Method, device, electronic device and medium for detecting abnormal chromosome events

Abstract

The embodiment of the present invention discloses a method, device, electronic device and storage medium for detecting abnormal chromosome events. The method comprises: obtaining a historical image of at least one historical moment in the target embryo's division process; inputting each historical image into a pre-established generation network to generate a predicted image of a future moment in the target embryo's division process relative to the historical moment; obtaining an actual image corresponding to the target embryo's division at the future moment, and determining whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generation network and the discriminant network include a neural network; if not, determining that a chromosomal abnormal event occurs in the target embryo's division process at a future moment. The technical solution of the embodiment of the present invention determines whether there is a chromosomal abnormal event by judging whether the predicted image is consistent with the actual image, without manual participation, thereby improving the efficiency and accuracy of detecting abnormal events.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of image recognition technology, and in particular to a method, device, electronic device and medium for detecting chromosomal abnormality events. Background Art

[0002] In the field of assisted reproductive technology (In Vitro Fertilization, IVF), the Time-lapse system can capture abnormal chromosomal nuclear structures during embryo division. In a normal primary nucleus, a single, relatively large nucleus with a complete nuclear membrane can be observed. Several different types of nuclear defects can occur in embryos, such as multinuclei, binuclei, micronuclei, and anucleus.

[0003] In order to help embryo research experts conduct biological research based on the time point and degree of abnormality of the cell nucleus, it is necessary to detect the chromosome images of the embryo division process captured by the time-lapse system to determine whether there are chromosomal abnormalities in the chromosome images. In the prior art, it is necessary to rely on experienced reproductive doctors to manually check the chromosome images to determine whether chromosomal abnormalities have occurred; this method has low detection efficiency and is prone to errors, so how to efficiently and accurately detect abnormal nuclei becomes increasingly important. Summary of the invention

[0004] The embodiments of the present invention provide a method, device, electronic device and storage medium for detecting chromosomal abnormality events, so as to determine whether a chromosomal abnormality event exists by judging whether a predicted image is consistent with an actual image, without the need for human intervention, thereby improving the efficiency and accuracy of detecting abnormal events.

[0005] In a first aspect, an embodiment of the present invention provides a method for detecting a chromosome abnormality event, comprising:

[0006] Acquire a historical image of at least one historical moment in the division process of the target embryo;

[0007] Inputting each of the historical images into a pre-established generation network to generate a predicted image of a future moment in the target embryo division process relative to the historical moment;

[0008] Acquire an actual image corresponding to the target embryo splitting at the future moment, and determine whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generating network and the discriminant network include neural networks;

[0009] If not, it is determined that a chromosomal abnormality event occurs during the division of the target embryo at the future time.

[0010] In a second aspect, an embodiment of the present invention further provides a device for detecting a chromosome abnormality event, the device comprising:

[0011] A historical image acquisition module is used to acquire a historical image of at least one historical moment in the cleavage process of the target embryo;

[0012] A prediction image generation module is used to input each of the historical images into a pre-established generation network to generate a prediction image of a future moment in the target embryo division process relative to the historical moment;

[0013] An actual image acquisition module is used to acquire an actual image corresponding to the target embryo splitting at the future moment, and determine whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generating network and the discriminant network include a neural network; if not, enter an abnormal event determination module;

[0014] The abnormal event determination module is used to determine the occurrence of a chromosomal abnormality event during the division of the target embryo at the future moment.

[0015] In a third aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0016] one or more processors;

[0017] a storage device for storing one or more programs,

[0018] When the one or more programs are executed by the one or more processors, the one or more processors implement the method for detecting chromosomal abnormality events provided by any embodiment of the present invention.

[0019] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for detecting chromosomal abnormality events provided by any embodiment of the present invention.

