Embryo dysplasia detection method and electronic equipment

Abstract

The invention provides an embryonic dysplasia detection method and electronic equipment, and relates to the technical field of embryo evaluation. Comprising the following steps: collecting a current embryo image of a target embryo in an embryo development period; determining the current development stage of the target embryo according to the current embryo image; according to the current development stage of the target embryo, at least one anomaly detection strategy corresponding to the current development stage is adopted, and a development anomaly detection result of the target embryo in the current development stage is determined; and determining the development result of the target embryo according to the development abnormality detection result of the target embryo in each specified development stage. According to the scheme, the abnormal development phenomena of the embryos in different development stages are identified by adopting the abnormal detection strategies corresponding to the development stages, so that the abnormal development information of the embryos in different development stages can be synthesized, and the development potential of the embryos can be comprehensively and reliably evaluated.

Description

Technical Field

[0001] The present application relates to the technical field of embryo assessment, and in particular to a method and electronic device for detecting abnormal embryo development. Background Art

[0002] During assisted reproductive treatment, embryologists need to culture and observe embryos in vitro after fertilization in order to detect abnormalities in embryonic development and assist in screening out high-quality embryos.

[0003] In the prior art, morphological characteristics of the embryo during development are often extracted to determine whether there are developmental abnormalities in the embryo during development.

[0004] However, this judgment method often relies on the rich experience of embryologists and cannot provide a quantitative and objective assessment of abnormal embryos. Summary of the invention

[0005] The purpose of the present application is to provide a method and electronic device for detecting abnormal embryo development in order to improve the accuracy of the results of abnormal embryo development detection in view of the deficiencies in the above-mentioned prior art.

[0006] To achieve the above purpose, the technical solution adopted in the embodiment of the present application is as follows:

[0007] In a first aspect, the present application provides a method for detecting abnormal embryonic development, comprising:

[0008] During the embryonic development cycle, a current embryonic image of the target embryo is acquired;

[0009] Determining the current developmental stage of the target embryo according to the current embryo image;

[0010] According to the current developmental stage of the target embryo, at least one abnormality detection strategy corresponding to the current developmental stage is adopted to determine a developmental abnormality detection result of the target embryo at the current developmental stage;

[0011] The developmental result of the target embryo is determined according to the developmental abnormality detection result of the target embryo in each designated developmental stage; the designated developmental stage includes: pronuclear stage, cleavage stage and blastocyst stage.

[0012] Optionally, the step of determining the developmental abnormality detection result of the target embryo at the current developmental stage by using at least one abnormality detection strategy corresponding to the current developmental stage according to the current developmental stage of the target embryo includes:

[0013] If the current development stage is the pronuclear stage, segmenting the current embryo image using an egg cytoplasm segmentation network to extract an egg cytoplasm image;

[0014] Determining the roundness of the egg cytoplasm according to the egg cytoplasm image, and determining the oval abnormality result of the egg cytoplasm according to the roundness of the egg cytoplasm;

[0015] Determine the gap distance between the egg cytoplasm and the zona pellucida according to the egg cytoplasm image, and determine the abnormal result of the perivitelline space of the egg cytoplasm according to the gap distance;

[0016] According to the egg cytoplasm image, an egg cytoplasm abnormality detection model is used to determine the abnormality detection result of the egg cytoplasm, and the abnormality detection result is used to indicate whether the egg cytoplasm is abnormal and the corresponding abnormality type.

[0017] Optionally, determining the gap distance between the egg cytoplasm and the zona pellucida according to the egg cytoplasm image comprises:

[0018] The blastocyst segmentation model is used to segment the zona pellucida of the target embryo, and a mask image of the embryo body is generated based on the outer contour of the zona pellucida;

[0019] Extracting the egg cytoplasm mask image according to the egg cytoplasm image;

[0020] Determining a difference image of the target embryo according to the embryo main body mask image and the egg cytoplasm mask image;

[0021] Marking the connected domain of the difference image to determine the mask image of the perivitelline space area;

[0022] The gap distance between the egg cytoplasm and the zona pellucida is determined according to the mask image of the perivitelline space area.

[0023] Optionally, the adopting of an oocyte cytoplasm abnormality detection model to determine an abnormality detection result of the oocyte cytoplasm comprises:

[0024] Using a first classification model in the egg cytoplasm abnormality detection model, performing a first detection on the egg cytoplasm image to determine a first detection result, wherein the first detection result is used to indicate whether the egg cytoplasm has an abnormality;

[0025] If an abnormality is present, a second classification model in the egg cytoplasm abnormality detection model is used to perform a second detection on the egg cytoplasm image to determine a second detection result, wherein the second detection result is used to indicate the abnormal type of the egg cytoplasm;

[0026] The abnormal detection result of the egg cytoplasm is determined according to the first detection result and the second detection result.

[0027] Optionally, the step of determining the developmental abnormality detection result of the target embryo at the current developmental stage by using at least one abnormality detection strategy corresponding to the current developmental stage according to the current developmental stage of the target embryo includes:

[0028] If the current development stage is the cleavage stage, extracting the initial developmental timing data of the target embryo in the cleavage stage, the initial developmental timing data including: the current culture time of the target embryo and the developmental stage identifier corresponding to the current culture time;

[0029] Determining the cleavage pattern of the target embryo according to the initial developmental timing data of the cleavage stage;

[0030] Using a fragment segmentation network to segment the current embryo image and extract a fragment region image;

[0031] The abnormal fragmentation result of the target embryo is determined according to the fragmentation area image.

[0032] Optionally, determining the cleavage pattern of the target embryo according to the initial developmental timing data of the cleavage stage comprises:

[0033] Performing smoothing on the initial developmental time series data to obtain target developmental time series data;

[0034] Performing jump point detection on the target development time series data to obtain jump point time series data;

[0035] Performing differential calculation on the jump point timing data to obtain a difference value sequence;

[0036] The cleavage pattern of the target embryo is determined based on the difference sequence.

[0037] Optionally, the smoothing process is performed on the initial developmental time series data to obtain target developmental time series data, including:

[0038] According to a preset smoothing window size, the initial developmental time series data is subjected to moving smoothing;

[0039] Perform a first smoothing process according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, to obtain developmental time series data after the first smoothing process; the first smoothing process indicates taking the minimum value of each developmental stage identifier within the smoothing window;

[0040] For the developmental time series data after the first smoothing processing, a second smoothing processing is performed according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, so as to obtain the target developmental time series data. The second smoothing processing indicates taking the average of the developmental stage identifiers within the smoothing window.

[0041] Optionally, determining the cleavage pattern of the target embryo according to the difference sequence includes:

[0042] According to the numerical attributes of the elements in the difference sequence, the cleavage type is determined, and the cleavage type is one of the following: normal cleavage, reverse cleavage or direct cleavage; the numerical attributes include one or more of the following: positive or negative value, value size, and position where the value appears.

[0043] Optionally, determining the abnormal fragmentation result of the target embryo according to the fragmentation region image includes:

[0044] Determining the diameter and roundness of each candidate fragment according to the fragment region image;

[0045] Determine a target fragment from each candidate fragment according to the diameter, roundness and current development stage of each candidate fragment, and determine a fragment proportion result according to the target fragment, wherein the fragment proportion result is used to indicate the proportion of the fragment in the target embryo body;

[0046] According to the fragment proportion result, the fragment abnormality result is determined.

[0047] Optionally, the step of determining the developmental abnormality detection result of the target embryo at the current developmental stage by using at least one abnormality detection strategy corresponding to the current developmental stage according to the current developmental stage of the target embryo includes:

[0048] If the current development stage is the blastocyst stage, extract the time series data of the cavity ratio of the target embryo in the blastocyst stage, wherein the time series data of the cavity ratio includes: the current culture time of the target embryo and the area ratio of the cavity area to the blastocyst main area corresponding to the current culture time;

[0049] According to the cavity proportion time series data, a time series downward trend detection is performed to obtain a time series downward trend detection result, wherein the time series downward trend detection result includes: the start time, the end time and the decline amplitude of each time series trend decline;

[0050] The shrinkage result of the target embryo is determined according to the detection result of the time series downward trend of the cavity proportion, and the shrinkage result includes: the number of shrinkage occurrences and the duration of each shrinkage.

[0051] Optionally, determining the shrinkage result of the target embryo according to the cavity proportion time series downward trend detection result includes:

[0052] Determine the duration of the current decline according to the start time and end time of the current cavity proportion time series decline trend;

[0053] If the duration of the current decrease in the cavity proportion meets the preset time, and the decrease amplitude of the current cavity proportion time series trend meets the preset amplitude, the number of shrinkages of the target embryo is increased by 1, and the duration of the current decrease in the cavity proportion is recorded.

[0054] In a second aspect, the embodiment of the present application further provides a device for detecting abnormal embryonic development, including: a collection module and a determination module;

[0055] The acquisition module is used to acquire the current embryo image of the target embryo during the embryo development cycle;

[0056] The determination module is used to determine the current development stage of the target embryo according to the current embryo image;

[0057] The determination module is used to determine the developmental abnormality detection result of the target embryo at the current developmental stage by adopting at least one abnormality detection strategy corresponding to the current developmental stage according to the current developmental stage of the target embryo;

[0058] The determination module is used to determine the developmental result of the target embryo according to the developmental abnormality detection results of the target embryo in each designated developmental stage; the designated developmental stages include: pronuclear stage, cleavage stage and blastocyst stage.

