A method and device for acquiring growth information of ovarian clear cell carcinoma organoids

Abstract

The present application discloses a method and device for acquiring growth information of ovarian clear cell carcinoma organoids. The method for acquiring growth information of ovarian clear cell carcinoma organoids comprises: acquiring an ovarian clear cell carcinoma organoid image; acquiring a trained target detection model; inputting the ovarian clear cell carcinoma organoid image into the trained target detection model, thereby acquiring target cells on the ovarian clear cell carcinoma organoid image; marking each target cell separately, thereby establishing a unique identifier for each target cell; acquiring target cell maps separately from the ovarian clear cell carcinoma organoid image; and acquiring the perimeter information and / or area information of the target cells according to the target cell maps. The present application acquires each target cell through a target detection model, and then acquires the perimeter information and / or area information of each target cell separately, thereby freeing up manpower, and the perimeter information and / or area information of the target cell can be acquired without the need for regular and fixed-point observation by humans.

Description

Technical Field

[0001] The present application relates to the field of image recognition technology, and in particular to a method for acquiring growth information of ovarian clear cell carcinoma organoids, a device for acquiring growth information of ovarian clear cell carcinoma organoids, and a method for observing ovarian clear cell carcinoma organoids. Background Art

[0002] Ovarian cancer is one of the diseases with the highest mortality rate in the world. Existing medical technologies usually obtain the growth process of ovarian clear cell carcinoma organoids through manual observation of ovarian clear cell carcinoma organoid images, which is a waste of manpower and material resources.

[0003] Therefore, it is hoped that a technical solution can be provided to solve or at least alleviate the above-mentioned deficiencies of the prior art. Summary of the invention

[0004] The object of the present invention is to provide a method for acquiring growth information of ovarian clear cell carcinoma organoids to at least solve one of the above-mentioned technical problems.

[0005] The present invention provides the following scheme:

[0006] According to one aspect of the present invention, a method for acquiring growth information of ovarian clear cell carcinoma organoids is provided, and the method for acquiring growth information of ovarian clear cell carcinoma organoids comprises:

[0007] Obtain images of ovarian clear cell carcinoma organoids;

[0008] Get the trained object detection model;

[0009] Inputting the ovarian clear cell carcinoma organoid image into the trained target detection model to obtain the target cells on the ovarian clear cell carcinoma organoid image;

[0010] Each target cell is labeled separately, thereby establishing a unique identity for each target cell;

[0011] Obtain target cell maps from ovarian clear cell carcinoma organoid images respectively;

[0012] The perimeter information and / or area information of the target is obtained according to the target cell map.

[0013] Optionally, the trained object detection model is trained using the following method:

[0014] Get the trained teacher model;

[0015] Acquire fluorescence image models and bright field image models;

[0016] Performing knowledge distillation on the fluorescence image model and the bright field image model according to the teacher model, thereby obtaining a distilled fluorescence image model and a distilled bright field image model;

[0017] The distilled bright field image model is made to learn the distilled fluorescence image model, thereby obtaining a trained target detection model.

[0018] Optionally, the input of the trained teacher model is a fluorescence image and a bright field image;

[0019] The input of the bright field image model is a bright field image;

[0020] The input of the fluorescence image is a fluorescence image.

[0021] Optionally, the fluorescence image model is pruned before training.

[0022] Optionally, pruning the fluorescence image model includes:

[0023] Training the fluorescence image model;

[0024] Calculate the L2 norm value of each convolution kernel in the trained fluorescence image model respectively;

[0025] According to the L2 norm value of each convolution kernel, the scatter plots of each convolution kernel of the fluorescence image model are obtained respectively;

[0026] Set multiple range areas;

[0027] Divide each norm point in the scatter plot of each convolution kernel according to each range area, so as to obtain the number of range areas occupied by each convolution kernel;

[0028] The convolution kernel with the largest number of range areas occupied by each convolution kernel is retained as the first retained convolution kernel;

[0029] Retain each convolution kernel whose number of range areas occupied by each convolution kernel exceeds a first preset threshold as a second reserved convolution kernel, and when in use, the feature channel of the second reserved convolution kernel is prohibited from expanding;

[0030] Delete other convolution kernels to obtain a pruned fluorescence image model.

[0031] Optionally, the trained target detection model is a YOLOv5 model.

