TWI761109B - Cell density grouping method, cell density grouping device, electronic device and storage media - Google Patents

Abstract

The present application provides a cell density grouping method, a cell density grouping device, an electronic device and storage medium. The method includes: inputting images into a pre-detected number of self-encoders of a density cluster model to a pre-detected number of reconstructed images to be detected, and each density cluster model corresponding to a density range; inputting the images to be detected and the reconstructed images to be detected into a twin network model of the density cluster model to calculate a first error value; determining a minimum first error value and determining a density range of the images to be detected based on the minimum first error value. The present application utilizes cell images with different densities to train the self-encoders and twin network models to distinguish the cell density range, thus accurately dividing the cell images with same density range, and improving efficiency of grouping cell image density.

Description

Cell density grouping method, device, electronic device and computer storage medium

The present application relates to the field of image processing, in particular to a cell density grouping method, device, electronic device and computer storage medium.

When performing image processing, it is necessary to calculate the number and volume of cells in the image for cell image density grouping, and then calculate the proportion of cells in the image. Because the calculation process of cell image density grouping is complicated, it is usually necessary to spend a lot of time.

In view of the above content, it is necessary to propose a cell density grouping method, device, electronic equipment and computer storage medium to realize the rapid classification of the cell density range.

A first aspect of the present application provides a cell density grouping method, the method comprising: inputting images to be detected into an autoencoder of a preset number of density grouping models, and obtaining a preset number of reconstructed images to be detected, each of the The density grouping model is composed of the self-encoder and the twin network model, each of the density grouping models corresponds to a density range, and each reconstructed image to be detected corresponds to a density grouping model; the to-be-detected image and the Input the Siamese network model of the density clustering model corresponding to each reconstructed image to be detected, and calculate the difference between the reconstructed image to be detected and the reconstructed image to be detected The first error value is composed of all the first error values to form a first error value set, and each first error value corresponds to a density clustering model; the minimum first error value in the first error value set is determined, and the minimum first error value is determined. The density range corresponding to the density grouping model corresponding to the first error value is used as the density range of the image to be detected.

Optionally, the method further includes: training the density clustering model, including: dividing the cell image set according to the density range to obtain a preset number of cell training image sets, each cell training image set corresponds to one Density range; perform a training operation on each cell training image set to obtain a preset number of density clustering models, each of which corresponds to a density range, and the training operation includes: collecting the cell training images into a set The cell training image is converted into a cell training image vector, and a cell training image vector set is composed of all cell training image vectors; the cell training image vector set is used to train an autoencoder, and the trained autoencoder is obtained. inputting the cell training image vector set into the autoencoder after the training of the density clustering model is completed to obtain a cell reconstruction image vector set, and the cell reconstruction image vector set includes the cell reconstruction image vector; Using the cell training image vector set and the cell reconstruction image vector set to train the twin network model, the trained twin network model is obtained, the training completed autoencoder and the training completed twin network The model constitutes the density clustering model.

Optionally, the step of using the cell training image vector set to train the autoencoder, and obtaining the trained autoencoder comprises: inputting the cell training image vector in the cell training image vector set into the autoencoder. The encoding layer of the encoder obtains the hidden vector of the cell training image vector; the hidden vector is input into the decoding layer of the autoencoder to obtain the reconstructed image vector of the cell training image vector; using the preset Calculate the second error value between the cell training image vector and the reconstructed image vector, adjust the parameters of the encoding layer and the decoding layer to minimize the error value, and obtain the density Autoencoders for swarming models.

Optionally, the twin network model includes a first neural network and a second neural network, and the twin network model is trained using the cell training image vector set and the cell reconstruction image vector set to obtain: The trained twin network model includes: inputting the cell training image vector in the cell training image vector set into the first neural network to obtain a first feature map; integrating the cell reconstruction image vector into the first neural network The reconstructed image vector of the cells is input into the second neural network to obtain a second feature map; the third error value between the first feature map and the second feature map is calculated, and a third error value between the first feature map and the second feature map is calculated. Three error values optimize the first neural network and the second neural network to obtain the trained twin network model.

Optionally, the weight of the first neural network and the weight of the second neural network are the same, and the structure of the first neural network and the structure of the second neural network are the same.

