally, with more than 10% annual increase. However, the live birth rate of in vitro fertilization has remained at around 30% over the past few decades, and varies among different clinics. This is because most in vitro fertilization programs rely heavily on the experience of clinical staff and involve significant subjectivity and inconsistency. For a long time, IVF organizations have hoped to adopt data-driven quantitative methods to transform and standardize IVF programs.

[0003] To achieve this goal, artificial intelligence (AI) technology will play a key role. For example, in semen analysis, artificial intelligence can analyze the image of a single sperm and quantitatively classify it as normal and abnormal, thereby providing more accurate male infertility diagnosis and guiding subsequent treatment programs. AI can also predict sperm DNA quality from sperm images, thereby selecting high-quality sperm for intracytoplasmic sperm injection (ICSI).

[0004] Among the various factors affecting IVF outcomes, the quality (i.e., developmental potential) of the selected embryos for transfer is the main factor determining IVF success. The existing method for evaluating and selecting embryos is based on manual observation of embryo morphology. Manual judgment only checks a limited number of morphological features (i.e., blastocyst size, number of cells in the inner cell mass (ICM), and number of trophoblast cells (TE)) to "grade" the embryo (Human Reproduction, page 26, 1270-1283), which has two obvious limitations. First, other morphological features, such as the shape and size of cells in the ICM (the structure that develops into the fetus), also reflect the developmental potential of the embryo, but are not considered in the current embryo grading method. Second, embryo developmental potential is not entirely determined by morphological features. Biochemical information / features related to the patient also greatly affect embryo developmental potential. For example, the patient's age, hormone levels, and sperm quality (such as the DNA fragmentation index (DFI)) also severely affect in vivo embryo development. These biochemical features cannot be reflected by embryo morphology and are also missing in the current embryo evaluation practices. IVF treatment requires a data-driven embryo selection / evaluation method that considers comprehensive embryo morphological features and patient information to select embryos with the highest developmental potential for transfer. SUMMARY

[0005] To this end, the technical problem to be solved by the present application is to provide a method for screening in vitro fertilization high development potential embryos with high accuracy.

[0006] To solve the above technical problem, the present application provides a method for screening in vitro fertilization high development potential embryos, comprising:

[0007] segmenting the embryo from the obtained multi-focus embryo image;

[0008] segmenting the TE image from the multi-focus embryo image after embryo segmentation, and unfolding the segmented TE image;

[0009] inputting the multi-focus embryo image after embryo segmentation, the unfolded TE image after segmentation, the biochemical characteristics of the patient couple and the maternal uterine state characteristics into the trained prediction model, and outputting the embryo development potential score;

[0010] selecting the embryo with the highest score as the highest development potential embryo.

[0011] In an embodiment of the present application, the segmentation of the embryo from the obtained multi-focus embryo image comprises:

[0012] binaryzation of the image at each focal plane in the obtained multi-focus embryo image to obtain a rough mask;

[0013] thinning the rough mask;

[0014] extracting all contours using the thinned mask;

[0015] keeping the white pixels of the contour with the largest size to achieve embryo segmentation.

[0016] In an embodiment of the present application, the prediction model comprises a CNN network, an attention module, a Vit network, a multi-layer perceptron and a score fusion module;

[0017] the CNN network is used to predict the embryo development potential score from the multi-focus embryo image after embryo segmentation, the attention module is used to generate the weight of the embryo development potential score obtained from the multi-focus embryo image, and the weighted sum of all embryo development potential scores is taken as the embryo development potential score predicted from the multi-focus image;

[0018] the Vit network is used to generate the embryo development potential score from the unfolded TE image after segmentation;

[0019] the multi-layer perceptron is used to generate the embryo development potential score from the biochemical characteristics of the patient couple and the maternal uterine state characteristics;

[0020] The score fusion module is used for fusing three kinds of embryo development potential scores respectively predicted from the multi-focus embryo image after embryo segmentation, the expanded TE image after segmentation, biochemical characteristics of the patient couple and maternal uterine state characteristics, and outputting a final embryo development potential score.