[0020] A method for detecting abnormal chromosome events provided by an embodiment of the present invention obtains historical images of at least one historical moment in the target embryo's division process, inputs each historical image into a pre-established generation network, and generates a predicted image of a future moment in the target embryo's division process relative to the historical moment; the predicted image can reflect the normal embryo division image under normal division conditions; obtains the actual image corresponding to the target embryo's division at the future moment, and determines whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generation network and the discriminant network include a neural network; if not, it is determined that a chromosomal abnormal event occurs in the target embryo's division process at the future moment. The embodiment of the present invention compares the predicted image with the actual image, that is, compares the normal embryo division image with the actual image, and determines whether the two are consistent to determine whether there is a chromosomal abnormal event, without the need for manual participation, thereby improving the efficiency and accuracy of detecting abnormal events.

[0021] In addition, a chromosome abnormality event detection device, electronic device and storage medium provided by the present invention correspond to the above method and have the same beneficial effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 A flowchart of a method for detecting a chromosome abnormality event provided by an embodiment of the present invention;

[0024] Figure 2 A flowchart of another method for detecting chromosome abnormality events provided by an embodiment of the present invention;

[0025] Figure 3 A structural diagram of a device for detecting chromosome abnormality events provided by an embodiment of the present invention;

[0026] Figure 4 A structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only parts related to the present invention, rather than all structures, are shown in the accompanying drawings.

[0028] It should also be noted that, for ease of description, only the part relevant to the present invention but not all content is shown in the accompanying drawings. It should be mentioned before discussing exemplary embodiments in more detail that some exemplary embodiments are described as processing or methods depicted as flow charts. Although the flow chart describes each operation (or step) as sequential processing, many operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of each operation can be rearranged. When its operation is completed, the processing can be terminated, but it can also have additional steps not included in the accompanying drawings. The processing can correspond to methods, functions, procedures, subroutines, subprograms, etc.

[0029] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0030] Embodiment 1

[0031] Figure 1 A flowchart of a chromosome abnormality event detection provided by an embodiment of the present invention. The method can be executed by a chromosome abnormality event detection device, which can be implemented by software and / or hardware and can be configured in a terminal and / or server to implement the chromosome abnormality event detection method in the embodiment of the present invention.

[0032] like Figure 1 As shown, the method of this embodiment may specifically include:

[0033] S101. Acquire a historical image of at least one historical moment in the cleavage process of a target embryo.

[0034] In a specific implementation, historical images of the target embryo at historical moments in the process of division can be obtained, and the embryo can be observed through a monochromatic light source. Exemplarily, historical images taken at t consecutive historical moments can be obtained, where t is a positive integer greater than 1. For each shooting moment, 11 images of different focal planes can be collected to ensure the accuracy and comprehensiveness of the collected historical images; the height resolution and width resolution of the historical images can be 1024 respectively.

[0035] S102, inputting each historical image into a pre-established generation network to generate a predicted image of a future moment in the target embryo division process relative to the historical moment.

[0036] Specifically, a generative network for generating predicted images can be pre-established, and the generative network includes a neural network. Exemplarily, the generative network is a U-Net network. Further, the neural network can be trained by using historical sample images at various historical moments and future sample images at future moments during the pre-established splitting process of the sample embryo to establish the generative network.

[0037] Optionally, each historical image is input into a pre-established generation network to generate a predicted image of a future moment in the target embryo division process relative to the historical moment, including: grayscale processing each historical image to generate a corresponding grayscale image; each grayscale image is input into a pre-established generation network, convolution calculation, downsampling calculation, deconvolution calculation and image jump connection processing are performed on each grayscale image, and the image obtained after processing is determined as the predicted image.

[0038] Specifically, in order to facilitate the generation network to analyze and process historical images, each historical image can be gray-scaled to generate a corresponding gray-scale image. The generation network can perform convolution calculation, downsampling calculation, deconvolution calculation and image jump connection processing on the gray-scale image, and determine the processed image as a predicted image of the future moment corresponding to the historical moment.

[0039] It should be noted that the future moment can be the next moment after the historical moment, and those skilled in the art can also determine the future moment as a moment after the historical moment, N moments apart, N being a positive integer greater than 1, and this embodiment of the present invention is not limited to this. The predicted image can reflect the embryonic state during the target embryo splitting and development process according to the current target embryo's splitting trend at each historical moment under normal circumstances.