Optionally, the inputting the cell training image vector in the cell training image vector set into the first neural network, and obtaining the first feature map includes: inputting the cell training image vector into the cell training image vector. A first neural network; calculating the cell training image vector and the matrix in the first neural network to obtain the features of the cell training image vector to form the first feature map.

Optionally, the calculating a third error value between the first feature map and the second feature map includes: using a mean absolute value error function to calculate the difference between the first feature map and the second feature map. The average absolute value error between the two, and the average absolute value error is used as the third error value between the first feature map and the second feature map.

A second aspect of the present application provides a cell density grouping device, the device includes: an image reconstruction module, which is used for inputting images to be detected into an autoencoder of a preset number of density clustering models, and obtains a preset number of to-be-detected images. Detecting and reconstructing images, each of the density clustering models is composed of the autoencoder and the twin network model, each of the density clustering models corresponds to a density range, and each reconstructed image to be detected corresponds to a density clustering model ; The error value calculation module is used to input the image to be detected and each reconstructed image to be detected into the twin network model of the density clustering model corresponding to each reconstructed image to be detected, and calculate the The first error value between the image to be detected and each reconstructed image to be detected is composed of all the first error values to form a first error value set, and each first error value corresponds to a density clustering model; the density range determines the model. group, which is used to determine the minimum first error value in the first error value set, and use the density range corresponding to the density grouping model corresponding to the minimum first error value as the density range of the image to be detected.

A third aspect of the present application provides an electronic device, the electronic device includes: a memory, which stores at least one instruction; and a processor, which executes the instructions stored in the memory to implement the cell density grouping method.

The fourth application of the present application provides a computer storage medium on which a computer program is stored, and when the computer program is executed by a processor, the cell density grouping method is implemented.

The present application can use cell images with different density ranges to train an autoencoder to reconstruct images and train a twin network model to distinguish the range of cell densities, so that accurate division of cell images with the same density range can be achieved, Improve the efficiency of cell image density clustering.

30: Cell Density Clustering Device

301: Image reconstruction module

302: Error value calculation module

303: Density range determination module

6: Electronic equipment

61: Memory

62: Processor

63: Computer Programs

S11~S13: Steps

FIG. 1 is a flowchart of a cell density grouping method according to an embodiment of the present application.

FIG. 2 is a structural diagram of a cell density grouping device according to an embodiment of the present application.

FIG. 3 is a schematic diagram of an electronic device for implementing a cell density grouping method according to an embodiment of the present application.

In order to more clearly understand the above objects, features and advantages of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments may be combined with each other in the case of no conflict.

In the following description, many specific details are set forth in order to facilitate a full understanding of the present invention, and the described embodiments are only some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terms used herein in the description of the present invention are for the purpose of describing specific embodiments only, and are not intended to limit the present invention.

Preferably, the cell density clustering method of the present application is applied in one or more electronic devices. The electronic device is a device that can automatically perform numerical calculation and/or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs) , Programmable Gate Array (Field-Programmable Gate Array, FPGA), Digital Signal Processor (Digital Signal Processor, DSP), embedded devices, etc.

The electronic device may be a computing device such as a desktop computer, a notebook computer, a tablet computer, and a cloud server. The electronic device can perform human-computer interaction with the user by means of a keyboard, a mouse, a remote control, a touch pad or a voice control device.

Example 1

FIG. 1 is a flowchart of a cell density grouping method in an embodiment of the present application. The cell density grouping method is applied in electronic devices. According to different requirements, the order of the steps in the flowchart can be changed, and some steps can be omitted.

Referring to Figure 1, the cell density grouping method specifically includes the following steps:

Step S11, input the image to be detected into the self-encoder of a preset number of density clustering models to obtain a preset number of reconstructed images to be detected, each of the density clustering models is composed of the self-encoder and the twin network model. Each of the density grouping models corresponds to a density range, and each reconstructed image to be detected corresponds to a density grouping model.

For example, the preset number may be 4, and 4 density clustering models correspond to 4 density ranges (0, 40%), (40%, 60%), (60%, 80%), (80%, 100 respectively) %), input the image to be detected into the autoencoder of 4 density clustering models, and obtain 4 reconstructed images to be detected, which are respectively the reconstructed image A to be detected, the reconstructed image B to be detected, and the reconstructed image C to be detected. , the reconstructed image D to be detected.