[0021] In an embodiment of the present application, the attention module generates the weight by performing sequential convolution, average merging and S-type operation on the highest level feature map in the CNN network.

[0022] In an embodiment of the present application, the biochemical characteristics of the patient couple include:

[0023] father's semen characteristics;

[0024] maternal age, body mass index and treatment history;

[0025] day of blastocyst transfer, antral follicle count and number of retrieved oocytes;

[0026] maternal hormone profile;

[0027] The maternal uterine state characteristics include: endometrial thickness and endometrial type.

[0028] In an embodiment of the present application, the embryo is segmented from the acquired multi-layer focus embryo image, and further includes:

[0029] The focal plane is changed along the Z axis of the microscope, the embryo image is captured under a plurality of focal planes, and the multi-layer focus embryo image is constructed.

[0030] In an embodiment of the present application, the segmented TE image is expanded, including: expanding the segmented TE image by polar coordinate deformation.

[0031] The present application also provides a computer device, including a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the method of any one of the above-mentioned embodiments when executing the program.

[0032] The present application also provides a computer readable storage medium, which stores a computer program, wherein the program is executed by a processor to implement the steps of the method of any one of the above-mentioned embodiments.

[0033] The present application also provides an in vitro fertilization high development potential embryo screening system, which includes:

[0034] The first segmentation module is used for segmenting the embryo from the acquired multi-layer focus embryo image; the second segmentation module is used for segmenting the TE image from the multi-layer focus embryo image after embryo segmentation,

[0035] and expanding the segmented TE image;

[0036] a model prediction module, configured to input the multi-focus embryo image after embryo segmentation, the TE image after segmentation, biochemical characteristics of the patient couple and maternal uterine state characteristics into the trained prediction model, and output an embryo development potential score;

[0037] 0 a selection module, configured to select the embryo with the highest score as the highest development potential embryo.

[0038] The above technical scheme of the present application has the following advantages compared with the prior art:

[0039] The in vitro fertilization high development potential embryo screening method and system of the present application comprehensively consider the multi-focus embryo image after embryo segmentation, the TE image after segmentation, biochemical characteristics of the patient couple and maternal uterine state characteristics, quantize the development potential of each embryo in multiple embryos of the same patient from multiple aspects, and can quickly and accurately screen the highest development potential embryo.

[0040] The above description is only a summary of the technical scheme of the present application, in order to more clearly understand the technical means of the present application, the content of the specification can be implemented, and in order to make the above and other purposes, characteristics and advantages of the present application more obvious and easy to understand, the following preferred embodiments are described in detail below, and the accompanying drawings are as follows. BRIEF DESCRIPTION OF DRAWINGS

[0041] In order to make the content of the present application more easily understood, the following is a further detailed description of the present application according to specific embodiments of the present application and in conjunction with the accompanying drawings, in which

[0042] Figure 1 is a flowchart of the in vitro fertilization high development potential embryo screening method in the embodiment of the present application;

[0043] Figure 2 is a hardware schematic diagram used to obtain the multi-layer focus embryo image in the embodiment of the present application;

[0044] Figure 3 is a schematic diagram of the multi-layer focus embryo image obtained in the embodiment of the present application;

[0045] Figure 4 is a schematic diagram of the TE image segmented from the multi-layer focus embryo image after embryo segmentation and the unfolded segmented TE image in the embodiment of the present application;

[0046] Figure 5 is a schematic diagram of the prediction model in the embodiment of the present application.

[0047] Legend: 201, microscope; 202, embryo container; 203, objective lens; 204, focusing motor; 205, vibration isolation; 206, image acquisition unit; 207, motorized positioner; 208, host computer. DETAILED DESCRIPTION

[0048] The present application will be further described with reference to the drawings and specific examples, so that those skilled in the art can better understand the present application and implement it.