[0040] S103, obtaining the actual image corresponding to the target embryo division at the future moment, and determining whether the actual image is consistent with the predicted image based on the pre-established discriminant network; if not, proceeding to S104.

[0041] Specifically, the actual image actually produced by the target embryo at a future moment in the division process and development process can be obtained; the actual image is the image that needs to be detected whether there is a chromosomal abnormality event. Further, a discriminant network can be established in advance, and the discriminant network can be used to determine whether the actual image and the predicted image are consistent. Among them, the discriminant network includes a neural network for realizing classification, and the discriminant network performs convolution calculation, downsampling operation and full connection operation on the input actual image and the predicted image, and finally determines whether the two are consistent through the softmax activation function. Exemplarily, the discriminant network includes a deep residual network ResNet and a deep convolutional network VGG.

[0042] Furthermore, if the actual image is consistent with the predicted image, it means that the embryo division reflected by the actual image is normal and no chromosomal abnormality occurs.

[0043] S104. Determine the occurrence of chromosomal abnormality events during the division of the target embryo at a future time.

[0044] Optionally, if the actual image is inconsistent with the predicted image, it means that the embryo division reflected by the actual image is abnormal, and the abnormal situation can be fed back to facilitate research experts to study. Further, after determining that a chromosomal abnormality event occurs in the target embryo division process at a future time, it also includes: in the actual image corresponding to the target embryo division at the future time, marking the position that is inconsistent with the predicted image.

[0045] Specifically, based on the discriminant results of the discriminant network, the image position where the actual image is different from the predicted image is determined, and the image position is marked. Exemplarily, the marking method can be a method of selecting a marking box and a method of displaying different colors, so that the research expert can intuitively and clearly determine the abnormal situation in the actual image and quickly find the problem.

[0046] A method for detecting abnormal chromosome events provided by an embodiment of the present invention obtains historical images of at least one historical moment in the target embryo's division process, inputs each historical image into a pre-established generation network, and generates a predicted image of a future moment in the target embryo's division process relative to the historical moment; the predicted image can reflect the normal embryo division image under normal division conditions; obtains the actual image corresponding to the target embryo's division at the future moment, and determines whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generation network and the discriminant network include a neural network; if not, it is determined that a chromosomal abnormal event occurs in the target embryo's division process at the future moment. The embodiment of the present invention compares the predicted image with the actual image, that is, compares the normal embryo division image with the actual image, and determines whether the two are consistent to determine whether there is a chromosomal abnormal event, without the need for manual participation, thereby improving the efficiency and accuracy of detecting abnormal events.

[0047] Embodiment 2

[0048] Figure 2 A flowchart of another method for detecting chromosomal abnormality events provided by an embodiment of the present invention. This embodiment is optimized based on the above-mentioned technical solutions. Optionally, before inputting each historical image into a pre-established generation network, it also includes: determining a loss function generated by the target embryo splitting process, and determining generation parameters of the predicted image based on the loss function; and establishing the generation network based on the generation parameters and the neural network structure. The explanations of the terms that are the same or corresponding to the above-mentioned embodiments are not repeated here.

[0049] S201. Acquire a historical image of at least one historical moment in the cleavage process of the target embryo.

[0050] S202, determining a loss function generated by the target embryo division process, and determining generation parameters of the predicted image based on the loss function; and establishing a generation network based on the generation parameters and the neural network structure.

[0051] In the specific implementation, in order to ensure that the accuracy of the predicted image is high, it is necessary to ensure that the established generation network has the characteristic of low loss. When establishing the generation network, the loss function generated by the target embryo division process can be determined, and the generation parameters corresponding to the minimum loss function value are determined as the generation parameters of the predicted image.

[0052] Optionally, determining a loss function generated by the target embryo division process includes: determining an adversarial loss function, a spatial loss function, and a temporal loss function generated by the target embryo division process.

[0053] Specifically, the loss function generated by the target embryo splitting process includes an adversarial loss function, a spatial loss function, and a temporal loss function, wherein the spatial loss function includes a pixel loss function and a gradient loss function. The adversarial loss function, the spatial loss function, and the temporal loss function can be determined respectively.