[0032] Optionally, obtaining target cell images from ovarian clear cell carcinoma organoid images respectively includes:

[0033] Get each target box in the target detection model identified by the YOLOv5 model;

[0034] According to each target frame, the image of the target frame area is cut out as the target cell image.

[0035] Optionally, acquiring the perimeter information and / or area information of the target according to the target cell map includes:

[0036] grayscale processing is performed on the target cell image, thereby obtaining a grayscale processed target cell image;

[0037] Performing noise reduction processing on the grayscale processed target cell image, thereby obtaining a target cell image after noise reduction processing;

[0038] Binarizing the target cell image after the noise reduction process, so as to obtain a target cell image after the binarization process;

[0039] Perform contour scanning on the binarized target cell image to obtain the cell boundary;

[0040] The perimeter information and / or area information of the target cell is obtained according to the cell boundary.

[0041] The present application also provides a device for acquiring growth information of ovarian clear cell carcinoma organoids, the device for acquiring growth information of ovarian clear cell carcinoma organoids comprising:

[0042] An image acquisition module, wherein the image acquisition module is used to acquire an image of an ovarian clear cell carcinoma organoid;

[0043] A target detection model acquisition module, wherein the target detection model acquisition module is used to acquire a trained target detection model;

[0044] A target recognition module, wherein the target recognition module is used to input the ovarian clear cell carcinoma organoid image into a trained target detection model, thereby obtaining target cells on the ovarian clear cell carcinoma organoid image;

[0045] A marking module, wherein the marking module is used to mark each target cell respectively, thereby establishing a unique identifier for each target cell;

[0046] A target cell image acquisition module, wherein the target cell image acquisition module is used to respectively acquire target cell images from ovarian clear cell carcinoma organoid images;

[0047] A growth information acquisition module is used to acquire the perimeter information and / or area information of the target according to the target cell map.

[0048] The present application also provides a method for observing ovarian clear cell carcinoma organoids, the method comprising:

[0049] Set the interval time;

[0050] Ovarian clear cell carcinoma organoid images were acquired according to the time interval;

[0051] The method for acquiring growth information of ovarian clear cell carcinoma organoids as described above is used to respectively acquire perimeter information and / or area information of each target cell image in the ovarian clear cell carcinoma organoid image;

[0052] The following operations are performed on each two adjacent ovarian clear cell carcinoma organoid images, where one ovarian clear cell carcinoma organoid image is called the first image and the other ovarian clear cell carcinoma organoid image is called the second image:

[0053] The positions of each target cell in the first image are compared with the positions of each target cell in the second image, so as to obtain target cells with the same position or within a preset range as paired cells. Every two successfully paired cells are the same cells and share a unique identifier. A target cell that only appears in the first image but not in the second image is called a dying cell; a target cell that only appears in the second image but not in the first image is called a newborn cell.

[0054] The perimeter information and / or area information of the same cells, the perimeter information and / or area information of dying cells, and the perimeter information and / or area information of newborn cells obtained from every two adjacent images of ovarian clear cell carcinoma organoids are stored in a database.

[0055] The method for acquiring growth information of ovarian clear cell carcinoma organoids of the present application acquires each target cell through a target detection model, and then acquires the perimeter information and / or area information of each target cell respectively, thereby freeing up manpower and eliminating the need for regular and fixed-point observation by humans to acquire the perimeter information and / or area information of the target cells. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a flowchart of a method for acquiring growth information of ovarian clear cell carcinoma organoids in one embodiment of the present application.

[0057] Figure 2 It is a schematic diagram of the structure of an electronic device in one embodiment of the present application.

[0058] Figure 3 is a schematic diagram of an ovarian clear cell carcinoma organoid image in an embodiment of the present application, wherein the framed cells are target cells detected by the target detection model.

[0059] Figure 4 It is a schematic diagram of the distribution of L2 norm values. DETAILED DESCRIPTION

[0060] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] Figure 1 It is a flowchart of a method for acquiring growth information of ovarian clear cell carcinoma organoids in one embodiment of the present application.