[0059] Optionally, the determination module is specifically configured to segment the current embryo image using an egg cytoplasm segmentation network to extract an egg cytoplasm image if the current development stage is the pronuclear stage;

[0060] Determining the roundness of the egg cytoplasm according to the egg cytoplasm image, and determining the oval abnormality result of the egg cytoplasm according to the roundness of the egg cytoplasm;

[0061] Determine the gap distance between the egg cytoplasm and the zona pellucida according to the egg cytoplasm image, and determine the abnormal result of the perivitelline space of the egg cytoplasm according to the gap distance;

[0062] According to the egg cytoplasm image, an egg cytoplasm abnormality detection model is used to determine the abnormality detection result of the egg cytoplasm, and the abnormality detection result is used to indicate whether the egg cytoplasm is abnormal and the corresponding abnormality type.

[0063] Optionally, the determination module is specifically used to segment the zona pellucida of the target embryo using a blastocyst segmentation model, and generate an embryo main body mask image based on the outer contour of the zona pellucida;

[0064] Extracting the egg cytoplasm mask image according to the egg cytoplasm image;

[0065] Determining a difference image of the target embryo according to the embryo main body mask image and the egg cytoplasm mask image;

[0066] Marking the connected domain of the difference image to determine the mask image of the perivitelline space area;

[0067] The gap distance between the egg cytoplasm and the zona pellucida is determined according to the mask image of the perivitelline space area.

[0068] Optionally, the determination module is specifically used to adopt a first classification model in the egg cytoplasm abnormality detection model to perform a first detection on the egg cytoplasm image to determine a first detection result, wherein the first detection result is used to indicate whether the egg cytoplasm has an abnormality;

[0069] If an abnormality is present, a second classification model in the egg cytoplasm abnormality detection model is used to perform a second detection on the egg cytoplasm image to determine a second detection result, wherein the second detection result is used to indicate the abnormal type of the egg cytoplasm;

[0070] The abnormal detection result of the egg cytoplasm is determined according to the first detection result and the second detection result.

[0071] Optionally, the determination module is specifically used to extract the initial developmental timing data of the target embryo in the cleavage stage if the current developmental stage is the cleavage stage, the initial developmental timing data including: the current culture time of the target embryo and the developmental stage identifier corresponding to the current culture time;

[0072] Determining the cleavage pattern of the target embryo according to the initial developmental timing data of the cleavage stage;

[0073] Using a fragment segmentation network to segment the current embryo image and extract a fragment region image;

[0074] The abnormal fragmentation result of the target embryo is determined according to the fragmentation area image.

[0075] Optionally, the determination module is specifically used to perform smoothing on the initial developmental time series data to obtain target developmental time series data;

[0076] Performing jump point detection on the target development time series data to obtain jump point time series data;

[0077] Performing differential calculation on the jump point timing data to obtain a difference value sequence;

[0078] The cleavage pattern of the target embryo is determined based on the difference sequence.

[0079] Optionally, the determination module is specifically used to perform moving smoothing on the initial development time series data according to a preset smoothing window size;

[0080] Perform a first smoothing process according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, to obtain developmental time series data after the first smoothing process; the first smoothing process indicates taking the minimum value of the identifiers within the smoothing window of each developmental stage;

[0081] For the developmental time series data after the first smoothing processing, a second smoothing processing is performed according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, so as to obtain the target developmental time series data. The second smoothing processing indicates taking the average of the identifiers within the smoothing window of each developmental stage.

[0082] Optionally, the determination module is specifically used to determine the cleavage type according to the numerical attributes of the elements in the difference sequence, and the cleavage type is one of the following: normal cleavage, reverse cleavage or direct cleavage; the numerical attributes include one or more of the following: the positive or negative value, the size of the value, and the position where the value appears.

[0083] Optionally, the determination module is specifically configured to determine the diameter and roundness of each candidate fragment according to the fragment region image;

[0084] Determine a target fragment from each candidate fragment according to the diameter, roundness and current development stage of each candidate fragment, and determine a fragment proportion result according to the target fragment, wherein the fragment proportion result is used to indicate the proportion of the fragment in the target embryo body;

[0085] According to the fragment proportion result, the fragment abnormality result is determined.

[0086] Optionally, the determination module is specifically used to extract the time series data of the cavity proportion of the target embryo in the blastocyst stage if the current development stage is the blastocyst stage, and the time series data of the cavity proportion includes: the current culture time of the target embryo and the area ratio of the cavity area to the blastocyst main body area corresponding to the current culture time;

[0087] According to the cavity proportion time series data, a time series downward trend detection is performed to obtain a cavity proportion time series downward trend detection result, wherein the cavity proportion time series downward trend detection result includes: the start time, the end time and the decline amplitude of each time when the cavity proportion time series trend declines;

[0088] The shrinkage result of the target embryo is determined according to the detection result of the time series downward trend of the cavity proportion, and the shrinkage result includes: the number of shrinkage occurrences and the duration of each shrinkage.

[0089] Optionally, the determination module is specifically used to determine the duration of the current cavity proportion decrease according to the start time and end time of the current cavity proportion time series trend decrease;

[0090] If the duration of the current decrease in the cavity proportion meets the preset time, and the decrease amplitude of the current cavity proportion time series trend meets the preset amplitude, the number of shrinkages of the target embryo is increased by 1, and the duration of the current decrease in the cavity proportion is recorded.

[0091] In a third aspect, an embodiment of the present application provides an electronic device comprising: a processor, a storage medium and a bus, wherein the storage medium stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the storage medium communicate via the bus, and the processor executes the machine-readable instructions to implement the method for detecting abnormal embryonic development as provided in the first aspect.

[0092] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for detecting abnormal embryonic development as provided in the first aspect is executed.

[0093] The beneficial effects of this application are:

[0094] The present application provides an embryonic development abnormality detection method and electronic device, including: collecting the current embryonic image of the target embryo during the embryonic development cycle; determining the current development stage of the target embryo according to the current embryonic image; according to the current development stage of the target embryo, using at least one abnormality detection strategy corresponding to the current development stage to determine the development abnormality detection result of the target embryo at the current development stage; according to the development abnormality detection result of the target embryo in each specified development stage, determining the development result of the target embryo. This solution monitors the embryonic development cycle to collect embryonic images at different development stages during the embryonic development process, and uses abnormality detection strategies corresponding to each development stage to identify the development abnormality of the embryo at different development stages, so that the abnormal development information of the embryo at different development stages can be integrated to comprehensively and reliably evaluate the development potential of the embryo. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0096] Figure 1A schematic diagram of a process for detecting abnormal embryonic development provided in an embodiment of the present application;

[0097] Figure 2 A schematic diagram of an embryonic development cycle provided in an embodiment of the present application;

[0098] Figure 3 A schematic diagram of another method for detecting abnormal embryonic development provided in an embodiment of the present application;

[0099] Figure 4 A schematic diagram of an embryo image provided in an embodiment of the present application;

[0100] Figure 5 A schematic diagram of a process for detecting an abnormal embryonic development provided in an embodiment of the present application;

[0101] Figure 6 A schematic diagram of another embryo image provided in an embodiment of the present application;

[0102] Figure 7 A schematic diagram of another method for detecting abnormal embryonic development provided in an embodiment of the present application;

[0103] Figure 8 This is a schematic diagram of the abnormal cytoplasmic region labeling result of an embodiment of the present application;

[0104] Fig. 9 A schematic diagram of a process for detecting an abnormal embryonic development provided in an embodiment of the present application;

[0105] Fig.10 A schematic diagram of a process for detecting an abnormal embryonic development provided in an embodiment of the present application;

[0106] Fig.11 A schematic diagram of another method for detecting abnormal embryonic development provided in an embodiment of the present application;

[0107] Fig.12 A schematic diagram of the cleavage pattern provided in the embodiment of the present application;

[0108] Fig.13 A schematic diagram of another method for detecting abnormal embryonic development provided in an embodiment of the present application;

[0109] Fig.14 A schematic diagram of a fragment area image adopted by an embodiment of the present application;

[0110] Fig.15 A schematic diagram of a process for detecting an abnormal embryonic development provided in an embodiment of the present application;

[0111] Fig.16A schematic diagram of a blastocyst segmentation result provided in an embodiment of the present application;

[0112] Fig.17 A schematic diagram of a timing curve provided in an embodiment of the present application;

[0113] Fig.18 A schematic diagram of a downward trend detection result provided in an embodiment of the present application;

[0114] Fig.19 A schematic diagram of another method for detecting abnormal embryonic development provided in an embodiment of the present application;

[0115] Fig. 20 A schematic diagram of a device for detecting abnormal embryonic development provided in an embodiment of the present application;

[0116] Fig.21 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0117] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of order, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.