[0054] Optionally, the generation parameters of the predicted image are determined based on the loss function, including: determining the weight coefficients corresponding to the adversarial loss function, the pixel loss function, the gradient loss function and the time loss function; based on the weight coefficients, performing a weighted sum calculation on the adversarial loss function, the pixel loss function, the gradient loss function and the time loss function, and determining the calculation result as the total loss generated by the target embryo division process; and generating parameters corresponding to the generation network of the predicted image when the total loss is minimized based on the gradient descent method.

[0055] Specifically, adversarial loss, pixel loss, gradient loss and time loss have different effects on the accuracy of establishing a generated network. Based on the magnitude of the impact, the weight coefficients corresponding to the adversarial loss function, pixel loss function, gradient loss function and time loss function can be set accordingly. The larger the weight coefficient, the greater the impact on the accuracy of the generated network. Those skilled in the art can set the weight coefficient according to the actual application situation, and the embodiment of the present invention does not limit this. Exemplarily, the sum of the weight coefficients corresponding to the adversarial loss function, pixel loss function, gradient loss function and time loss function can be 1.

[0056] Furthermore, the functional relationship between the actual image and the predicted image can be determined from the aspects of adversarial loss, pixel loss, gradient loss and time loss, respectively, to form adversarial loss function, pixel loss function, gradient loss function and time loss function; based on each weight coefficient, the adversarial loss function, pixel loss function, gradient loss function and time loss function are weighted and summed, and the calculation result is determined as the total loss generated by the target embryo division process. Based on the gradient descent method, the corresponding generation parameters of the generation network can be determined to minimize the total loss, and the generation network can be constructed based on the generation parameters.

[0057] Exemplarily, the total loss is calculated as follows:

[0058]

[0059] in, represents the total loss, λ int , gd , op , adv They are the weight coefficients of pixel loss function, gradient loss function, time loss function, and adversarial loss function respectively; is the predicted image at time t+1, I t+1 The actual image at time t+1, L int , L gd , L op , They are pixel loss function, gradient loss function, time loss function, and adversarial loss function respectively.

[0060] Optionally, determining the time loss function generated by the target embryo division process includes: determining the equivalent function of the regularization network, obtaining the actual image corresponding to the previous moment of the future moment; based on the equivalent function, calculating the first equivalent function value between the predicted image and the actual image corresponding to the previous moment, and calculating the second equivalent function value between the actual image at the future moment and the actual image at the previous moment; calculating the norm of the difference between the first equivalent function value and the second equivalent function value, and determining the calculation result as the time loss function.

[0061] Specifically, in order to fully consider the temporal dimension correlation in the target embryo division process, a regularization network can be introduced to measure the temporal dimension difference between the predicted image at the future moment and the actual image at the historical moment, and between the actual image at the future moment and the actual image at the historical moment. Exemplarily, the regularization network includes an optical flow network, the predicted image can be used as the calculated synthetic optical flow in the optical flow network, and the actual image can be used as the real optical flow in the optical flow network. The formula of the time loss function is as follows:

[0062]

[0063] Among them, L op represents the time loss function, Represents the predicted image at time t+1, I t represents the actual image at time t, I t+1 represents the actual image at time t+1, represents the first equivalent function value of the regularized network between the predicted image at time t+1 and the actual image at time t, f(I t+1 ,I t ) represents the second equivalent function value of the regularized network between the actual image at time t+1 and the actual image at time t; L op equal With f(I t+1 ,I t ) is the 1-norm of the difference between .

[0064] Specifically, the calculation formula of the pixel loss function is as follows:

[0065]

[0066] in, The pixel loss function is the 2-norm of the predicted image at time t+1 and the actual image at time t+1; The smaller the value obtained, the closer the pixel amplitudes of the predicted image and the actual image are.