[0062] like Figure 1 The method for obtaining growth information of ovarian clear cell carcinoma organoids shown includes:

[0063] Step 1: Obtain ovarian clear cell carcinoma organoid images;

[0064] Step 2: Get the trained object detection model;

[0065] Step 3: Input the ovarian clear cell carcinoma organoid image into the trained target detection model to obtain the target cells on the ovarian clear cell carcinoma organoid image;

[0066] Step 4: Mark each target cell separately to establish a unique identifier for each target cell;

[0067] Step 5: Obtain target cell images from the ovarian clear cell carcinoma organoid images respectively;

[0068] Step 6: Obtain the perimeter information and / or area information of the target cell according to the target cell map.

[0069] The method for acquiring growth information of ovarian clear cell carcinoma organoids of the present application acquires each target cell through a target detection model, and then acquires the perimeter information and / or area information of each target cell respectively, thereby freeing up manpower and eliminating the need for regular and fixed-point observation by humans to acquire the perimeter information and / or area information of the target cells.

[0070] In this embodiment, the trained target detection model is trained using the following method:

[0071] Get the trained teacher model;

[0072] Acquire fluorescence image models and bright field image models;

[0073] Performing knowledge distillation on the fluorescence image model and the bright field image model according to the teacher model, thereby obtaining a distilled fluorescence image model and a distilled bright field image model;

[0074] The distilled bright field image model is made to learn the distilled fluorescence image model, thereby obtaining a trained target detection model.

[0075] In this embodiment, the input of the trained teacher model is a fluorescence image and a bright field image;

[0076] The input of the bright field image model is a bright field image;

[0077] The input of the fluorescence image is a fluorescence image.

[0078] The method for acquiring ovarian clear cell carcinoma organoid growth information of the present application is through two distillations, wherein the input of the teacher model is a fluorescent image and a bright field image (i.e., the characteristics of the fluorescent image and the characteristics of the bright field image comprehensively learned by the teacher model), the input of the bright field image model is a bright field image, and the input of the fluorescent image model is a fluorescent image. Knowledge distillation is used to allow the bright field image model to learn the teacher model's perception ability of fluorescent images and the perception ability of bright field images. In addition, the information of the fluorescent image is further learned through the second learning of the fluorescent image, thereby significantly improving its depth perception ability.

[0079] In this way, when the bright field image model of the present application is actually used, only the bright field image needs to be input, but at the same time, the knowledge of the fluorescent image is also taken into account in the bright field graphic model of the present application.

[0080] In this embodiment, the fluorescence image model is pruned before training.

[0081] join Figure 4 In this embodiment, pruning the fluorescence image model includes:

[0082] Training the fluorescence image model;

[0083] Calculate the L2 norm value of each convolution kernel in the trained fluorescence image model respectively;

[0084] According to the L2 norm value of each convolution kernel, the scatter plots of each convolution kernel of the fluorescence image model are obtained respectively;

[0085] Set multiple range areas;

[0086] Divide each norm point in the scatter plot of each convolution kernel according to each range area, so as to obtain the number of range areas occupied by each convolution kernel;

[0087] The convolution kernel with the largest number of range areas occupied by each convolution kernel is retained as the first retained convolution kernel;

[0088] Retain each convolution kernel whose number of range areas occupied by each convolution kernel exceeds a first preset threshold as a second reserved convolution kernel, and when in use, the feature channel of the second reserved convolution kernel is prohibited from expanding;

[0089] Delete other convolution kernels to obtain a pruned fluorescence image model.

[0090] In this way, in order to solve the problem of parameter redundancy of the original model and the problem that the traditional L2 norm-based weight threshold model pruning method ignores the distribution of weight parameters, the above pruning method is used to prune the original model. By retaining the sparsity of the model to a greater extent, the purpose of retaining the model detection accuracy to a greater extent is achieved.

[0091] See also Figure 4 In this embodiment, the range area can be set as an interval every 0.5. For example, 0-0.5 is a range area, and 0.5 to 1 is a range area.

[0092] See also Figure 4 In this embodiment, the fluorescence image model of the present application is divided into 10 intervals (interval A, interval B, interval C, interval D, interval E, interval F, interval G, interval H, interval I, interval J).

[0093] See also Figure 4 In this embodiment, the convolution kernel corresponding to interval A is the one with the largest number of range areas occupied by various convolution kernels.

[0094] The first preset threshold can be set as needed. For example, in this embodiment, the first preset threshold is 3. Figure 4 It can be seen that the interval E and the interval H both exceed the first preset threshold, then the convolution kernel corresponding to the interval E and the interval H is the second reserved convolution kernel.