Step S12, input the image to be detected and each reconstructed image to be detected into the twin network model of the density clustering model corresponding to each reconstructed image to be detected, and calculate the image to be detected With the first error value of each reconstructed image to be detected, a first error value set is composed of all the first error values, and each first error value corresponds to a density clustering model.

For example, the image to be detected and the reconstructed image A to be detected are input into the Siamese network model of the density clustering model corresponding to the density range (0, 40%), and the image to be detected and the reconstructed image B to be detected are input The twin network model of the density clustering model corresponding to the density range (40%, 60%), the image to be detected and the reconstructed image C to be detected are the twins of the density clustering model corresponding to the input density range (60%, 80%) The network model, the twin network model of the density clustering model corresponding to the input density range (80%, 100%) of the image to be detected and the reconstructed image to be detected D, calculate the image to be detected and each of the The first error value of the reconstructed image to be detected is composed of all the first error values to form a first error value set, and each first error value corresponds to a density clustering model.

Step S13: Determine the minimum first error value in the first error value set, and use the density range corresponding to the density grouping model corresponding to the minimum first error value as the density range of the image to be detected.

For example, when the minimum first error value in the first error value set is 10%, the density range (60%, 80%) corresponding to the density grouping model corresponding to the minimum first error value is used as the to-be-detected The density range of the image.

In at least one embodiment of the present application, the method further includes: training the density clustering model, including: dividing the cell image set according to the density range to obtain a preset number of cell training image sets, each cell The training image set corresponds to a density range; Perform a training operation on each cell training image set to obtain a preset number of density clustering models, each of which corresponds to a density range, and the training operation includes: The cell training image is converted into a cell training image vector, and a cell training image vector set is composed of all cell training image vectors; the autoencoder is trained by using the cell training image vector set, and the trained autoencoder is obtained; The cell training image vector set is input into the autoencoder after the training of the density clustering model is completed, and a cell reconstruction image vector set is obtained, and the cell reconstruction image vector set includes the cell reconstruction image vector; using the cell training The image vector set and the cell reconstruction image vector set train the twin network model to obtain the trained twin network model, and the trained self-encoder and the trained twin network model form the density Clustering model.

For example, divide the cell image set by 4 density ranges (0, 40%), (40%, 60%), (60%, 80%), (80%, 100%) to get four cell training Image set, perform the training operation on each cell training image set, and obtain 4 density clustering models, each of which corresponds to a density range.

In at least one embodiment of the present application, the step of using the cell training image vector set to train an autoencoder, and obtaining a trained autoencoder includes: adding the cell training map in the cell training image vector set Input the image vector into the coding layer of the autoencoder to obtain the hidden vector of the cell training image vector; input the hidden vector into the decoding layer of the autoencoder to obtain the reconstruction map of the cell training image vector image vector; use a preset error function to calculate a second error value between the cell training image vector and the reconstructed image vector, and adjust the parameters of the encoding layer and the decoding layer to minimize the error value to obtain the autoencoder of the density clustering model.

，其中，MAE 為所述最小絕對值函數，y _i 為所述細胞訓練圖像向量的第i個向量，

為重建圖 像向量的第i個向量，n為所述細胞訓練圖像向量與所述重建圖像向量的向量維度。 In at least one embodiment of the present application, when the preset error function is a minimum absolute value error function, the minimum absolute value function error is used to calculate the difference between the cell training image vector and the reconstructed image vector The second error value, the calculation formula is

, where MAE is the minimum absolute value function, y _i is the ith vector of the cell training image vector,

is the ith vector of the reconstructed image vector, and n is the vector dimension of the cell training image vector and the reconstructed image vector.

，其中，RSE為所述均方 差函數，y _i 為所述細胞訓練圖像向量的第i個向量，

為重建圖像向量的第i個 向量，n為所述細胞訓練圖像向量與所述重建圖像向量的向量維度。 In at least one embodiment of the present application, when the preset error function is a mean square error function, the mean square error function error is used to calculate the second error between the cell training image vector and the reconstructed image vector value, calculated as

, where RSE is the mean square error function, y _i is the ith vector of the cell training image vector,

is the ith vector of the reconstructed image vector, and n is the vector dimension of the cell training image vector and the reconstructed image vector.