[0067] Specifically, the calculation formula of the gradient loss function is as follows:

[0068]

[0069] in, is the gradient loss function of the predicted image at time t+1 and the actual image at time t+1, which represents the sum of the L1 norms of the absolute values ​​of the gradient differences between the predicted image and the actual image in the horizontal and vertical directions respectively; and The same is the predicted image at time t+1; I and I t+1 The same is the actual image at time t+1; i represents the horizontal axis pixel position in the image, and j represents the vertical axis pixel position in the image; and are the grayscale values ​​at the pixel positions (i, j), (i-1, j), and (i, j-1) in the predicted image at time t+1 respectively; I i,j ,I i-1,j ,I i,j-1They are the grayscale values ​​at the pixel positions (i, j), (i-1, j), and (i, j-1) in the actual image at time t+1. It should be noted that the gradient loss function can measure the detail difference between the actual image and the predicted image. The smaller the calculation result, the smaller the difference between the two. The spatial loss function is determined by the pixel loss function and the gradient loss function.

[0070] Specifically, the calculation formula of the adversarial loss function is as follows:

[0071]

[0072] Among them, L MSE represents the mean square error, The variable is The function mapping of the discriminant network has a value range of {0,1}; i represents the horizontal axis pixel position in the image, and j represents the vertical axis pixel position in the image; and The same is the predicted image at time t+1. The adversarial loss is used to optimize the generative network so that the predicted image generated by the generative network is more accurate.

[0073] S203, inputting each historical image into a pre-established generation network to generate a predicted image of a future moment in the target embryo division process relative to the historical moment.

[0074] S204, obtaining the actual image corresponding to the target embryo division at the future moment, and determining whether the actual image is consistent with the predicted image based on the pre-established discriminant network; if not, proceeding to S205.

[0075] Specifically, before the predicted image and the actual image are input into the discriminant network, a loss function for optimizing the discriminant network can be determined, and the discriminant network is established based on the loss function; the formula of the loss function generated by the discriminant process is as follows:

[0076]

[0077] in, is the loss function used to optimize the discriminant network, L MSE represents the mean square error, is the function mapping of the discriminant network with variable I, The variable is The function mapping of the discriminant network.

[0078] Exemplarily, when training the generator network and the discriminator network, the training process is:

[0079] 1. Construct an image dataset of images generated by the sample embryo during the division process, and obtain historical sample images of the division process of the sample embryo taken at t consecutive moments in the image dataset.

[0080] 2. Determine the adversarial loss, spatial loss and time loss of the generative network, determine the total loss based on the adversarial loss, spatial loss and time loss, use the gradient descent method to determine the generation parameters of the generative network when the total loss is minimum, and build the generative network based on the generation parameters. Input the historical sample image into the generative network composed of the neural network to generate the predicted image at time t+1.

[0081] 3. The total loss corresponding to the discriminant network can be determined, and the corresponding discriminant network when the total loss is minimized can be determined using the gradient descent method. The predicted image and the actual image at time t+1 are simultaneously input into the discriminant network to determine whether there are abnormal events in the actual image.

[0082] 4. Repeat steps 1, 2, and 3 until the entire data set is traversed, which is recorded as iteration 1.

[0083] 5. Repeat steps 1, 2, 3, and 4, count the number of iterations, and when the number of iterations is greater than the set maximum number, the training of the generation network and the discrimination network is completed.

[0084] S205. Determine whether a chromosomal abnormality event will occur at a future time during the division of the target embryo.

[0085] The method for detecting chromosomal abnormality events provided in an embodiment of the present invention can train the generation network and the discriminant network based on the loss function, thereby improving the accuracy of the prediction image generated by the generation network and the accuracy of the discrimination results produced by the discriminant network, and can ensure the effectiveness of the detection results of abnormal events.

[0086] Embodiment 3

[0087] Figure 3 A structural diagram of a detection device for chromosome abnormality events provided in an embodiment of the present invention; the device is used to execute the detection method for chromosome abnormality events provided in any of the above embodiments. The device and the detection method for chromosome abnormality events in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiment of the detection device for chromosome abnormality events, reference can be made to the embodiment of the detection method for chromosome abnormality events described above. The device may specifically include:

[0088] A historical image acquisition module 10 is used to acquire a historical image of at least one historical moment in the cleavage process of the target embryo;

[0089] A prediction image generation module 11 is used to input each historical image into a pre-established generation network to generate a prediction image of a future moment in the target embryo division process relative to the historical moment;

[0090] The actual image acquisition module 12 is used to acquire the actual image corresponding to the target embryo splitting at a future moment, and determine whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generating network and the discriminant network include a neural network; if not, then enter the abnormal event determination module;

[0091] The abnormal event determination module 13 is used to determine the occurrence of a chromosomal abnormality event at a future time during the division of the target embryo.