[0095] In this embodiment, each interval represents a convolutional layer, a connection layer or a detect layer. For example, in this embodiment, there are 10 intervals in total, which means that the sum of the number of convolutional layers, connection layers and detect layers is 10. For example, interval A to interval E are convolutional layers, interval E, interval F, and interval G are connection layers, and interval I and interval J are detect layers.

[0096] In this way, in order to solve the problem of parameter redundancy of the original model and the problem that the traditional L2 norm-based weight threshold model pruning method ignores the distribution of weight parameters, the above pruning method is used to prune the original model. By retaining the sparsity of the model to a greater extent, the purpose of retaining the model detection accuracy to a greater extent is achieved.

[0097] In this embodiment, the trained target detection model is a YOLOv5 model.

[0098] See also Figure 3 In this embodiment, the step of obtaining target cell images from ovarian clear cell carcinoma organoid images includes:

[0099] Get each target box in the target detection model identified by the YOLOv5 model;

[0100] According to each target frame, the image of the target frame area is cut out as the target cell image.

[0101] In this embodiment, the following method is used to cut out the image of the target frame area according to each target frame as the target cell image:

[0102] According to the target box, the area containing the corresponding target cells is cut out in the ovarian clear cell carcinoma organoid image. In order to restore the true scale, the center coordinates (x, y) of the target box of the target cell image, as well as the width and height w and h of the normalized target box are restored respectively. The specific formula is:

[0103]

[0104] Where W and H are the height and width of the ovarian clear cell carcinoma organoid image, respectively.

[0105] In this embodiment, obtaining the perimeter information and / or area information of the target according to the target cell map includes:

[0106] grayscale processing is performed on the target cell image, thereby obtaining a grayscale processed target cell image;

[0107] Performing noise reduction processing on the grayscale processed target cell image, thereby obtaining a target cell image after noise reduction processing;

[0108] Binarizing the target cell image after the noise reduction process, so as to obtain a target cell image after the binarization process;

[0109] Perform contour scanning on the binarized target cell image to obtain the cell boundary;

[0110] The perimeter information and / or area information of the target cell is obtained according to the cell boundary.

[0111] In this embodiment, the Ramer-Douglas-Peucker (RDP) algorithm is used to perform contour approximation and calculate the perimeter.

[0112] In this embodiment, the area within the contour may be calculated using Green's formula.

[0113] The present application also provides a device for acquiring growth information of ovarian clear cell carcinoma organoids, which comprises an image acquisition module, a target detection model acquisition module, a target recognition module, a marking module, a target cell map acquisition module and a growth information acquisition module, wherein:

[0114] The image acquisition module is used to acquire images of ovarian clear cell carcinoma organoids;

[0115] The target detection model acquisition module is used to obtain the trained target detection model;

[0116] The target recognition module is used to input the ovarian clear cell carcinoma organoid image into the trained target detection model, so as to obtain the target cells on the ovarian clear cell carcinoma organoid image;

[0117] The marking module is used to mark each target cell respectively, thereby establishing a unique identifier for each target cell;

[0118] The target cell map acquisition module is used to respectively acquire target cell maps from ovarian clear cell carcinoma organoid images;

[0119] The growth information acquisition module is used to acquire the perimeter information and / or area information of the target cells according to the target cell map.

[0120] The present application also provides a method for observing ovarian clear cell carcinoma organoids, the method comprising:

[0121] Set the interval time;

[0122] Ovarian clear cell carcinoma organoid images were acquired according to the time interval;

[0123] Using the method for acquiring growth information of ovarian clear cell carcinoma organoids as described in any one of claims 1 to 8, respectively acquiring perimeter information and / or area information of each target cell image in the ovarian clear cell carcinoma organoid image;

[0124] The following operations are performed on each two adjacent ovarian clear cell carcinoma organoid images, where one ovarian clear cell carcinoma organoid image is called the first image and the other ovarian clear cell carcinoma organoid image is called the second image:

[0125] The positions of each target cell in the first image are compared with the positions of each target cell in the second image, so as to obtain target cells with the same position or within a preset range as paired cells. Every two successfully paired cells are the same cells and share a unique identifier. A target cell that only appears in the first image but not in the second image is called a dying cell; a target cell that only appears in the second image but not in the first image is called a newborn cell.