[0118] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0119] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0120] In recent years, my country's assisted reproductive technology has developed rapidly, with the total number of technical service cycles exceeding 1 million and the number of births exceeding 300,000 each year. In the in vitro fertilization-embryo transfer (IVF-ET) assisted pregnancy cycle, the potential assessment of embryos and the selection of transplanted embryos play a decisive role. Usually, the assessment and selection of embryos are determined by regular observation and scoring by embryologists. For embryologists, this is a time-consuming and energy-consuming process. With the increase in the number of patients treated with assisted reproductive treatment and the increase in the number of cycles in the center, it means that a large number of experienced embryologists are needed to devote themselves to this work. In addition, embryonic development is a relatively complex process. For older women or those with poor egg quality, it is often accompanied by abnormal effects in the embryonic development process. These abnormal phenomena in the development process are often related to outcomes such as pregnancy failure and miscarriage. Therefore, the automatic identification and labeling of these abnormal phenomena has become a key issue in improving the efficiency of assisted reproductive treatment.

[0121] Existing embryo abnormality detection mainly relies on two methods. One is through embryo biopsy, which can detect potential genetic and chromosomal abnormalities of embryos, but biopsy requires sampling, which is a situation that will cause damage to in vitro embryos. The other is to judge whether there are abnormalities in the embryo development process through morphological characterization of the embryo development process. The determination and degree of abnormalities will affect the embryo transplantation outcome. These tasks often require experienced embryologists to make judgments. Existing technical solutions for embryo development abnormality detection based on time-difference incubators are mostly focused on the embryo cleavage process, ignoring the detection of abnormalities in pronuclear and blastocyst stage embryos, and ignoring the identification and automatic recording of key data such as the occurrence and duration of abnormalities, the degree of abnormalities, and the number of times abnormalities occur. The lack of these key data makes it impossible to make a quantitative and objective evaluation of abnormal embryos.

[0122] Based on this, this solution provides a method for detecting abnormal embryo development, which collects images of embryos in the development process in real time, identifies the developmental stage of the embryo based on the collected embryo images, and detects abnormalities of the embryos at the current developmental stage according to the abnormality handling strategy corresponding to the developmental stage, thereby detecting abnormalities that may occur in the embryos at different developmental stages, and comprehensively analyzing the abnormalities at each developmental stage to achieve an evaluation of the embryo's development. Compared with the prior art, this solution provides a means of detecting abnormalities at different developmental stages, which can comprehensively analyze all abnormalities of the embryo during development, thereby achieving a quantitative and objective evaluation of the embryo.

[0123] Next, some terms that may be involved in the following embodiments of this solution are explained:

[0124] TLs: time difference incubator.

[0125] SERa: Smooth endoplasmic reticulum aggregation abnormality (SER: smooth endoplasmic reticulum) refers to the aggregation of a large number of smooth endoplasmic reticulum in the center of the oocyte cytoplasm to form a clear, mostly round, flat, microscopic disc. This aggregation will cause abnormal release of Ca2+ to a certain extent, disrupting the energy supply and demand balance in the oocyte, which may have an adverse effect on the development of the early embryo.

[0126] CLCG: Centrally located cytoplasmic granularity.

[0127] N2PN: Two pronuclei are not present, non-2PN (Non-2PN) 0PN, 1PN, 3PN, and 4PN are all abnormal states.

[0128] Vacuole: Vacuole.

[0129] DA: Developmental arrest, unfertilized embryo or developmental arrest.

[0130] DC: Direct cleavage.

[0131] DUC: Direct Unequal Cleavages.

[0132] RC: Reverse cleavage.

[0133] BF: The fragment content exceeds 50% (Big Frament).

[0134] BS: Blastocyst shrunken.

[0135] Figure 1 A schematic diagram of a process for detecting abnormal embryonic development provided in an embodiment of the present application; the execution subject of the method may be a computer device, such as Figure 1 As shown, the method may include:

[0136] S101. During the embryonic development cycle, a current embryo image of a target embryo is acquired.

[0137] The target embryo can be cultured in a time difference incubator, and during the embryo development cycle, the current embryo image of the target embryo can be collected from the time difference incubator at a preset interval, wherein the embryo development cycle can include the entire cycle from the beginning to the end of the embryo development, that is, from the pronuclear stage-cleavage stage-blastocyst stage.

[0138] Figure 2 A schematic diagram of an embryonic development cycle is provided in an embodiment of the present application, and the embryonic development process is defined as the following periods: t1 (pronuclear stage), t2 (2-cell stage), t3 (3-cell stage), t4 (4-cell stage), t5 (5-cell stage), t6 (6-cell stage), t7 (7-cell stage), t8 (8-cell stage), t9 (9-cell stage and above), tM (fusion stage), and tB (blastocyst stage), wherein t1 to tm represent the starting time of the corresponding period, and tSB is the starting time of the tB period.

[0139] S102: Determine the current developmental stage of the target embryo according to the current embryo image.

[0140] Based on the acquired current embryo image, image recognition can be performed to determine the current developmental stage of the target embryo.

[0141] It is worth mentioning that T2 (2-cell stage), T3 (3-cell stage), T4 (4-cell stage), T5 (5-cell stage), T6 (6-cell stage), T7 (7-cell stage), T8 (8-cell stage), and T9 (9-cell stage and above) all belong to the cleavage stage.

[0142] S103. According to the current developmental stage of the target embryo, at least one abnormality detection strategy corresponding to the current developmental stage is adopted to determine the developmental abnormality detection result of the target embryo at the current developmental stage.

[0143] For different developmental stages, this solution provides different abnormality detection strategies to identify abnormal developmental phenomena unique to different developmental stages.

[0144] According to the current developmental stage of the identified target embryo, at least one abnormality detection strategy corresponding to the current developmental stage can be used to detect and obtain a developmental abnormality detection result of the target embryo at the current developmental stage.

[0145] S104, determining the developmental outcome of the target embryo according to the developmental abnormality detection results of the target embryo in each designated developmental stage.

[0146] Designated developmental stages include: pronuclear stage, cleavage stage, and blastocyst stage.

[0147] Optionally, the current developmental stage in this embodiment can be one of the specified developmental stages, mainly including: pronuclear stage, cleavage stage and blastocyst stage. By comprehensively analyzing the developmental abnormality detection results of the target embryo in the pronuclear stage, cleavage stage and blastocyst stage, the developmental potential of the target embryo can be quantitatively and objectively evaluated, and the analysis results are more comprehensive and reliable.

[0148] In summary, the embryonic development abnormality detection method provided by this embodiment includes: collecting the current embryonic image of the target embryo during the embryonic development cycle; determining the current development stage of the target embryo according to the current embryonic image; according to the current development stage of the target embryo, using at least one abnormality detection strategy corresponding to the current development stage to determine the development abnormality detection result of the target embryo at the current development stage; according to the development abnormality detection result of the target embryo in each specified development stage, determine the development result of the target embryo. This scheme monitors the embryonic development cycle to collect embryonic images at different developmental stages during the embryonic development process, and uses abnormality detection strategies corresponding to each developmental stage to identify the developmental abnormalities of the embryo at different developmental stages, so that the abnormal development information of the embryo at different developmental stages can be integrated to comprehensively and reliably evaluate the developmental potential of the embryo.

[0149] Figure 3 A flow chart of another method for detecting embryonic developmental abnormalities provided in an embodiment of the present application; Optionally, in step S103, according to the current developmental stage of the target embryo, at least one abnormality detection strategy corresponding to the current developmental stage is adopted to determine the developmental abnormality detection result of the target embryo at the current developmental stage, which may include:

[0150] S301. If the current development stage is the pronuclear stage, the egg cytoplasm segmentation network is used to segment the current embryo image and extract the egg cytoplasm image.

[0151] The oocyte cytoplasm segmentation network is used to extract the oocyte cytoplasm region of the embryo. This module is composed of an instance segmentation network model. By inputting the current embryo image into the oocyte cytoplasm segmentation network, the oocyte cytoplasm image can be extracted from the current embryo image. Afterwards, the extracted oocyte cytoplasm image can be filled with the background mean.

[0152] Figure 4 A schematic diagram of an embryo image provided in an embodiment of the present application, Figure 4 Figure a shows the extracted current embryo image at the pronuclear stage. Figure 4 Panel b shows the oocyte cytoplasm image after background mean filling.

[0153] S302: Determine the roundness of the egg cytoplasm according to the egg cytoplasm image, and determine the oval abnormality result of the egg cytoplasm according to the roundness of the egg cytoplasm.

[0154] Optionally, based on the oocyte cytoplasm image, the perimeter L of the oocyte cytoplasm can be calculated first. i With area S i , through the area and perimeter, the following formula can be used: The roundness of the ovum cytoplasm was calculated.

[0155] In some embodiments, the roundness of the egg cytoplasm can be compared with a preset threshold to obtain an abnormal oval shape result, wherein the abnormal oval shape result can be: abnormal oval shape or normal oval shape.

[0156] Generally, when the roundness of the egg cytoplasm is less than 0.75, it can be considered that the target embryo has an abnormal egg shape.

[0157] S303, determining the gap distance between the egg cytoplasm and the zona pellucida according to the egg cytoplasm image, and determining the abnormal result of the perivitelline space of the egg cytoplasm according to the gap distance.

[0158] Optionally, the embryo main body mask image Embryo_Mask and the mask image Oocyte_Mask of the egg cytoplasm can be extracted first. Then, the gap distance between the egg cytoplasm and the zona pellucida is determined according to the embryo main body mask image and the mask image of the egg cytoplasm, which can also be called the periovum gap width, preferably the maximum gap distance between the egg cytoplasm and the zona pellucida. According to the periovum gap width, the periovum gap abnormal result is determined, and the periovum gap abnormal result is used to indicate whether the target embryo has abnormality.