In at least one embodiment of the present application, the twin network model includes a first neural network and a second neural network, and the training is performed using the cell training image vector set and the cell reconstruction image vector set. The twin network model, the trained twin network model includes: Inputting the cell training image vector in the cell training image vector set into the first neural network to obtain a first feature map; inputting the cell reconstruction image vector in the cell reconstruction image vector set The second neural network obtains a second feature map; calculates a third error value between the first feature map and the second feature map, and optimizes the first neural network according to the third error value network and the second neural network to obtain the trained twin network model.

In at least one embodiment of the present application, the weight of the first neural network is the same as the weight of the second neural network, and the structure of the first neural network is the same as the weight of the second neural network. The structure is the same.

In at least one embodiment of the present application, the inputting the cell training image vector in the cell training image vector set into the first neural network, and obtaining the first feature map includes: training the cell The image vector is input to the first neural network; the cell training image vector and the matrix in the first neural network are calculated to obtain the characteristics of the cell training image vector, forming the first neural network. feature map.

In at least one embodiment of the present application, the calculating a third error value between the first feature map and the second feature map includes: using an average absolute value error function to calculate the difference between the first feature map and the second feature map. The average absolute value error between the second feature maps is determined, and the average absolute value error is used as a third error value between the first feature map and the second feature map.

In other embodiments of the present application, the calculating a third error value between the first feature map and the second feature map includes: calculating the first feature map and the second feature map using a mean square error function mean square error between feature maps, and the mean square error value is used as a third error value between the first feature map and the second feature map.

The present application can use cell images with different density ranges to train an autoencoder to reconstruct images and train a twin network model to distinguish the range of cell densities, so that accurate division of cell images with the same density range can be achieved, Improve the efficiency of cell image density clustering.

Example 2

FIG. 2 is a structural diagram of a cell density clustering device 30 according to an embodiment of the present application.

In some embodiments, the cell density clustering device 30 operates in an electronic device. The cell density grouping device 30 may include a plurality of functional modules composed of code segments. The code of each program segment in the cell density clustering device 30 can be stored in memory and executed by at least one processor.

In this embodiment, the cell density grouping device 30 can be divided into a plurality of functional modules according to the functions it performs. Referring to FIG. 3 , the cell density grouping device 30 may include an image reconstruction module 301 , an error value calculation module 302 and a density range determination module 303 . A module referred to in this application refers to a series of computer program segments that can be executed by at least one processor and can perform fixed functions, and are stored in a memory. In some embodiments, the functions of each module will be described in detail in subsequent embodiments.

The image reconstruction module 301 inputs the images to be detected into the autoencoder of a preset number of density clustering models, and obtains a preset number of reconstructed images to be detected, and each of the density clustering models is determined by the autoencoder. It is composed of a twin network model, each density grouping model corresponds to a density range, and each reconstructed image to be detected corresponds to a density grouping model.

The error value calculation module 302 inputs the to-be-detected image and each of the to-be-detected reconstructed images into the twin network model of the density clustering model corresponding to each of the to-be-detected reconstructed images, and calculates the The first error value of the image to be detected and each reconstructed image to be detected is composed of all the first error values to form a first error value set, and each first error value corresponds to a density clustering model.

The density range determination module 303 determines the minimum first error value in the first error value set, and uses the density range corresponding to the density grouping model corresponding to the minimum first error value as the image to be detected density range.

In at least one embodiment of the present application, the apparatus further includes a density clustering model training module.

The density clustering model training module is used to train the density clustering model, including: dividing the cell image set according to the density range to obtain a preset number of cell training image sets, each cell training image set corresponds to one Density range; perform a training operation on each cell training image set to obtain a preset number of density clustering models, each of which corresponds to a density range, and the training operation includes: collecting the cell training images into a set The cell training image is converted into a cell training image vector, and a cell training image vector set is composed of all cell training image vectors; the cell training image vector set is used to train an autoencoder, and the trained autoencoder is obtained. inputting the cell training image vector set into the autoencoder after the training of the density clustering model is completed to obtain a cell reconstruction image vector set, the cell reconstruction image vector set including the cell reconstruction image vector; using the Described cell training image vector set and described cell reconstruction image vector set training twin network model, obtains the twin network model that the training completes, and the autoencoder that the described training completes and the twin network model that the described training completes are formed The density clustering model.