[0092] Based on any optional technical solution in the embodiment of the present invention, optionally, before inputting each historical image into the pre-established generation network, the method further includes:

[0093] A generative network module is established to determine the loss function generated by the target embryo division process, and the generation parameters of the predicted image are determined based on the loss function; a generative network is established based on the generation parameters and the neural network structure.

[0094] Based on any optional technical solution in the embodiments of the present invention, optionally, a generation network module is established, including:

[0095] The loss function determination unit is used to determine the adversarial loss function, spatial loss function and temporal loss function generated by the target embryo division process; wherein the spatial loss function includes a pixel loss function and a gradient loss function.

[0096] Based on any optional technical solution in the embodiments of the present invention, optionally, a generation network module is established, including:

[0097] A weighted summation unit is used to determine the weight coefficients corresponding to the adversarial loss function, pixel loss function, gradient loss function and time loss function; based on the weight coefficients, the adversarial loss function, pixel loss function, gradient loss function and time loss function are weightedly summed up and the calculation result is determined as the total loss generated by the target embryo division process; based on the gradient descent method, when the total loss is minimized, the generation parameters corresponding to the generation network of the predicted image are generated.

[0098] Based on any optional technical solution in the embodiments of the present invention, optionally, determining a loss function unit includes:

[0099] Determine an actual image unit, which is used to determine the equivalent function of the regularization network, and obtain the actual image corresponding to the previous moment of the future moment; based on the equivalent function, calculate the first equivalent function value between the predicted image and the actual image corresponding to the previous moment, and calculate the second equivalent function value between the actual image at the future moment and the actual image at the previous moment; calculate the norm of the difference between the first equivalent function value and the second equivalent function value, and determine the calculation result as the time loss function.

[0100] Based on any optional technical solution in the embodiment of the present invention, optionally, the prediction image generation module 11 includes:

[0101] Generate a grayscale image, which is used to grayscale each historical image to generate a corresponding grayscale image; input each grayscale image into a pre-established generation network, perform convolution calculation, downsampling calculation, deconvolution calculation and image jump connection processing on each grayscale image, and determine the processed image as a predicted image.

[0102] Based on any optional technical solution in the embodiments of the present invention, optionally, the following further includes:

[0103] The marking position module is used to mark the position inconsistent with the predicted image in the actual image corresponding to the target embryo division at the future moment after determining that a chromosomal abnormality event occurs during the target embryo division process at the future moment.

[0104] The detection device for abnormal chromosome events provided by the embodiment of the present invention can perform the following method: obtaining a historical image of at least one historical moment in the target embryo's division process; inputting each historical image into a pre-established generation network to generate a predicted image of a future moment in the target embryo's division process relative to the historical moment; obtaining an actual image corresponding to the target embryo's division at the future moment, and determining whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generation network and the discriminant network include a neural network; if not, determining that a chromosomal abnormal event occurs in the target embryo's division process at a future moment. The embodiment of the present invention compares the predicted image with the actual image, that is, compares the normal embryo division image with the actual image, and by judging whether the two are consistent, it can be determined whether there is a chromosomal abnormal event, without the need for manual intervention, thereby improving the efficiency and accuracy of detecting abnormal events.

[0105] It is worth noting that in the embodiment of the above-mentioned device for detecting chromosomal abnormality events, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0106] Embodiment 4

[0107] Figure 4 A structural diagram of an electronic device provided by an embodiment of the present invention. Figure 4 A block diagram of an exemplary electronic device 20 suitable for implementing an embodiment of the present invention is shown. The electronic device 20 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0108] like Figure 4 As shown, the electronic device 20 is in the form of a general computing device. The components of the electronic device 20 may include but are not limited to: one or more processors or processing units 201, a system memory 202, and a bus 203 connecting different system components (including the system memory 202 and the processing unit 201).