[0159] S304: According to the oocyte cytoplasm image, an oocyte cytoplasm abnormality detection model is used to determine an oocyte cytoplasm abnormality detection result.

[0160] The abnormality test results are used to indicate whether there is an abnormality in the egg cytoplasm and the corresponding abnormality type.

[0161] The egg cytoplasm abnormality detection model used in this embodiment can be composed of a multi-task learning network, which can be used to determine whether there is an abnormality in the egg cytoplasm, and can also be used to determine the abnormality category when an abnormality exists in the egg cytoplasm.

[0162] Therefore, the abnormality detection result obtained may be: there is no abnormality in the egg cytoplasm; or the abnormality detection result obtained may be: there is an abnormality in the egg cytoplasm and the corresponding abnormality category.

[0163] Figure 5 A schematic flow chart of another method for detecting abnormal embryonic development provided in an embodiment of the present application; Optionally, in step S303, determining the gap distance between the egg cytoplasm and the zona pellucida according to the egg cytoplasm image may include:

[0164] S501, segmenting the zona pellucida of the target embryo using a blastocyst segmentation model, and generating an embryo main body mask image based on the outer contour of the zona pellucida.

[0165] Optionally, the zona pellucida of the current embryo image at the pronuclear stage may be segmented using a blastocyst segmentation model, and an embryo main body mask image Embryo_Mask may be generated based on the outer contour of the zona pellucida.

[0166] S502: extracting an oocyte cytoplasm mask image according to the oocyte cytoplasm image.

[0167] Similarly, based on the oocyte cytoplasm image, image segmentation is performed to extract the oocyte cytoplasm mask image Oocyte_Mask.

[0168] Figure 6 Schematic diagram of another embryo image provided in the embodiment of the present application. Figure 6 Figure a shows the extracted oocyte mask image Oocyte_Mask. Figure 6 Figure b shows the extracted embryo body mask image Embryo_Mask.

[0169] S503: Determine a difference image of the target embryo according to the embryo main body mask image and the egg cytoplasm mask image.

[0170] The difference map of the embryo image can be calculated by the oocyte mask image Oocyte_Mask and the embryo body mask image Embryo_Mask, such as Figure 6 The difference image shown in d.

[0171] S504: Mark the connected domain of the difference image to determine the mask image of the perivitelline space area.

[0172] Next, the connected domain of the difference image is marked, and the area with the largest connected domain is retained as the perivitelline space area, and the mask image Pv_Mask of the perivitelline space area is obtained as follows: Figure 6 As shown in f.

[0173] S505: Determine the gap distance between the egg cytoplasm and the zona pellucida based on the mask image of the perivitelline space region.

[0174] The mask image Pv_Mask of the perivitelline space is transformed into polar coordinates, and the annular structure is transformed into a rectangular structure to obtain Pv_Mask_Polar. The width of the converted Pv_Mask_Polar is the width of the perivitelline space, that is, the gap distance between the egg cytoplasm and the zona pellucida. Figure 6 The g in the figure shows the Pv_Mask_Polar obtained after conversion, in which the width of the perivitelline space is marked.

[0175] In some embodiments, when the width of the perivitelline space exceeds a certain threshold, the embryo may be considered abnormal.

[0176] Figure 7 A schematic flow chart of another method for detecting abnormal embryonic development provided in an embodiment of the present application; Optionally, in step S304, using an oocyte cytoplasm abnormality detection model to determine an abnormal oocyte cytoplasm detection result may include:

[0177] S701. Use the first classification model in the egg cytoplasm abnormality detection model to perform a first detection on the egg cytoplasm image and determine a first detection result.

[0178] The first test result is used to indicate whether there is an abnormality in the egg cytoplasm.

[0179] The oocyte abnormality detection model can be composed of an MTL (Multi-Task Learning) network. The framework is based on MTL, and the backbone network uses ResNet-34. Multi-task learning is a form of transfer learning. It learns multiple related tasks at the same time by sharing feature weights, so that these tasks can achieve better results than training a task alone, and the model has better generalization.

[0180] In deep learning models, the most direct way to implement multi-task learning is to have multiple tasks share the underlying multi-layer network parameters, and configure the underlying Task-specific parameters for different tasks at the model output layer. The underlying network here refers to the same hidden layers shared by multiple tasks. Different tasks learn some common low-level features by sharing the bottom hidden layers, and then start to branch out in the network near the output layer (Task-specific layers) to implement different tasks.

[0181] In this embodiment, the egg cytoplasm abnormality detection model can be divided into two classification models, that is, used to implement two tasks. The first classification model is used to detect whether the egg cytoplasm is abnormal, and the second classification model is used to detect the abnormal type of the egg cytoplasm.

[0182] Optionally, the first classification model in the egg cytoplasm abnormality detection model can be used to perform a first detection and identification on the egg cytoplasm image to obtain a first detection result, and the first detection result is used to indicate whether there is an abnormality in the egg cytoplasm.

[0183] S702: If an abnormality exists, a second classification model in the egg cytoplasm abnormality detection model is used to perform a second detection on the egg cytoplasm image to determine a second detection result.

[0184] The second test result is used to indicate the type of abnormality in the egg cytoplasm.

[0185] When there is an abnormality in the ovum cytoplasm, a second test can be performed through the second classification model in the ovum cytoplasm abnormality detection model to obtain a second test result, and the second test result is used to indicate the abnormal type of the ovum cytoplasm. That is, the second test result is used to determine whether the ovum cytoplasm belongs to the three abnormalities of SER (abnormal smooth endoplasmic reticulum aggregation), Vacuole (vacuole) or CLCG (central granulation of oocyte cytoplasm).

[0186] Of course, if the first test result indicates that there is no abnormality in the egg cytoplasm, the second test will not be performed.

[0187] S703: Determine the abnormality detection result of the egg cytoplasm according to the first detection result and the second detection result.

[0188] By combining the first test result and the second test result, an abnormal test result of the egg cytoplasm can be obtained.

[0189] Illustratively, the abnormality detection result of the egg cytoplasm may be: there is no abnormality in the egg cytoplasm.

[0190] The abnormal test result of the egg cytoplasm may also be: there is an abnormality in the egg cytoplasm, and the abnormality type is SER.

[0191] It is worth noting that the first classification model can use two types of oocyte cytoplasm images, one with abnormal oocyte cytoplasm and one with normal oocyte cytoplasm, as training samples to train the model to accurately classify abnormal oocyte cytoplasm and normal oocyte cytoplasm. The second classification model can use oocyte cytoplasm images with three abnormal categories of SER, Vacuole, and CLCG as training samples to train the model to accurately classify different abnormal categories.

[0192] In some embodiments, the abnormal detection results of the above-mentioned oocyte cytoplasm can also be visualized by a visualization processing method.

[0193] Optionally, the abnormal lesion area of ​​the oocyte cytoplasm can be identified by two feature visualization methods, Grad_CAM (Gradient-weighted Class Activation Mapping), which stands for Gradient-weighted Class Activation Mapping, and Grad_CAM++ (which is an enhanced version of Grad_CAM). Grad_CAM stands for Gradient-weighted Class Activation Mapping. This is a method used for feature region identification in neural network classification models, especially convolutional neural networks, to understand which areas of the input image are important for the network to predict a specific class.

[0194] Grad_CAM and Grad_CAM++ will perform gradient calculation on the feature parameters of the last convolutional layer (layer4) of feature recognition in the oocyte abnormality detection model, and then map the results of the gradient calculation to the original image in the form of a thermal image. The high-temperature area indicates the abnormal feature area inferred by the anomaly detection model through training and learning.

[0195] The difference between Grad_CAM++ and Grad_CAM is that they use different methods to calculate feature weights. The highlighted areas of the heatmaps displayed by the two methods may be different. Grad_CAM calculates the weights by taking the global average of the gradients of the feature map. Grad_CAM++ obtains the weights of the feature map by taking the weighted average of the pixel-level gradients. In terms of effect, Grad_CAM++ will provide a more comprehensive visualization of features than Grad_CAM.

[0196] Figure 8 FIG. 1 is a schematic diagram of the marking result of an abnormal region of the oocyte cytoplasm in an embodiment of the present application. By marking the abnormal region using the above-mentioned visualization method, the following can be obtained: Figure 8 The marking results are shown. Figure 8 Shown from top to bottom are the three categories of oocyte cytoplasmic abnormalities, SER, CLCG, and Vacuole, as well as the visualization results of the lesion characteristic areas by the Grad_CAM and Grad_CAM++ methods, Grad_CAM_Map and Grad_CAM++_Map.

[0197] Fig. 9 A flow chart of another method for detecting embryonic developmental abnormalities provided in an embodiment of the present application; Optionally, in step S103, according to the current developmental stage of the target embryo, at least one abnormality detection strategy corresponding to the current developmental stage is adopted to determine the developmental abnormality detection result of the target embryo at the current developmental stage, which may include:

[0198] S901. If the current developmental stage is the cleavage stage, extract the initial developmental timing data of the target embryo in the cleavage stage.