In at least one embodiment of the present application, the density clustering model training module uses the cell training image vector set to train an autoencoder, and obtaining a trained autoencoder includes: integrating the cell training image vector set The cell training image vector is input into the coding layer of the autoencoder to obtain the hidden vector of the cell training image vector; the hidden vector is input into the decoding layer of the autoencoder to obtain the cell training The reconstructed image vector of the image vector; use a preset error function to calculate the second error value between the cell training image vector and the reconstructed image vector, and adjust the parameters of the encoding layer and the decoding layer To minimize the error value, an autoencoder for the density clustering model is obtained.

，其中，MAE 為所述最小絕對值函數，y _i 為所述細胞訓練圖像向量的第i個向量，

為重建圖 像向量的第i個向量，n為所述細胞訓練圖像向量與所述重建圖像向量的向量維度。 In at least one embodiment of the present application, when the preset error function is a minimum absolute value error function, the minimum absolute value function error is used to calculate the difference between the cell training image vector and the reconstructed image vector The second error value, the calculation formula is

, where MAE is the minimum absolute value function, y _i is the ith vector of the cell training image vector,

is the ith vector of the reconstructed image vector, and n is the vector dimension of the cell training image vector and the reconstructed image vector.

，其中，RSE為所述均方 差函數，y _i 為所述細胞訓練圖像向量的第i個向量，

為重建圖像向量的第i個 向量，n為所述細胞訓練圖像向量與所述重建圖像向量的向量維度。 In at least one embodiment of the present application, when the preset error function is a mean square error function, the mean square error function error is used to calculate the second error between the cell training image vector and the reconstructed image vector value, calculated as

, where RSE is the mean square error function, y _i is the ith vector of the cell training image vector,

is the ith vector of the reconstructed image vector, and n is the vector dimension of the cell training image vector and the reconstructed image vector.

In at least one embodiment of the present application, the twin network model includes a first neural network and a second neural network.

In at least one embodiment of the present application, the density clustering model training module uses the cell training image vector set and the cell reconstruction image vector set to train a twin network model to obtain a trained twin network model The method includes: inputting the cell training image vector in the cell training image vector set into the first neural network to obtain a first feature map; inputting the cell reconstruction image in the cell reconstruction image vector set The vector is input into the second neural network to obtain a second feature map; the third error value between the first feature map and the second feature map is calculated, and the third error value is optimized according to the third error value. A neural network and the second neural network obtain the trained twin network model.

In at least one embodiment of the present application, the weight of the first neural network is the same as the weight of the second neural network, and the structure of the first neural network is the same as the weight of the second neural network. The structure is the same.

In at least one embodiment of the present application, the density clustering model training module inputs the cell training image vector in the cell training image vector set into the first neural network, and obtaining the first feature map includes the following steps: : input the cell training image vector into the first neural network; calculate the cell training image vector and the matrix in the first neural network to obtain the characteristics of the cell training image vector , forming the first feature map.

In at least one embodiment of the present application, calculating the third error value between the first feature map and the second feature map by the density clustering model training module includes: using a mean absolute value error function to calculate the The average absolute value error between the first feature map and the second feature map, and the average absolute value error is used as the third error value between the first feature map and the second feature map.

In other embodiments of the present application, the calculating a third error value between the first feature map and the second feature map includes: calculating the first feature map and the second feature map using a mean square error function mean square error between feature maps, and the mean square error value is used as a third error value between the first feature map and the second feature map.

The present application can use cell images with different density ranges to train an autoencoder to reconstruct images and train a twin network model to distinguish the range of cell densities, so that accurate division of cell images with the same density range can be achieved, Improve the efficiency of cell image density clustering.

Example 3

FIG. 3 is a schematic diagram of an electronic device 6 in an embodiment of the present application.