[0126] The perimeter information and / or area information of the same cells, the perimeter information and / or area information of dying cells, and the perimeter information and / or area information of newborn cells obtained from every two adjacent images of ovarian clear cell carcinoma organoids are stored in a database.

[0127] Through the above method, we can obtain various databases. For example, if there are a total of 4 images of ovarian clear cell carcinoma organoids, namely image A (taken on January 1, 2023), image B, (taken on January 2, 2023) image C (taken on January 3, 2023) and image D (taken on January 4, 2023), among which image A and image B are a group of adjacent images, image B and image C are a group of adjacent images, and image C and image D are a group of adjacent images. In this way, a group of adjacent images corresponds to a group of databases, and 3 databases will be obtained. By observing these three databases, researchers can understand the growth of target cells in these few days, for example, which cells have been added, which cells have disappeared, and the changes in the circumference area of ​​​​each cell.

[0128] Figure 2 It is a structural block diagram of an electronic device provided by one or more embodiments of the present invention.

[0129] like Figure 2 As shown, the present application also discloses an electronic device, including: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the method for acquiring growth information of ovarian clear cell carcinoma organoids.

[0130] The present application also provides a computer-readable storage medium storing a computer program executable by an electronic device. When the computer program runs on the electronic device, the steps of the method for acquiring growth information of ovarian clear cell carcinoma organoids can be implemented.

[0131] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0132] The electronic device includes a hardware layer, an operating system layer running on the hardware layer, and an application layer running on the operating system. The hardware layer includes hardware such as a central processing unit (CPU), a memory management unit (MMU), and a memory. The operating system can be any one or more computer operating systems that implement electronic device control through a process, such as a Linux operating system, a Unix operating system, an Android operating system, an iOS operating system, or a Windows operating system. In addition, in an embodiment of the present invention, the electronic device can be a handheld device such as a smart phone or a tablet computer, or can be an electronic device such as a desktop computer or a portable computer, which is not particularly limited in the embodiment of the present invention.

[0133] The execution subject of the electronic device control in the embodiment of the present invention may be an electronic device, or a functional module in the electronic device that can call and execute a program. The electronic device may obtain the firmware corresponding to the storage medium. The firmware corresponding to the storage medium is provided by the supplier. The firmware corresponding to different storage media may be the same or different, which is not limited here. After the electronic device obtains the firmware corresponding to the storage medium, the firmware corresponding to the storage medium may be written into the storage medium, specifically, the firmware corresponding to the storage medium may be burned into the storage medium. The process of burning the firmware into the storage medium may be implemented using existing technology, which will not be described in detail in the embodiment of the present invention.

[0134] The electronic device may also obtain a reset command corresponding to the storage medium. The reset command corresponding to the storage medium is provided by the supplier. The reset commands corresponding to different storage media may be the same or different, and are not limited here.

[0135] At this time, the storage medium of the electronic device is a storage medium in which the corresponding firmware is written, and the electronic device can respond to the reset command corresponding to the storage medium in the storage medium in which the corresponding firmware is written, so that the electronic device resets the storage medium in which the corresponding firmware is written according to the reset command corresponding to the storage medium. The process of resetting the storage medium according to the reset command can be implemented by the existing technology and will not be described in detail in the embodiments of the present invention.

[0136] For the convenience of description, the above devices are described in terms of functions and are divided into various units and modules. Of course, when implementing the present application, the functions of each unit and module can be implemented in the same or multiple software and / or hardware.

[0137] Those skilled in the art will appreciate that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined.

[0138] For the method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0139] It can be known from the description of the above implementation modes that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server or a network device, etc.) to execute the methods described in the various implementation modes of the present application or certain parts of the implementation modes.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for acquiring growth information of ovarian clear cell carcinoma organoids, characterized in that: The method for obtaining growth information of ovarian clear cell carcinoma organoids comprises: Obtain images of ovarian clear cell carcinoma organoids; Get the trained object detection model; Inputting the ovarian clear cell carcinoma organoid image into the trained target detection model to obtain the target cells on the ovarian clear cell carcinoma organoid image; Each target cell is labeled separately, thereby establishing a unique identity for each target cell; Obtain target cell maps from ovarian clear cell carcinoma organoid images respectively; Acquire perimeter information and / or area information of the target cell according to the target cell map; The trained object detection model is trained using the following method: Get the trained teacher model; Acquire fluorescence image models and bright field image models; Performing knowledge distillation on the fluorescence image model and the bright field image model according to the teacher model, thereby obtaining a distilled fluorescence image model and a distilled bright field image model; Allowing the distilled bright field image model to learn the distilled fluorescence image model, thereby obtaining a trained target detection model; The input of the trained teacher model is a fluorescence image and a bright field image; The input of the bright field image model is a bright field image; The input of the fluorescence image model is a fluorescence image.