[0199] The initial developmental time series data includes: the current culture time of the target embryo and the developmental stage identifier corresponding to the current culture time.

[0200] When it is identified that the current development stage is in the t2-t9+ stage, that is, when the current development stage is the cleavage stage, the initial developmental timing data of the target embryo during the cleavage process will be collected in real time.

[0201] The initial developmental time series data may include: the current culture time of the target embryo (ie, the culture time after fertilization), and the developmental stage identifier corresponding to the current culture time. The same developmental stage is marked with the same developmental stage identifier.

[0202] by Figure 2 Taking the embryonic development cycle diagram as an example, the horizontal axis represents the embryonic culture time after fertilization, and the vertical axis represents the developmental stage. In the process of embryonic development from t1 (pronuclear stage) to t2 (2-cell stage), the t1 period will last for a period of time, until 26.0 hours, when the embryo truly enters the t2 period. Therefore, the developmental stage labels corresponding to different post-fertilization culture times may be the same.

[0203] S902. Determine the cleavage pattern of the target embryo based on the initial developmental timing data of the cleavage period.

[0204] The initial developmental time series data may be, for example: {t1:x0, t2:x1, ... tM:xn}, where t1-tM may represent developmental stage identifiers, and x0-xn may represent post-fertilization culture time.

[0205] Based on the initial developmental timing data, a series of processing such as timing smoothing and jump point detection can be performed to ultimately determine the cleavage pattern of the target embryo.

[0206] Among them, the cleavage mode can be: normal cleavage, reverse cleavage, direct cleavage, etc.

[0207] S903, using a fragment segmentation network to segment the current embryo image and extract fragment region images.

[0208] In addition to determining the cleavage pattern, the fragments produced during the cleavage of the target embryo can also be extracted to determine if the fragments are abnormal.

[0209] Embryonic fragmentation usually refers to the formation of fragments due to the death of some cells of the embryo due to certain reasons during the embryonic development. Embryonic fragmentation is usually more common in the early stages of pregnancy and is caused by poor quality of the fertilized egg, poor development of the blastomere, and too short a time for the fertilized egg to form. These reasons may cause the embryo to have morphological abnormalities or die during development, forming fragments. If the fragmentation rate exceeds 30%, it is generally considered an abnormal phenomenon. If the fragmentation rate exceeds 50%, it is considered a severe fragmentation rate. If the fragmentation rate is very high, it may lead to serious consequences such as miscarriage.

[0210] In some embodiments, the current embryo image can be segmented by an instance segmentation network MaskRCNN to extract a fragment area image.

[0211] S904. Determine the abnormal fragmentation result of the target embryo according to the fragmentation area image.

[0212] Optionally, based on the extracted fragment area image, data such as the size, quantity, and roundness of the fragments may be analyzed to obtain fragment abnormality results.

[0213] Fig.10 A schematic flow chart of another method for detecting abnormal embryonic development provided in an embodiment of the present application; Optionally, in step S902, determining the cleavage pattern of the target embryo according to the initial developmental timing data of the cleavage period may include:

[0214] S1001. Smoothing the initial developmental time series data to obtain target developmental time series data.

[0215] First, the initial developmental time series data can be smoothed to eliminate the problem of incorrect recognition of embryonic development stages caused by embryo occlusion, imaging abnormalities, etc. After smoothing, the target developmental time series data is obtained.

[0216] S1002: Perform jump point detection on the target development time series data to obtain jump point time series data.

[0217] Next, the target developmental time series data is subjected to jump point detection. A jump point refers to a rising edge or a falling edge in the time series data, that is, detecting when a change in developmental stage occurs in the embryo, that is, for example: when the embryo jumps from the t1 developmental stage to the t2 developmental stage, or when the embryo jumps from the t2 developmental stage to the t3 developmental stage.

[0218] In some embodiments, the obtained jump point timing data may include an identifier of the embryonic development stage at the jump point moment.

[0219] Usually, the embryonic development stage identifier can be represented by numbers, and the t1-tM development stages can correspond to 1-10 respectively. When the embryonic development stage is the t1 stage, the embryonic development stage identifier corresponding to the time series data is 1; when the embryonic development stage is the t2 stage, the embryonic development stage identifier corresponding to the time series data is 2, and so on.

[0220] According to the embryonic development stage identifiers corresponding to two adjacent culture times in the target development time series data, the jump point can be determined and the jump point time series data can be obtained.

[0221] S1003: Perform differential calculation on the jump point timing data to obtain a difference sequence.

[0222] The jumping point time series data implies the cleavage pattern during the embryonic cleavage process. By performing differential calculation on the jumping point time series data, a difference sequence can be obtained.

[0223] S1004. Determine the cleavage pattern of the target embryo based on the difference sequence.

[0224] By analyzing the difference sequence, the cleavage pattern of the target embryo can be determined. When the cleavage pattern is direct cleavage, the specific pattern of direct cleavage can be further determined.

[0225] Fig.11 A schematic flow chart of another method for detecting abnormal embryonic development provided in an embodiment of the present application; Optionally, in step S1001, the initial development time series data is smoothed to obtain target development time series data, which may include:

[0226] S1101. Perform moving smoothing on the initial developmental time series data according to a preset smoothing window size.

[0227] Optionally, the initial developmental time series data may be smoothed using a moving average method according to a preset smoothing window size.

[0228] By calculating the average of data points within a time window in the initial developmental time series data, short-term fluctuations can be eliminated, allowing long-term trends to be better observed.

[0229] S1102, performing a first smoothing process according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, to obtain developmental time series data after the first smoothing process.

[0230] The first smoothing process indicates taking the minimum value of the mark in the smoothing window of each development stage.

[0231] The principle of moving average is roughly as follows:

[0232]

[0233] MA t is the moving average at time point t, and n is the window size

[0234] The selection of the timing smoothing window n is particularly important. If n is set larger, the overall lag of the smooth timing will be caused. If n is set smaller, the effect of the smoothing process will be affected, and there is no way to eliminate the jump with a longer duration. In the present embodiment, the smoothing window can be set in the range of 5 to 7, the smoothing model is set to the center window smoothing, and two smoothing iterations are performed to control the timing lag error within ± 0.5h. This is also considering that the duration of abnormal cleavage is usually greater than 1h. In addition, it is also considered that the jump abnormalities occurring in different time periods are different, such as t1 to t5 periods are prone to rising jumps, that is, the embryonic period of the low stage is identified as the high stage, and t6 to t9+ is prone to falling jumps, that is, the embryonic period of the high stage is identified as the low stage. Therefore, in the timing processing process in the present embodiment, the timing can be divided into two sections according to the period, when t1 to t5 periods are smooth, the sliding window is taken to the minimum value, and the rising edge jump is eliminated. When in the t6 to t9+ period, the sliding window is averaged to eliminate the falling edge jump, and finally the smoothing window is merged to obtain the target development timing data.

[0235] Optionally, for the data segment to which the sliding window currently moves, the minimum value of the developmental stage identifiers in the data in the data segment may be taken, and the minimum value may be used to replace the data located at the center of the data segment.

[0236] S1103, performing a second smoothing process on the developmental time series data after the first smoothing process according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, to obtain target developmental time series data.

[0237] The second smoothing process indicates taking the mean of the marks within the smoothing window of each developmental stage.

[0238] For the developmental time series data after the first smoothing process, the developmental stage identifiers in the data in the data segment to which the sliding window currently slides are averaged, and the average is used to replace the data at the center of the data segment.

[0239] After two smoothing processes, the target development time series data can be obtained, thereby eliminating the rising edge and falling edge jumps.

[0240] Optionally, in step S1004, determining the cleavage pattern of the target embryo according to the difference sequence may include: determining the cleavage type according to the numerical attributes of the elements in the difference sequence, the cleavage type being one of the following: normal cleavage, reverse cleavage or direct cleavage; the numerical attributes including one or more of the following: the positive or negative value, the size of the value, and the position at which the value appears.

[0241] Usually, the normal cleavage pattern should be (t1->t2->t3->t4->t5->t6->t7->t8->t9+), that is, under the normal cleavage pattern, the developmental stage identifiers in the time series data are monotonically increasing.

[0242] Optionally, the difference between two adjacent elements in the jump point time series data can be calculated to obtain a difference sequence.

[0243] In some embodiments, based on the positive and negative values ​​of each value in the difference sequence, when a negative number appears in the sequence, the cleavage type is determined to be reverse cleavage. Based on the size of each value in the difference sequence, when a value greater than 2 appears in the sequence, the cleavage type is determined to be direct cleavage. In the case of direct cleavage, the position of the value greater than 2 in the sequence can be further determined, thereby determining the specific pattern of direct cleavage.

[0244] When all values ​​in the difference sequence are 1, the cleavage pattern is determined to be normal cleavage.

[0245] It is worth noting that the above embodiment is implemented based on the developmental stage identifiers 1-10 corresponding to the set embryonic developmental stages t1-tm. When the representation method of the developmental stage identifiers changes, although the cleavage pattern can still be judged based on the positive and negative values, the size of the values, and the positions where the values ​​appear in the difference sequence, the judgment results will be different.