The electronic device 6 includes a memory 61 , a processor 62 and a computer program 63 stored in the memory 61 and executable on the processor 62 . When the processor 62 executes the computer program 63, the steps in the above embodiments of the cell density grouping method are implemented, for example, steps S11 to S13 shown in FIG. 1 . Alternatively, when the processor 62 executes the computer program 63, the functions of the modules/units in the above embodiments of the cell density clustering device are implemented, such as modules 301-303 in FIG. 2 .

Exemplarily, the computer program 63 can be divided into one or more modules/units, and the one or more modules/units are stored in the memory 61 and executed by the processor 62 , to complete this application. The one or more modules/units may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program 63 in the electronic device 6 . For example, the computer program 63 can be divided into the image reconstruction module 301, the error value calculation module 302 and the density range determination module 303 in FIG.

In this embodiment, the electronic device 6 may be a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud terminal device. Those skilled in the art can understand that the schematic diagram is only an example of the electronic device 6, and does not constitute a limitation to the electronic device 6, and may include more or less components than the one shown, or combine some components, or different Components such as the electronic device 6 may also include input and output devices, network access devices, bus bars, and the like.

The processor 62 may be a central processing unit (CPU), other general-purpose processors, a digital signal processor (DSP), or an application specific integrated circuit (ASIC). ), off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor 62 can also be any conventional processor, etc. The processor 62 is the control center of the electronic device 6, and uses various interfaces and lines to connect the entire electronic device 6. various parts.

The memory 61 can be used to store the computer programs 63 and/or modules/units, and the processor 62 runs or executes the computer programs and/or modules/units stored in the memory 61, And call the data stored in the memory 61 to realize various functions of the electronic device 6 . The memory 61 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; storage data The area may store data (such as audio data, phone book, etc.) created according to the use of the electronic device 6, and the like. In addition, the memory 61 may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, Smart Media Card (SMC), Secure Digital (Secure Digital, SD) card, flash memory card (Flash Card), at least one disk memory device, flash memory device, or other volatile solid state memory device.

If the modules/units integrated in the electronic device 6 are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the present application realizes all or part of the processes in the methods of the above embodiments, and can also be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, When the computer program is executed by the processor, the steps of the above method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code may be in the form of original code, object code, executable file, or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a flash drive, a removable hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM, Read-Only Memory); Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the device embodiments described above are only illustrative. For example, the division of the modules is only a logical function division, and other division methods may be used in actual implementation.

In addition, each functional module in each embodiment of the present application may be integrated in the same processing module, or each module may exist physically alone, or two or more modules may be integrated in the same module. The above-mentioned integrated modules can be implemented in the form of hardware, or can be implemented in the form of hardware plus software function modules.

It will be apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, but that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments are to be regarded in all respects as illustrative and not restrictive, and the scope of the present invention is defined by the appended claims rather than the foregoing description, and is therefore intended to fall within the scope of the claims. All changes within the meaning and range of the equivalents of , are included in the present invention. Any reference sign in a claim should not be construed as limiting the claim to which it relates. Furthermore, it is clear that the word "comprising" does not exclude other modules or steps, and the singular does not exclude the plural. Multiple modules or electronic devices stated in the electronic device claim may also be implemented by the same module or electronic device through software or hardware. The terms first, second, etc. are used to denote names and do not denote any particular order.

To sum up, the present invention complies with the requirements of an invention patent, and a patent application can be filed in accordance with the law. However, the above descriptions are only the preferred embodiments of the present invention, and for those who are familiar with the techniques of this case, equivalent modifications or changes made in accordance with the creative spirit of this case shall be included in the scope of the following patent application.

S11~S13: Steps

Claims (10)