[0049] Example 1

[0050] Referring to Figure 1 The present embodiment discloses a method for screening high developmental potential embryos in in vitro fertilization, which comprises the following steps:

[0051] Step S1, segmenting the embryo from the acquired multi-focal embryo image;

[0052] Optionally, before step S1, the following step is further included:

[0053] Changing the focal plane along the Z-axis of the microscope, capturing the embryo image under multiple focal planes, and constructing the multi-focal embryo image.

[0054] Wherein, the multi-focal embryo image is captured by a camera installed on the microscope. The microscope is equipped with a computer-controlled focusing motor for changing the focal plane along the Z-axis of the microscope while capturing the multi-focal embryo image.

[0055] Referring to Figure 2 In one embodiment, the hardware used to acquire the multi-focal embryo image is as follows:

[0056] Microscope 201, for example, inverted optical microscope. The microscope can also include differential interference contrast (DIC) or phase contrast optics;

[0057] Host computer 208, which can include software and processors with instructions for controlling the elements of the system and image processing, including: controlling the focusing motor 204 to change the focal plane along the Z-axis of the microscope, controlling the image acquisition unit 206 to capture the image of the focal plane; retrieving the biochemical characteristics of the patient couple and the maternal uterine state characteristics from the interface or database; predicting the embryo developmental potential and training the embryo development prediction model;

[0058] Motorized positioner 207, which controls the movement of the embryo container 202 to position the embryo at the center of the field of view. The motorized positioner 207 is operatively connected or linked to the host computer 208 through a wired connection or a wireless connection;

[0059] High magnification (e.g. 100X) objective lens 203;

[0060] An image acquisition unit 206, for example, a CCD camera, which can be mounted on the microscope 201 and operatively connected or linked to the host computer 208 through a wired connection or a wireless connection;

[0061] A focusing motor 204, which can be mounted on the focusing knob and connected to the host computer 208 through a wired connection or a wireless connection;

[0062] The embryo container 202 can contain multiple embryos of the same patient, and the motorized positioner 207 can be controlled by the host computer 208 to position each embryo in the field of view. Further, a vibration isolation table 205 can also be included to minimize the vibration of the microscope 201, the motorized positioner 207, and the image acquisition unit 206.

[0063] The diameter of a blastocyst is 100-200 microns, which exceeds the depth of field of a microscope. Therefore, an embryo image captured at a single focal plane only partially reveals the morphological features of the embryo. The present invention proposes to use embryo images captured from multiple focal planes, which contain more comprehensive morphological features of the embryo and can more accurately predict the developmental potential of the embryo.

[0064] The multi-layered focal embryo image stack can be automatically captured by a computer-controlled camera (206) and a computer-controlled focus adjustment motor (204) mounted on the microscope (201). The computer controls the focusing motor to set the microscope focal plane to a series of predetermined values and controls the camera to capture images at each focal plane to build the multi-layered focal embryo image stack.

[0065] Referring to Figure 3 , an example of a stack of multi-layered focal embryo images captured at seven different focal planes ranging from -45 microns to 45 microns along the z-axis of the microscope. Compared to a single image captured at a single focal plane, the embryo image stack reveals more morphological features of the embryo. For example, the inner cell mass (red circle) is visible in the embryo image stack (at focal plane Z=45). However, it is not visible or less visible in other images from single focal planes (e.g., at focal planes Z=-15, -30, -45) because it is out of focus.

[0066] Referring to Figure 4 , in one embodiment, step S1 comprises:

[0067] The images at each focal plane in the acquired multi-layered focal embryo images (Referring to Figure 4 (a)) are binarized to obtain a coarse mask (Referring to Figure 4 (b)); wherein white pixels all belong to the embryo or cell fragments, and black pixels represent the embryo, cell fragments, and background.

[0068] Refining the coarse mask (see Figure 4 (c)) ; optionally, using GrubCut segmentation algorithm to refine the coarse mask.

[0069] Extracting all contours using the refined mask (see Figure 4 (d)) ;

[0070] Retaining white pixels of the contour with the largest size (see Figure 4 (e)), achieving embryo segmentation (see Figure 4 (f)).