[0246] In some embodiments, the period when the value greater than 2 appears can be determined based on the position at which the value greater than 2 appears. If the value greater than 2 appears in the t1 period, the cleavage pattern can be determined to be: DUC1, that is, the t1 period directly divides into the t3 period; if the value greater than 2 appears in the t2 period, the cleavage pattern is DUC2, that is, it directly divides from the t2 period to the t5 or t6 period; if the value greater than 2 appears in the t4 period, the cleavage pattern is DUC3, that is, it divides from the t4 period to t9+.

[0247] Fig.12 Schematic diagram of cleavage pattern provided in the embodiment of the present application. RC: indicates that the embryo in the cleavage stage has undergone reverse cleavage, for example, the 2-cell stage has fused into the 1-cell stage; DUC: indicates that the embryo has undergone unequal cleavage, and this cleavage pattern can be subdivided into the following types:

[0248] DUC1: direct cleavage 1->3 (non-diploid cleavage);

[0249] DUC2: 1-2-5 / 6 (asymmetric cleavage at the two-cell stage);

[0250] DUC3: 1-2-4-9 (asymmetric cleavage at the four-cell stage).

[0251] Fig.13A flow chart of another method for detecting abnormal embryonic development provided in an embodiment of the present application; Optionally, in step S904, determining the abnormal fragmentation result of the target embryo according to the fragmentation region image may include:

[0252] S1301: Determine the diameter and roundness of each candidate fragment according to the fragment region image.

[0253] Fig.14 A schematic diagram of a fragment area image adopted in an embodiment of the present application is shown in FIG. Fig.14 The image a in the figure shows the original image of the embryo at the cleavage stage. Fig.14 The area circled by the black line in b is the detected debris area image.

[0254] Usually, the fragments are irregular in shape, but there are also fragments that are similar in shape to blastomeres. In order to distinguish blastomeres from fragments, in this embodiment, the roundness and diameter of all candidate fragments in the fragment area can be calculated.

[0255] The calculation method of the fragment circularity can refer to the calculation formula of the above-mentioned egg cytoplasm circularity.

[0256] S1302. Determine a target fragment from each candidate fragment according to the diameter, roundness and current development stage of each candidate fragment, and determine a fragment proportion result according to the target fragment. The fragment proportion result is used to indicate the proportion of the fragment in the target embryo body.

[0257] Usually, the diameter of the fragments on the second day of embryo culture is less than 45um, and the diameter of the fragments on the third day of embryo culture is less than 40um.

[0258] The candidate fragments can be judged first based on their roundness. When the roundness is greater than 0.75, the diameter of the candidate fragments can be judged. Combined with the developmental stage of the embryo, the candidate fragments that meet the fragment diameter requirements can be used as target fragments.

[0259] Exemplarily, according to the current fertilization and culture time of the embryo, when the fertilization time is less than 45h, and the roundness of the candidate fragment is less than 0.75 and the diameter is less than 45um, the candidate fragment is judged to be the target fragment, otherwise the candidate fragment may be a blastomere or a vacuoles; when the fertilization and culture time is greater than 45h and less than 72h, and the roundness of the candidate fragment is less than 0.75 and the diameter is less than 40um, the candidate fragment is judged to be the target fragment, otherwise the candidate fragment may be a blastomere or a vacuoles.

[0260] Then, the area of ​​the target fragment is determined according to the target fragment, and the proportion of the target fragment in the embryo body is calculated to obtain the proportion result.

[0261] Among them, the embryo image will obtain a binary image of the fragment mask (either 0 or 1) after passing through the segmentation network. The pixels of the non-fragmented area are 0, and the pixels of the fragmented area are 1. The fragment masks are summed, and the sum result is the area of ​​the fragmented area.

[0262] S1303. Determine the abnormal fragment result according to the fragment proportion result.

[0263] Optionally, when the fragmentation ratio exceeds 50%, it can be considered that there is a fragmentation anomaly.

[0264] Fig.15 A flow chart of another method for detecting embryonic developmental abnormalities provided in an embodiment of the present application; Optionally, in step S103, according to the current developmental stage of the target embryo, at least one abnormality detection strategy corresponding to the current developmental stage is adopted to determine the developmental abnormality detection result of the target embryo at the current developmental stage, which may include:

[0265] S1501. If the current developmental stage is the blastocyst stage, extract the time series data of the cavity ratio of the target embryo in the blastocyst stage.

[0266] The cavity ratio time series data includes: the current culture time of the target embryo and the area ratio of the cavity area to the blastocyst main body area corresponding to the current culture time.

[0267] If the current developmental stage is the blastocyst stage, the time series data of the cavity ratio of the target embryo in the blastocyst stage can be extracted.

[0268] Optionally, the blastocyst cavity region and the blastocyst main body region of the target embryo can be segmented using a blastocyst segmentation model, and the cavity ratio time series data can be recorded.

[0269] Among them, the cavity ratio time series data can record the area ratio of the cavity area to the blastocyst main area corresponding to multiple culture times of the target embryo during the entire blastocyst stage. Multiple culture times can be determined according to preset intervals. For example, the area ratio of the cavity area to the blastocyst main area can be recorded every 15 minutes.

[0270] Exemplarily, the cavity ratio time series data may be: {culture time 1: the ratio of the area of ​​the cavity region to the area of ​​the blastocyst body; culture time 2: the ratio of the area of ​​the cavity region to the area of ​​the blastocyst body…culture time n: the ratio of the area of ​​the cavity region to the area of ​​the blastocyst body}.

[0271] Fig.16 A schematic diagram of blastocyst segmentation results provided in an embodiment of the present application, Fig.16 The a in the figure is the original picture of the blastocyst stage embryo. Fig.16Figure b is the segmentation result of the blastocyst segmentation module: ① shows the zona pellucida, ② shows the trophoblast layer, ③ shows the cavity, ④ shows the inner cell mass, and ⑤ shows the main area of ​​the blastocyst.

[0272] S1502. Perform cavity ratio time series decreasing trend detection based on cavity ratio time series data to obtain cavity ratio time series decreasing trend detection results.

[0273] The cavity proportion time series decreasing trend detection result includes: the starting time, the end time and the decreasing amplitude of each cavity proportion time series decreasing trend.

[0274] Optionally, based on the acquired cavity ratio time series data, time series trend detection may be performed, wherein the open source algorithm library Pytrendseries may be used to detect an upward trend or a downward trend of the cavity ratio time series data to obtain a detection result.

[0275] Fig.17 A timing curve diagram provided in an embodiment of the present application is shown in FIG. Fig.17 The time series curve shown in ① is the time series of the area ratio of the cavity area to the blastocyst main area. Fig.17 ② shows the curve of blastocyst area change. During the development of the blastocyst, the cavity expands, and the blastocyst area and the cavity area increase at the same time. When the blastocyst shrinks or collapses, the proportion of the cavity decreases rapidly and lasts for a certain period of time.

[0276] Therefore, this embodiment focuses on detecting the downward trend of the cavity proportion time series to obtain the cavity proportion time series downward trend detection result.

[0277] The cavity proportion temporal decreasing trend detection result may include: the starting time, the end time and the decreasing amplitude of each cavity proportion temporal decreasing trend.

[0278] S1503. Determine the shrinkage result of the target embryo based on the detection result of the time series downward trend of the cavity proportion.

[0279] The shrinkage results include: the number of shrinkages that occurred and the duration of each shrinkage.

[0280] Fig.18 A schematic diagram of a downward trend detection result provided in an embodiment of the present application is shown in FIG. Fig.18 As shown in , the shaded area (gray filled area) indicates the time period when the area ratio of the cavity area to the blastocyst body area is detected to decrease, and this time period corresponds to the start and end time of the blastocyst shrinkage.

[0281] Generally, a downward trend in the ratio of the area of ​​the cavity region to the main area of ​​the blastocyst does not necessarily mean that blastocyst shrinkage has occurred. Further judgment is needed based on the duration of the decrease in the cavity ratio and the magnitude of the decrease in the cavity ratio corresponding to the downward trend in the cavity ratio.

[0282] Fig.19 A flow chart of another method for detecting abnormal embryonic development provided in an embodiment of the present application; Optionally, in step S1503, determining the shrinkage result of the target embryo according to the time series downward trend detection result of the cavity proportion may include:

[0283] S1901. Determine the duration of the current decrease in cavity proportion according to the start time and end time of the current decrease in cavity proportion time series trend.

[0284] Optionally, the duration of each decrease in the cavity proportion timing trend may be determined according to the start time and the end time of each decrease in the cavity proportion timing trend detection result.

[0285] S1902. If the duration of the current decrease in the cavity ratio meets the preset time, and the decrease in the current cavity ratio time series trend meets the preset range, then the number of shrinkages of the target embryo is increased by 1, and the duration of the current decrease is recorded.

[0286] The current time may be any time of cavity ratio time series trend decrease in the cavity ratio time series decrease trend detection result. When the cavity ratio decrease lasts for more than 1.5 hours and the decrease is greater than 25%, it can be considered that a blastocyst shrinkage has occurred.

[0287] Then, based on the specific data of each decrease in the cavity ratio time series trend in the cavity ratio time series decrease trend detection results, it can be determined which cavity ratio time series trend decreases represent actual blastocyst shrinkage, thereby ultimately determining the total number of shrinkages, and taking the duration of the cavity ratio decrease corresponding to the shrinkage as the shrinkage duration.