A cell density grouping method, wherein the method comprises: inputting the images to be detected into an autoencoder of a preset number of density grouping models to obtain a preset number of reconstructed images to be detected, each of the density grouping models is composed of The self-encoder is composed of a twin network model, each of the density grouping models corresponds to a density range, and each reconstructed image to be detected corresponds to a density grouping model, wherein the preset number is the number of density grouping models, The density range is the proportion of the number of cells and the volume in the image; the image to be detected and each reconstructed image to be detected are input into the density clustering model corresponding to each reconstructed image to be detected. Siamese network model, and calculate the first error value between the image to be detected and each reconstructed image to be detected, and form a first error value set by all the first error values, each first error value corresponds to one Density clustering model; determining the minimum first error value in the first error value set, and using the density range corresponding to the density clustering model corresponding to the minimum first error value as the density range of the image to be detected. The cell density grouping method according to claim 1, wherein the method further comprises: training the density grouping model, including: dividing the cell image set according to the density range to obtain a preset number of cell training image sets , each cell training image set corresponds to a density range; perform a training operation on each cell training image set to obtain a preset number of density clustering models, each of the density clustering models corresponds to a density range, and the training operation Including: converting the cell training image in the cell training image set into a cell training image vector, and forming a cell training image vector set from all cell training image vectors; using the cell training image vector set to train Autoencoder to obtain a trained autoencoder; input the cell training image vector set into the trained autoencoder of the density clustering model to obtain a cell reconstruction image vector set, the cell reconstruction image vector The set includes a cell reconstruction image vector; use the cell training image vector set and the cell reconstruction image vector set to train a twin network model, and obtain a trained twin network model, and the trained autoencoder is The trained twin network model constitutes the density clustering model. The cell density clustering method according to claim 2, wherein the training an autoencoder by using the cell training image vector set to obtain a trained autoencoder comprises: combining all the cells in the cell training image vector set The cell training image vector is input into the coding layer of the autoencoder to obtain the hidden vector of the cell training image vector; the hidden vector is input into the decoding layer of the autoencoder to obtain the cell training image The reconstructed image vector of the vector; use a preset error function to calculate the second error value between the cell training image vector and the reconstructed image vector, and adjust the parameters of the encoding layer and the decoding layer to minimize Quantize the error value to obtain the autoencoder of the density clustering model. The cell density clustering method according to claim 2, wherein the twin network model includes a first neural network and a second neural network, and the training image vector set using the cells is the same as the The cell reconstruction image vector set trains the twin network model, and obtaining the trained twin network model includes: inputting the cell training image vector in the cell training image vector set into the first neural network, obtaining a first feature map; inputting the cell reconstruction image vector in the cell reconstruction image vector set into the second neural network to obtain a second feature map; calculating the first feature map and the second feature map The third error value between the feature maps is obtained, and the first neural network and the second neural network are optimized according to the third error value to obtain the trained twin network model. The cell density clustering method according to claim 4, wherein the weight of the first neural network and the weight of the second neural network are the same, and the structure of the first neural network is the same as the weight of the second neural network. The structure of the neural network is the same. The cell density grouping method according to claim 4, wherein the inputting the cell training image vector in the cell training image vector set into the first neural network, and obtaining the first feature map comprises: The cell training image vector is input to the first neural network; the cell training image vector and the matrix in the first neural network are calculated to obtain the characteristics of the cell training image vector, which is composed of the first feature map. The cell density grouping method according to claim 4, wherein the calculating the third error value between the first feature map and the second feature map comprises: calculating the first error value using a mean absolute value error function The average absolute value error between the feature map and the second feature map, and the average absolute value error is taken as the third error value between the first feature map and the second feature map. A cell density grouping device, comprising: an image reconstruction module, which is used to input images to be detected into an autoencoder of a preset number of density grouping models to obtain a preset number of reconstructed images to be detected, each of which is The density clustering model is composed of the self-encoder and the twin network model, each density clustering model corresponds to a density range, and each reconstructed image to be detected corresponds to a density clustering model, wherein the preset number is the density The number of clustering models, the density range is the proportion of the number of cells and the volume in the image; the error value calculation module is used to input the image to be detected and each reconstructed image to be detected with each The twin network model of the density clustering model corresponding to the reconstructed image to be detected, and calculate the first error value between the reconstructed image to be detected and each reconstructed image to be detected, and all the first error values compose the first error value. An error value set, each first error value corresponds to a density clustering model; a density range determination module is used to determine the minimum first error value in the first error value set, and assign the minimum first error value to the corresponding The density range corresponding to the density clustering model is taken as the density range of the image to be detected. An electronic device, wherein the electronic device includes: a memory that stores at least one instruction; and The processor executes the instructions stored in the memory to implement the cell density grouping method according to any one of claim 1 to 7. A computer storage medium on which a computer program is stored, wherein: the computer program implements the cell density grouping method according to any one of claim 1 to 7 when the computer program is executed by a processor.

Images