[0109] Bus 203 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus structures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnect (PCI) bus.

[0110] The electronic device 20 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 20, including volatile and non-volatile media, removable and non-removable media.

[0111] The system memory 202 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 204 and / or cache memory 205. The electronic device 20 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 206 may be used to read and write non-removable, non-volatile magnetic media. A disk drive for reading and writing removable non-volatile disks (such as "floppy disks"), and an optical disk drive for reading and writing removable non-volatile optical disks (such as CD-ROM, DVD-ROM or other optical media) may be provided. In these cases, each drive may be connected to the bus 203 via one or more data medium interfaces. The memory 202 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of various embodiments of the present invention.

[0112] A program / utility 208 having a set (at least one) of program modules 207 may be stored, for example, in the memory 202, such program modules 207 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 207 generally perform the functions and / or methods of the embodiments described herein.

[0113] The electronic device 20 may also communicate with one or more external devices 209 (e.g., keyboards, pointing devices, displays 210, etc.), may communicate with one or more devices that enable a user to interact with the electronic device 20, and / or may communicate with any device that enables the electronic device 20 to communicate with one or more other computing devices (e.g., network cards, modems, etc.). Such communication may be performed via an input / output (I / O) interface 211. Furthermore, the electronic device 20 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 212. As shown, the network adapter 212 communicates with other modules of the electronic device 20 via a bus 203. It should be understood that other hardware and / or software modules may be used in conjunction with the electronic device 20, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0114] The processing unit 201 executes various functional applications and data processing by running programs stored in the system memory 202 .

[0115] An electronic device provided by the present invention can implement the following method: obtaining a historical image of at least one historical moment in the target embryo's division process; inputting each historical image into a pre-established generation network to generate a predicted image of a future moment in the target embryo's division process relative to the historical moment; obtaining an actual image corresponding to the target embryo's division at the future moment, and determining whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generation network and the discriminant network include a neural network; if not, determining that a chromosomal abnormality event occurs in the target embryo's division process at a future moment. The embodiment of the present invention compares the predicted image with the actual image, that is, compares the normal embryo division image with the actual image, and by judging whether the two are consistent, it can be determined whether there is a chromosomal abnormality event, without the need for manual intervention, thereby improving the efficiency and accuracy of detecting abnormal events.

[0116] Embodiment 5

[0117] An embodiment of the present invention provides a storage medium containing computer executable instructions, wherein the computer executable instructions are used to perform a method for detecting a chromosome abnormality event when executed by a computer processor, the method comprising:

[0118] Obtain a historical image of at least one historical moment in the target embryo's division process; input each historical image into a pre-established generation network to generate a predicted image of a future moment in the target embryo's division process relative to the historical moment; obtain an actual image corresponding to the target embryo's division at the future moment, and determine whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generation network and the discriminant network include a neural network; if not, determine that a chromosomal abnormality event occurs in the target embryo's division process at a future moment. The embodiment of the present invention compares the predicted image with the actual image, that is, compares the normal embryo division image with the actual image, and by judging whether the two are consistent, it can be determined whether there is a chromosomal abnormality event, without the need for manual intervention, thereby improving the efficiency and accuracy of detecting abnormal events.

[0119] Of course, the computer executable instructions of a storage medium containing computer executable instructions provided in an embodiment of the present invention are not limited to the method operations described above, and can also execute related operations in the method for detecting chromosomal abnormality events provided in any embodiment of the present invention.

[0120] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, a device or a device or used in combination with it.