[0288] In some embodiments, the present application also provides a function of visually displaying abnormal detection results. The detection results of abnormal detection of embryos using the above method can be associated with the embryos and stored in a database. When the user needs to view the abnormal detection results of the embryos, the abnormal results can be visualized and confirmed through the client software.

[0289] The abnormal prompt information that can be displayed in this program includes but is not limited to: the number of abnormalities that occur in the embryo; whether the embryo has been manually reviewed; abnormal event identification at the corresponding time on the timeline; picture display of abnormal embryos; a list of all abnormal event levels of the current embryo; detailed content of the current abnormal event, including the abnormal level, time and detailed description, etc.

[0290] Generally, the morphology or structure of the cytoplasm affects the development of oocytes, fertilization, and later embryo quality and pregnancy outcomes. Abnormal cleavage in the early stage of the preimplantation embryo will affect the formation and quality of the blastocyst, and the earlier the abnormal cleavage occurs, the greater the impact on subsequent embryo development. Abnormal fragmentation means that there are a large number of fragments in the embryo during the cleavage period. When the amount of fragments is high, it means that the embryo's developmental potential, morphology and quality are poor. Not only is the transplantation success rate low, it will also increase the chance of embryonic chromosomal abnormalities and affect pregnancy outcomes. Blastocyst shrinkage may be accompanied by problems such as reduced implantation ability, genetic abnormalities and adverse pregnancy outcomes.

[0291] Based on the abnormality detection method of this scheme, all abnormal phenomena that may affect the developmental outcomes of embryos in various developmental periods can be detected. Therefore, combined with the abnormality detection results in each developmental period, a comprehensive and accurate assessment of the developmental potential of the embryo before implantation can be made, so that embryos of poor quality can be eliminated in a timely manner.

[0292] In summary, the embryonic development abnormality detection method provided by this embodiment includes: collecting the current embryonic image of the target embryo during the embryonic development cycle; determining the current development stage of the target embryo according to the current embryonic image; according to the current development stage of the target embryo, using at least one abnormality detection strategy corresponding to the current development stage to determine the development abnormality detection result of the target embryo at the current development stage; according to the development abnormality detection result of the target embryo in each specified development stage, determine the development result of the target embryo. This scheme monitors the embryonic development cycle to collect embryonic images at different developmental stages during the embryonic development process, and uses abnormality detection strategies corresponding to each developmental stage to identify the developmental abnormalities of the embryo at different developmental stages, so that the abnormal development information of the embryo at different developmental stages can be integrated to comprehensively and reliably evaluate the developmental potential of the embryo.

[0293] The following describes the apparatus, device, storage medium, etc. used to execute the method for detecting abnormal embryonic development provided by the present application. The specific implementation process and technical effects are described above and will not be repeated below.

[0294] Fig. 20A schematic diagram of an embryonic development abnormality detection device provided in an embodiment of the present application, the functions implemented by the embryonic development abnormality detection device correspond to the steps performed by the above method. The device can be understood as the above-mentioned computer device, or server, or server processor, or can also be understood as a component independent of the above-mentioned server or processor that implements the functions of the present application under the control of the server, such as Fig. 20 As shown, the device may include: a collection module 100 and a determination module 200;

[0295] The acquisition module 100 is used to acquire the current embryo image of the target embryo during the embryo development cycle;

[0296] A determination module 200 is used to determine the current developmental stage of the target embryo based on the current embryo image;

[0297] A determination module 200 is used to determine the developmental abnormality detection result of the target embryo at the current developmental stage by using at least one abnormality detection strategy corresponding to the current developmental stage according to the current developmental stage of the target embryo;

[0298] The determination module 200 is used to determine the developmental result of the target embryo according to the developmental abnormality detection results of the target embryo in each designated developmental stage; the designated developmental stages include: pronuclear stage, cleavage stage and blastocyst stage.

[0299] Optionally, the determination module 200 is specifically configured to segment the current embryo image using an egg cytoplasm segmentation network to extract an egg cytoplasm image if the current development stage is the pronuclear stage;

[0300] According to the egg cytoplasm image, the roundness of the egg cytoplasm is determined, and according to the roundness of the egg cytoplasm, the abnormal oval shape result of the egg cytoplasm is determined;

[0301] According to the oocyte cytoplasm image, the gap distance between the oocyte cytoplasm and the zona pellucida is determined, and according to the gap distance, the abnormal result of the perivitelline space of the oocyte cytoplasm is determined;

[0302] According to the egg cytoplasm image, an egg cytoplasm abnormality detection model is used to determine the egg cytoplasm abnormality detection result, and the abnormality detection result is used to indicate whether the egg cytoplasm is abnormal and the corresponding abnormality type.

[0303] Optionally, the determination module 200 is specifically configured to segment the zona pellucida of the target embryo using a blastocyst segmentation model, and generate a mask image of the embryo body based on the outer contour of the zona pellucida;

[0304] Extracting an oocyte cytoplasm mask image according to the oocyte cytoplasm image;

[0305] Determine a difference image of the target embryo according to the embryo main body mask image and the egg cytoplasm mask image;

[0306] Marking the connected domain of the difference image and determining the mask image of the perivitelline space area;

[0307] The gap distance between the oocyte cytoplasm and the zona pellucida was determined based on the mask image of the perivitelline space area.

[0308] Optionally, the determination module 200 is specifically configured to use a first classification model in the egg cytoplasm abnormality detection model to perform a first detection on the egg cytoplasm image and determine a first detection result, wherein the first detection result is used to indicate whether there is an abnormality in the egg cytoplasm;

[0309] If there is an abnormality, a second classification model in the egg cytoplasm abnormality detection model is used to perform a second detection on the egg cytoplasm image to determine a second detection result, and the second detection result is used to indicate the abnormal type of the egg cytoplasm;

[0310] Based on the first test result and the second test result, an abnormal test result of the egg cytoplasm is determined.

[0311] Optionally, the determination module 200 is specifically used to extract the initial developmental timing data of the target embryo in the cleavage stage if the current developmental stage is the cleavage stage, the initial developmental timing data including: the current culture time of the target embryo and the developmental stage identifier corresponding to the current culture time;

[0312] Determine the cleavage pattern of the target embryo based on the initial developmental timing data of the cleavage stage;

[0313] The fragment segmentation network is used to segment the current embryo image and extract the fragment area image;

[0314] Based on the fragmentation area image, the fragmentation abnormality results of the target embryo are determined.

[0315] Optionally, the determination module 200 is specifically used to perform smoothing processing on the initial development time series data to obtain target development time series data;

[0316] Perform jump point detection on the target development time series data to obtain jump point time series data;

[0317] Perform differential calculation on the jump point time series data to obtain a difference sequence;

[0318] Based on the difference sequence, the cleavage pattern of the target embryo is determined.

[0319] Optionally, the determination module 200 is specifically used to perform moving smoothing on the initial development time series data according to a preset smoothing window size;

[0320] According to the developmental stage identifiers corresponding to the current culture time within the smoothing range currently moved to, a first smoothing process is performed to obtain the developmental time series data after the first smoothing process; the first smoothing process indicates taking the minimum value of the identifiers within the smoothing window of each developmental stage;

[0321] For the developmental time series data after the first smoothing process, a second smoothing process is performed according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, to obtain the target developmental time series data. The second smoothing process indicates taking the average of the identifiers within the smoothing window of each developmental stage.

[0322] Optionally, the determination module 200 is specifically used to determine the cleavage type according to the numerical attributes of the elements in the difference sequence, and the cleavage type is one of the following: normal cleavage, reverse cleavage or direct cleavage; the numerical attributes include one or more of the following: positive or negative value, value size, and position where the value appears.

[0323] Optionally, the determination module 200 is specifically configured to determine the diameter and roundness of each candidate fragment according to the fragment region image;

[0324] Determine the target fragment from each candidate fragment according to the diameter, roundness and current development stage of each candidate fragment, and determine the fragment proportion result according to the target fragment, and the fragment proportion result is used to indicate the proportion of the fragment in the target embryo body;

[0325] According to the fragment ratio results, the abnormal fragment results are determined.

[0326] Optionally, the determination module 200 is specifically used to extract the time series data of the cavity ratio of the target embryo in the blastocyst stage if the current development stage is the blastocyst stage, and the time series data of the cavity ratio includes: the current culture time of the target embryo and the area ratio of the cavity area to the blastocyst main area corresponding to the current culture time;

[0327] According to the cavity proportion time series data, a cavity proportion time series downward trend detection is performed to obtain a cavity proportion time series downward trend detection result, and the cavity proportion time series downward trend detection result includes: the start time, the end time and the decline amplitude of each cavity proportion time series trend decline;

[0328] The shrinkage result of the target embryo is determined according to the detection result of the time series downward trend of the cavity proportion, and the shrinkage result includes: the number of shrinkage occurrences and the duration of each shrinkage.

[0329] Optionally, the determination module 200 is specifically used to determine the duration of the current cavity proportion decrease according to the start time and end time of the current cavity proportion time series trend decrease;

[0330] If the duration of the current decrease in the cavity ratio meets the preset time, and the decrease in the current cavity ratio time series trend meets the preset range, the number of shrinkages of the target embryo is increased by 1, and the duration of the current decrease in the cavity ratio is recorded.