[0071] Step S2, segmenting the TE image from the multi-focus embryo image after embryo segmentation (see Figure 4 (g)), and unfolding the segmented TE image (see Figure 4 (h)) ; optionally, unfolding the segmented TE image by polar coordinate deformation.

[0072] As shown in Figure 3 , the image captured from a single focal plane only contains in-focus morphological features, while out-of-focus morphological features are mostly missing. Therefore, the present application uses the image stack captured from different focal planes to include more comprehensive embryo morphological features as model input.

[0073] Step S3, inputting the multi-focus embryo image after embryo segmentation, the unfolded TE image after segmentation, the biochemical features of the patient couple and the maternal uterine state features into the trained prediction model, and outputting the embryo development potential score;

[0074] Referring to Figure 5 , in one embodiment, the prediction model includes a CNN network, an Attention Module, a Vit network (Vision Transformer), a multi-layer perceptron (MLP) and a score fusion module.

[0075] CNN network is the most advanced method for solving image-based classification problems. Existing cell neural networks use feature concatenation or majority voting method to make classification decisions using multiple input images. However, in the embryo image stack, the cells in the embryo have different sizes. Large cells can appear on multiple focal planes, while small cells can only appear on one focal plane. Therefore, directly connecting the features extracted from the multi-focus embryo image will undesirably enhance the features from large cells and weaken the features from small cells. In the majority voting method, the prediction results from the multi-focus embryo image are equally weighted. However, the present application finds that the multi-focus embryo image has different contributions to the prediction of embryo development potential.

[0076] Therefore in the present application, the CNN network predicts the embryo development potential score from the multi-focus embryo image after segmentation of the embryo, and the attention module is used to generate the weight of the embryo development potential score obtained from the multi-layer focus embryo image, and the weighted sum of all embryo development potential scores is taken as the embryo development potential score predicted from the multi-focus image; Specifically, the attention module generates the weight by performing sequential convolution, average merging and S operation on the highest level feature map in the CNN network. Because, the highest level feature map is directly related to the embryo development potential score, therefore is used to generate the weight.

[0077] The Vit network is used to generate the embryo development potential score from the TE image after segmentation and unfolding; the morphological features of TE cells are important predictors of embryo development potential. However, TE cells occupy a narrow annular band in the outer layer of the embryo image, and CNN performs poorly in processing such features. Therefore, the present application segments TE from the embryo image, unfolds the TE image into an elongated rectangular shape, and uses Vit to process the unfolded TE image. The reason why the present application chooses Vit instead of CNN is that Vit has a self-attention mechanism, which enables it to more accurately capture the cohesiveness (and tightness) of TE cells to predict embryo development potential.

[0078] The multi-layer perceptron is used to generate the embryo development potential score from the biochemical features and maternal uterine state features of the patient couple;

[0079] Among them, the patient's biochemical features that cannot be displayed by the embryo image are also important for predicting embryo development potential. The patient's biochemical features include the father's semen characteristics (e.g., raw semen volume, A-grade sperm ratio after semen processing, and sperm DNA fragmentation index); maternal age, body mass index, and treatment history; blastocyst transfer day, antral follicle count, and oocyte number; maternal hormone profile (such as progesterone, estradiol, luteinizing hormone, and free thyroxine). Maternal uterine state features include endometrial thickness and endometrial type.

[0080] The inclusion of maternal uterine state features in the model input helps to prevent bias / mislabeling, i.e., in the case of implantation failure, to mitigate the impact of adverse uterine conditions on the evaluation of biased embryos.

[0081] Therefore, in addition to morphological features, the addition of biochemical features related to the patient can more comprehensively reveal embryo development potential. It is well known that as a woman's age increases, the frequency of genetic abnormalities in oocytes also increases, thereby reducing embryo development potential and leading to poor clinical outcomes. Sperm DFI reflects the extent of damage to sperm DNA, and sperm with high DFI have lower embryo development potential and poorer clinical outcomes after fertilization with oocytes.