2. The method for acquiring growth information of ovarian clear cell carcinoma organoids according to claim 1, characterized in that: The fluorescence image model is pruned before training.

3. The method for acquiring growth information of ovarian clear cell carcinoma organoids according to claim 2, characterized in that: The trained target detection model is a YOLOv5 model.

4. The method for acquiring growth information of ovarian clear cell carcinoma organoids according to claim 3, characterized in that: The method of obtaining target cell images from ovarian clear cell carcinoma organoid images comprises: Get each target box in the target detection model identified by the YOLOv5 model; According to each target frame, the image of the target frame area is cut out as the target cell image.

5. The method for acquiring growth information of ovarian clear cell carcinoma organoids according to claim 4, characterized in that: The step of obtaining the perimeter information and / or area information of the target according to the target cell map includes: grayscale processing is performed on the target cell image, thereby obtaining a grayscale processed target cell image; Performing noise reduction processing on the grayscale processed target cell image, thereby obtaining a target cell image after noise reduction processing; Binarizing the target cell image after the noise reduction process, so as to obtain a target cell image after the binarization process; Perform contour scanning on the binarized target cell image to obtain the cell boundary; The perimeter information and / or area information of the target cell is obtained according to the cell boundary.

6. A device for acquiring growth information of ovarian clear cell carcinoma organoids, characterized in that: The device for acquiring growth information of ovarian clear cell carcinoma organoids comprises: An image acquisition module, wherein the image acquisition module is used to acquire an image of an ovarian clear cell carcinoma organoid; A target detection model acquisition module, wherein the target detection model acquisition module is used to acquire a trained target detection model; A target recognition module, wherein the target recognition module is used to input the ovarian clear cell carcinoma organoid image into a trained target detection model, thereby obtaining target cells on the ovarian clear cell carcinoma organoid image; A marking module, wherein the marking module is used to mark each target cell respectively, thereby establishing a unique identifier for each target cell; A target cell image acquisition module, wherein the target cell image acquisition module is used to respectively acquire target cell images from ovarian clear cell carcinoma organoid images; A growth information acquisition module, the growth information acquisition module is used to acquire perimeter information and / or area information of the target cell according to the target cell map; The trained object detection model is trained using the following method: Get the trained teacher model; Acquire fluorescence image models and bright field image models; Performing knowledge distillation on the fluorescence image model and the bright field image model according to the teacher model, thereby obtaining a distilled fluorescence image model and a distilled bright field image model; Allowing the distilled bright field image model to learn the distilled fluorescence image model, thereby obtaining a trained target detection model; The input of the trained teacher model is a fluorescence image and a bright field image; The input of the bright field image model is a bright field image; The input of the fluorescence image model is a fluorescence image.

7. A method for observing ovarian clear cell carcinoma organoids, characterized in that: The ovarian clear cell carcinoma organoid observation method comprises: Set the interval time; Ovarian clear cell carcinoma organoid images were acquired according to the time interval; Using the method for acquiring growth information of ovarian clear cell carcinoma organoids as described in any one of claims 1 to 5, respectively acquiring perimeter information and / or area information of each target cell image in the ovarian clear cell carcinoma organoid image; The following operations are performed on each two adjacent ovarian clear cell carcinoma organoid images, where one ovarian clear cell carcinoma organoid image is called the first image and the other ovarian clear cell carcinoma organoid image is called the second image: The positions of each target cell in the first image are compared with the positions of each target cell in the second image, so as to obtain target cells with the same position or within a preset range as paired cells. Every two successfully paired cells are the same cells and share a unique identifier. A target cell that only appears in the first image but not in the second image is called a dying cell; a target cell that only appears in the second image but not in the first image is called a newborn cell. The perimeter information and / or area information of the same cells, the perimeter information and / or area information of dying cells, and the perimeter information and / or area information of newborn cells obtained from every two adjacent images of ovarian clear cell carcinoma organoids are stored in a database.

Images