[0121] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0122] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0123] Computer program code for performing the operation of embodiments of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0124] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A method for detecting a chromosomal abnormality event, characterized in that: include: Acquire a historical image of at least one historical moment in the division process of the target embryo; Inputting each of the historical images into a pre-established generation network to generate a predicted image of a future moment in the target embryo division process relative to the historical moment; Acquire an actual image corresponding to the target embryo splitting at the future moment, and determine whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generating network and the discriminant network include neural networks; If not, determining that a chromosomal abnormality event occurs during the division of the target embryo at the future time; The step of inputting each of the historical images into a pre-established generation network to generate a predicted image of a future moment in the target embryo division process relative to the historical moment includes: Performing grayscale processing on each of the historical images to generate a corresponding grayscale image; Inputting each of the grayscale images into the pre-established generation network, performing convolution calculation, downsampling calculation, deconvolution calculation and image jump connection processing on each of the grayscale images, and determining the image obtained after the processing as the predicted image; The determining whether the actual image is consistent with the predicted image based on a pre-established discriminant network includes: Based on the pre-established discriminant network, convolution calculation, downsampling operation and full connection operation are performed on the input actual image and the predicted image to obtain the output result of the discriminant network, and whether the actual image is consistent with the predicted image is determined according to the output result of the discriminant network.

2. The method according to claim 1, characterized in that Before inputting each of the historical images into the pre-established generation network, the method further includes: Determining a loss function generated during the target embryo division process, and determining generation parameters of the predicted image based on the loss function; The generation network is established based on the generation parameters and the neural network structure.

3. The method according to claim 2, characterized in that The step of determining the loss function generated during the target embryo splitting process comprises: Determine the adversarial loss function, spatial loss function and temporal loss function generated by the target embryo splitting process; The spatial loss function includes a pixel loss function and a gradient loss function.

4. The method according to claim 3, characterized in that The determining the generation parameters of the predicted image based on the loss function comprises: Determine weight coefficients corresponding to the adversarial loss function, the pixel loss function, the gradient loss function, and the time loss function; Based on each of the weight coefficients, a weighted sum calculation is performed on the adversarial loss function, the pixel loss function, the gradient loss function and the time loss function, and the calculation result is determined as the total loss generated during the division process of the target embryo; When the total loss is determined to be minimum based on the gradient descent method, generation parameters corresponding to the generation network of the predicted image are generated.

5. The method according to claim 3, characterized in that: Determining the time loss function generated by the target embryo division process includes: Determine an equivalent function of the regularized network to obtain an actual image corresponding to a previous moment of the future moment; Based on the equivalent function, a first equivalent function value between the predicted image and the actual image corresponding to the previous moment is calculated, and a second equivalent function value between the actual image at the future moment and the actual image at the previous moment is calculated; The norm of the difference between the first equivalent function value and the second equivalent function value is calculated, and the calculation result is determined as the time loss function.

6. The method according to claim 1, characterized in that After determining that a chromosomal abnormality event occurs during the division of the target embryo at the future time, the method further includes: In the actual image corresponding to the target embryo splitting at the future moment, the position inconsistent with the predicted image is marked.

7. A device for detecting chromosome abnormality events, characterized in that: include: A historical image acquisition module is used to acquire a historical image of at least one historical moment in the cleavage process of the target embryo; A prediction image generation module is used to input each of the historical images into a pre-established generation network to generate a prediction image of a future moment in the target embryo division process relative to the historical moment; An actual image acquisition module is used to acquire an actual image corresponding to the target embryo splitting at the future moment, and determine whether the actual image is consistent with the predicted image based on a pre-established discriminant network; wherein the generating network and the discriminant network include a neural network; if not, enter an abnormal event determination module; An abnormal event determination module is used to determine the occurrence of a chromosomal abnormality event during the division of the target embryo at the future moment; The module for generating predicted images comprises: A grayscale image generation unit is used to perform grayscale processing on each historical image to generate a corresponding grayscale image; each grayscale image is input into a pre-established generation network, and convolution calculation, downsampling calculation, deconvolution calculation and image jump connection processing are performed on each grayscale image, and the processed image is determined as a predicted image; The actual image acquisition module is specifically used for: Based on the pre-established discriminant network, convolution calculation, downsampling operation and full connection operation are performed on the input actual image and the predicted image to obtain the output result of the discriminant network, and whether the actual image is consistent with the predicted image is determined according to the output result of the discriminant network.

8. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method for detecting chromosomal abnormality events as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for detecting a chromosome abnormality event as described in any one of claims 1 to 6 is implemented.

Images