[0082] At the same time, uterine status features (i.e. endometrial thickness and endometrial pattern) are added to the model input. When an embryo with high developmental potential is transferred to a uterus with poor status, poor clinical outcome (failure) can be attributed to poor uterine status. Therefore, including uterine status features helps to prevent biased / wrong labeling, i.e. mitigates the effect of poor uterine status on biased embryo assessment in case of implantation failure.

[0083] To incorporate biochemical features and maternal uterine status features into the model input, the present application integrates the numerical features (biochemical features and uterine status features) with CNN and Vit; therefore, biochemical features, maternal uterine status features, embryo image stack and TE image are considered together for assessing the embryo. Classical methods like decision tree and SVM are not suitable for processing biochemical features here, as they cannot be easily integrated with CNN and Vit. Instead, a multi-layer perceptron (MLP) can be used, which predicts embryo developmental potential score from biochemical features and maternal uterine status.

[0084] The score fusion module is used to fuse three embryo developmental potential scores respectively predicted from the multi-focus embryo image after embryo segmentation, the unfolded TE image after segmentation, the biochemical features of the patient couple and the maternal uterine status features, and outputs the final embryo developmental potential score.

[0085] Specifically, the embryo developmental potential score predicted from the multi-layer focus embryo image, the embryo developmental potential score predicted from the TE, the embryo developmental potential score predicted from the biochemical features and uterine status features are added, and a constant value is taken as the final developmental potential score. The constant value is calculated as -ln(a), where a is the ratio of embryos with positive clinical outcome to embryos with negative clinical outcome. The constant value is used to mitigate the bias of model prediction for negative clinical outcome, which constitutes a large portion (e.g. 70%) of the dataset used to train the model.

[0086] Step S4, select the embryo with the highest score as the highest developmental potential embryo.

[0087] In addition, when the proposed embryo selection method is applied to different IVF clinics, the number of biochemical features and uterine status features of patients that can be used in the method can be different. In order to adapt the proposed method to this situation, a set of models are trained using different combinations of biochemical features and uterine status features of patients, and then the model matching the available biochemical features and uterine status features of patients is selected for predicting embryo developmental potential.

[0088] Embodiment two

[0089] The embodiment discloses a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method in the embodiment one when executing the program.

[0090] Embodiment three

[0091] The embodiment discloses a computer readable storage medium, which stores a computer program, and the program implements the steps of the method in the embodiment one when executed by a processor.

[0092] Embodiment four

[0093] The embodiment discloses a screening system for high development potential embryos of in vitro fertilization, comprising:

[0094] A first segmentation module is configured to segment the embryo from the acquired multi-focus embryo image;

[0095] A second segmentation module is configured to segment the TE image from the multi-focus embryo image after the embryo is segmented, and to unfold the segmented TE image;

[0096] A model prediction module is configured to input the multi-focus embryo image after the embryo is segmented, the unfolded TE image after the segmentation, biochemical characteristics of the patient couple and maternal uterine state characteristics into the trained prediction model, and to output an embryo development potential score;

[0097] A selection module is configured to select the embryo with the highest score as the highest development potential embryo.

[0098] The screening system for high development potential embryos of in vitro fertilization in the embodiment of the application is used to implement the aforementioned screening method for high development potential embryos of in vitro fertilization, and therefore the specific implementation manner of the system can be seen from the aforementioned embodiment part of the screening method for high development potential embryos of in vitro fertilization, and therefore the specific implementation manner can be referred to the description of the corresponding embodiment part, and will not be introduced here.

[0099] In addition, since the screening system for high development potential embryos of in vitro fertilization in the embodiment is used to implement the aforementioned screening method for high development potential embryos of in vitro fertilization, the function thereof corresponds to the function of the aforementioned method, and will not be described here.

[0100] Those skilled in the art should understand that the embodiments of the application can be provided as a method, a system or a computer program product. Therefore, the application can adopt a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the application can adopt a computer program product in the form of one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes.