[0331] The above-mentioned device is used to execute the method provided by the aforementioned embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0332] The above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more digital singnal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0333] The above modules can be connected or communicated with each other via a wired connection or a wireless connection. The wired connection may include a metal cable, an optical cable, a hybrid cable, etc., or any combination thereof. The wireless connection may include a connection in the form of a LAN, a WAN, Bluetooth, a ZigBee, or NFC, or any combination thereof. Two or more modules may be combined into a single module, and any one module may be divided into two or more units. Those skilled in the art will clearly understand that for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the method embodiment, and will not be repeated in this application.

[0334] Fig.21 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application, which may be a computing device with data processing capabilities.

[0335] The device includes: a processor 801 and a storage medium 802 .

[0336] The storage medium 802 is used to store programs, and the processor 801 calls the programs stored in the storage medium 802 to execute the above method embodiment. The specific implementation method and technical effect are similar and will not be repeated here.

[0337] Among them, the storage medium 802 stores program code, and when the program code is executed by the processor 801, the processor 801 executes the various steps of the embryonic development abnormality detection method according to various exemplary embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0338] Processor 801 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor (DigitalSignal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, and can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present application can be directly embodied as a hardware processor to be executed, or a combination of hardware and software modules in the processor can be executed.

[0339] Storage medium 802, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules. Storage medium may include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type storage medium, random access storage medium (Random Access Memory, RAM), static random access storage medium (Static Random Access Memory, SRAM), programmable read-only storage medium (Programmable Read Only Memory, PROM), read-only storage medium (Read Only Memory, ROM), electrically erasable programmable read-only storage medium (Electrically Erasable Programmable Read-Only Memory, EEPROM), magnetic storage medium, disk, optical disk, etc. Storage medium is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The storage medium 802 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0340] Optionally, the present application also provides a program product, such as a computer-readable storage medium, including a program, which is used to execute the above method embodiment when executed by a processor.

[0341] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0342] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0343] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0344] The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a computer-readable storage medium. The above-mentioned software functional unit is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (English: processor) to perform some steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only storage medium (English: Read-Only Memory, referred to as: ROM), random access storage medium (English: Random Access Memory, referred to as: RAM), disk or optical disk and other media that can store program code.

Claims

1. A method for detecting abnormal embryonic development, characterized in that: include: During the embryonic development cycle, a current embryonic image of the target embryo is acquired; Determining the current developmental stage of the target embryo according to the current embryo image; According to the current developmental stage of the target embryo, at least one abnormality detection strategy corresponding to the current developmental stage is adopted to determine a developmental abnormality detection result of the target embryo at the current developmental stage; Determining the developmental outcome of the target embryo according to the developmental abnormality detection results of the target embryo in each designated developmental stage; The designated developmental stages include: pronuclear stage, cleavage stage and blastocyst stage.

2. The method according to claim 1, characterized in that The step of determining the developmental abnormality detection result of the target embryo at the current developmental stage by using at least one abnormality detection strategy corresponding to the current developmental stage according to the current developmental stage of the target embryo comprises: If the current development stage is the pronuclear stage, segmenting the current embryo image using an egg cytoplasm segmentation network to extract an egg cytoplasm image; Determining the roundness of the egg cytoplasm according to the egg cytoplasm image, and determining the oval abnormality result of the egg cytoplasm according to the roundness of the egg cytoplasm; Determine the gap distance between the egg cytoplasm and the zona pellucida according to the egg cytoplasm image, and determine the abnormal result of the perivitelline space of the egg cytoplasm according to the gap distance; According to the egg cytoplasm image, an egg cytoplasm abnormality detection model is used to determine the abnormality detection result of the egg cytoplasm, and the abnormality detection result is used to indicate whether the egg cytoplasm is abnormal and the corresponding abnormality type.

3. The method according to claim 2, characterized in that Determining the gap distance between the egg cytoplasm and the zona pellucida according to the egg cytoplasm image includes: The blastocyst segmentation model is used to segment the zona pellucida of the target embryo, and a mask image of the embryo body is generated based on the outer contour of the zona pellucida; Extracting the egg cytoplasm mask image according to the egg cytoplasm image; Determining a difference image of the target embryo according to the embryo main body mask image and the egg cytoplasm mask image; Marking the connected domain of the difference image to determine the mask image of the perivitelline space area; The gap distance between the egg cytoplasm and the zona pellucida is determined according to the mask image of the perivitelline space area.

4. The method according to claim 2, characterized in that: The method of using an egg cytoplasm abnormality detection model to determine an abnormality detection result of the egg cytoplasm includes: Using a first classification model in the egg cytoplasm abnormality detection model, performing a first detection on the egg cytoplasm image to determine a first detection result, wherein the first detection result is used to indicate whether the egg cytoplasm has an abnormality; If an abnormality is present, a second classification model in the egg cytoplasm abnormality detection model is used to perform a second detection on the egg cytoplasm image to determine a second detection result, wherein the second detection result is used to indicate the abnormal type of the egg cytoplasm; The abnormal detection result of the egg cytoplasm is determined according to the first detection result and the second detection result.

5. The method according to claim 1, characterized in that The step of determining the developmental abnormality detection result of the target embryo at the current developmental stage by using at least one abnormality detection strategy corresponding to the current developmental stage according to the current developmental stage of the target embryo comprises: If the current development stage is the cleavage stage, extracting the initial developmental timing data of the target embryo in the cleavage stage, the initial developmental timing data including: the current culture time of the target embryo and the developmental stage identifier corresponding to the current culture time; Determining the cleavage pattern of the target embryo according to the initial developmental timing data of the cleavage stage; Using a fragment segmentation network to segment the current embryo image and extract a fragment region image; The abnormal fragmentation result of the target embryo is determined according to the fragmentation area image.

6. The method according to claim 5, characterized in that Determining the cleavage pattern of the target embryo according to the initial developmental time series data of the cleavage period comprises: Performing smoothing on the initial developmental time series data to obtain target developmental time series data; Performing jump point detection on the target development time series data to obtain jump point time series data; Performing differential calculation on the jump point timing data to obtain a difference value sequence; The cleavage pattern of the target embryo is determined based on the difference sequence.

7. The method according to claim 6, characterized in that The step of smoothing the initial developmental time series data to obtain target developmental time series data includes: According to a preset smoothing window size, the initial developmental time series data is subjected to moving smoothing; Perform a first smoothing process according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, to obtain developmental time series data after the first smoothing process; the first smoothing process indicates taking the minimum value of the identifiers within the smoothing window of each developmental stage; For the developmental time series data after the first smoothing processing, a second smoothing processing is performed according to the developmental stage identifiers corresponding to the current culture times within the smoothing range currently moved to, so as to obtain the target developmental time series data. The second smoothing processing indicates taking the average of the identifiers within the smoothing window of each developmental stage.

8. The method according to claim 6, characterized in that Determining the cleavage pattern of the target embryo according to the difference sequence includes: According to the numerical attributes of the elements in the difference sequence, the cleavage type is determined, and the cleavage type is one of the following: normal cleavage, reverse cleavage or direct cleavage; the numerical attributes include one or more of the following: positive or negative value, value size, and position where the value appears.

9. The method according to claim 5, characterized in that Determining the abnormal fragmentation result of the target embryo according to the fragmentation region image includes: Determining the diameter and roundness of each candidate fragment according to the fragment region image; Determine a target fragment from each candidate fragment according to the diameter, roundness and current development stage of each candidate fragment, and determine a fragment proportion result according to the target fragment, wherein the fragment proportion result is used to indicate the proportion of the fragment in the target embryo body; According to the fragment proportion result, the fragment abnormality result is determined.

10. The method according to claim 1, characterized in that The step of determining the developmental abnormality detection result of the target embryo at the current developmental stage by using at least one abnormality detection strategy corresponding to the current developmental stage according to the current developmental stage of the target embryo comprises: If the current development stage is the blastocyst stage, extract the time series data of the cavity ratio of the target embryo in the blastocyst stage, wherein the time series data of the cavity ratio includes: the current culture time of the target embryo and the area ratio of the cavity area to the blastocyst main area corresponding to the current culture time; According to the cavity proportion time series data, a time series downward trend detection is performed to obtain a cavity proportion time series downward trend detection result, wherein the cavity proportion time series downward trend detection result includes: the start time, the end time and the decline amplitude of each time when the cavity proportion time series trend declines; The shrinkage result of the target embryo is determined according to the detection result of the time series downward trend of the cavity proportion, and the shrinkage result includes: the number of shrinkage occurrences and the duration of each shrinkage.

11. The method according to claim 10, characterized in that Determining the shrinkage result of the target embryo according to the detection result of the time series downward trend of the cavity proportion includes: Determine the duration of the current cavity ratio decrease according to the start time and end time of the current cavity ratio decrease time series trend; If the duration of the current decrease in the cavity proportion meets the preset time, and the decrease amplitude of the current cavity proportion time series trend meets the preset amplitude, the number of shrinkages of the target embryo is increased by 1, and the duration of the current decrease in the cavity proportion is recorded.

12. An electronic device, characterized in that: include: A processor, a storage medium and a bus, wherein the storage medium stores program instructions executable by the processor. When the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the program instructions to implement the method for detecting abnormal embryonic development as described in any one of claims 1 to 11.

Images