[0101] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0102] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0103] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0104] Obviously, the above-described embodiments are only examples and are not intended to limit the present application. Based on the above description, one of ordinary skill in the art can further make other variations and changes to the present application. Here, it is not necessary or possible to enumerate all the embodiments. The obvious changes or variations derived therefrom are still within the scope of the present application.

Claims

1. A method for screening embryos with high developmental potential through in vitro fertilization, characterized in that: include: Segmenting the embryo from the acquired multi-layer focused embryo images; Segmenting a TE image from the multi-layer focused embryo image after embryo segmentation, and unfolding the segmented TE image; The multi-focus embryo image after embryo segmentation, the TE image expanded after segmentation, the biochemical characteristics of the patient couple and the maternal uterine state characteristics are input into the trained prediction model, and the embryo development potential score is output; the prediction model includes a CNN network, an attention module, a Vit network, a multi-layer perceptron and a score fusion module; the CNN network predicts the embryo development potential score from the multi-focus embryo image after embryo segmentation; the attention module is used to generate the weight of the embryo development potential score obtained from the multi-layer focused embryo image, and the weighted sum of all embryo development potential scores is used as the embryo development potential score predicted from the multi-focus image; the Vit network is used to generate the embryo development potential score from the TE image expanded after segmentation; the multi-layer perceptron is used to generate the embryo development potential score from the biochemical characteristics of the patient couple and the maternal uterine state characteristics; the score fusion module is used to fuse the three embryo development potential scores predicted from the multi-focus embryo image after embryo segmentation, the TE image expanded after segmentation, the biochemical characteristics of the patient couple and the maternal uterine state characteristics, and output the final embryo development potential score; The embryos with the highest scores were selected as the embryos with the highest developmental potential.

2. The method for screening in vitro fertilized high developmental potential embryos according to claim 1, characterized in that: The step of segmenting the embryo from the acquired multi-layer focused embryo image comprises: The image at each focal plane in the acquired multi-layer focused embryo image is binarized to obtain a rough mask; Refine the coarse mask; Extract all contours using the refined mask; The white pixels of the contour with the largest size are retained to achieve embryo segmentation.

3. The method for screening in vitro fertilized high developmental potential embryos according to claim 1, characterized in that: The attention module generates weights by performing sequential convolution, average pooling, and sigmoid operations on the highest-level feature map in the CNN network.

4. The method for screening in vitro fertilized high developmental potential embryos according to claim 1, characterized in that: The biochemical profile of the patient couple included: Father's semen characteristics; maternal age, body mass index, and treatment history; blastocyst transfer day, number of antral follicles, and number of retrieved oocytes; maternal hormone profile; The maternal uterine status characteristics include: endometrial thickness and endometrial type.

5. The method for screening in vitro fertilized high developmental potential embryos according to claim 1, characterized in that: The step of segmenting the embryo from the acquired multi-layer focused embryo image also includes: The focal plane is changed along the Z-axis of the microscope, embryo images are captured at multiple focal planes, and multi-layer focused embryo images are constructed.

6. The method for screening in vitro fertilized high developmental potential embryos according to claim 1, characterized in that: Expanding the segmented TE image includes expanding the segmented TE image by polar coordinate deformation.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method according to any one of claims 1 to 6 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A screening system for in vitro fertilized embryos with high developmental potential, characterized in that: The system is used to implement the steps of the method according to any one of claims 1 to 6, including: a first segmentation module, for segmenting the embryo from the acquired multi-layer focused embryo image; a second segmentation module for segmenting a TE image from the multi-layer focused embryo image after embryo segmentation, and unfolding the segmented TE image; The model prediction module is used to input the multi-focus embryo image after embryo segmentation, the expanded TE image after segmentation, the biochemical characteristics of the patient couple, and the maternal uterine status characteristics into the trained prediction model and output the embryo development potential score; The selection module is used to select the embryos with the highest scores as the embryos with the highest developmental potential.

Images