Red blood cell detection area selection method and device, computer equipment and storage medium

Abstract

The invention relates to a red blood cell detection area selection method and device, computer equipment and a storage medium, a cell image is input into a pre-trained deep neural network to predict the cell distribution condition and generate a first thermodynamic diagram, and the more uniform the cell distribution is, the higher the thermodynamic value corresponding to the area is; carrying out filtering processing on the first heat map by adopting N preset filtering frames with different sizes to obtain N second thermodynamic diagrams; extracting a coordinate position with the maximum thermodynamic value in each second thermodynamic diagram, and generating a selected area frame with the same size as the corresponding filtering frame by taking the coordinate position with the maximum thermodynamic value on the cell image as a center; extracting cell edges in each selected area frame, and detecting circular features of the cells based on the cell edges; the cells in the selected area frames are screened based on the circular features, the number of red blood cells in each selected area frame is counted, the selected area frame with the red blood cell number closest to the target number serves as the red blood cell detection area, and the red blood cell analysis and detection efficiency is effectively improved.

Description

Technical Field

[0001] The present application relates to the field of cell detection technology, and in particular to a method, device, computer equipment and storage medium for selecting a red blood cell detection area. Background Art

[0002] Red blood cell morphology analysis is an important diagnostic method in medical testing, especially in the diagnosis of blood diseases and anemia. Doctors need to make diagnoses based on the size, morphology and distribution of red blood cells in cell images. Of course, in order to improve diagnostic efficiency and the accuracy of diagnostic results, it is usually necessary to detect the target number of red blood cells in an area with a relatively uniform cell distribution, that is, to use this area with a relatively uniform cell distribution as the red blood cell detection area. As one of the key steps in red blood cell morphology analysis, it is particularly important to select the red blood cell detection area.

[0003] In the prior art, a suitable red blood cell detection area is selected by manual observation and counting, but manual observation and counting are easily affected by the subjective judgment of the observer and are time-consuming, resulting in the problem of low efficiency of red blood cell analysis and detection in the prior art.

[0004] In the prior art, selecting a suitable red blood cell detection area by manual observation and counting has the problem of low efficiency in red blood cell analysis and detection, and no effective solution has been proposed yet. Summary of the invention

[0005] Based on this, it is necessary to provide a red blood cell detection area selection method, device, computer equipment and computer-readable storage medium to address the above-mentioned technical problems, so as to solve the problem of low efficiency of red blood cell analysis and detection in the related technology of manually observing and counting to select a suitable red blood cell detection area.

[0006] In a first aspect, an embodiment of the present application provides a method for selecting a red blood cell detection area, the method being used to select a red blood cell detection area from a cell image, the method comprising the following steps:

[0007] Inputting the cell image into a pre-trained deep neural network, obtaining a prediction result of the cell distribution of the cell image, and generating a first thermodynamic map corresponding to the cell image based on the prediction result, wherein a region in the cell image with more uniform cell distribution has a higher corresponding thermodynamic value;

[0008] Using preset N filter frames of different sizes to perform mean filtering and smoothing processing on the first heat map, respectively, to obtain N second heat maps after filtering processing; wherein N is an integer greater than or equal to 1;

[0009] Extract the coordinate position with the largest thermal value in each of the second thermal maps, and generate a selection box with the same size as the corresponding filter box with the coordinate position with the largest thermal value on the cell image as the center;

[0010] Extracting the cell edge in each of the selection boxes, and detecting the circular feature of the cell based on the cell edge;

[0011] The cells in the selection box are screened based on the detected circular features, the number of red blood cells in each selection box is counted, and the selection box with the red blood cell number closest to the target number is used as the red blood cell detection area.

[0012] In one embodiment, extracting the cell edge in each of the selection boxes comprises the following steps:

[0013] The Canny edge detection algorithm is used to extract the cell edges in each selection box.

[0014] In one embodiment, the method of detecting the circular feature of a cell based on the cell edge comprises the following steps:

[0015] The circular features of cells are detected based on cell edges using Hough transform.

[0016] In one embodiment, the pre-trained deep neural network is based on the MobileNetV3 small network model.

[0017] In a second aspect, the present application also provides a device for selecting a red blood cell detection area, the device comprising a generation module, a filtering module, an extraction module, a detection module and a screening module;

[0018] The generating module is used to input the cell image into a pre-trained deep neural network, obtain a prediction result of the cell distribution of the cell image, and generate a first thermodynamic map corresponding to the cell image based on the prediction result, wherein in the first thermodynamic map, a region in the cell image where the cell distribution is more uniform has a higher corresponding thermodynamic value;

[0019] The filtering module is used to perform mean filtering and smoothing processing on the first heat map using preset N filter frames of different sizes to obtain N second heat maps after filtering processing; wherein N is an integer greater than or equal to 1;

[0020] The extraction module is used to extract the coordinate position with the maximum thermal value in each of the second thermal maps, and to generate a selection frame with the same size as the corresponding filter frame, with the coordinate position with the maximum thermal value on the cell image as the center;

[0021] The detection module is used to extract the cell edge in each of the selection boxes, and detect the circular feature of the cell based on the cell edge;

[0022] The screening module is used to screen the cells in the selection box based on the detected circular features, count the number of red blood cells in each selection box, and use the selection box with the red blood cell number closest to the target number as the red blood cell detection area.

[0023] In a third aspect, a computer device is provided in this embodiment, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in the first aspect when executing the computer program.

[0024] In a fourth aspect, in this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described in the first aspect are implemented.

[0025] The above-mentioned red blood cell detection area selection method, device, computer equipment and computer-readable storage medium are used to select a red blood cell detection area from a cell image, obtain a prediction result of the cell distribution of the cell image by inputting the cell image into a pre-trained deep neural network, and generate a first thermodynamic map corresponding to the cell image based on the prediction result. In the first thermodynamic map, the more uniform the cell distribution in the cell image, the higher the corresponding thermodynamic value; respectively use preset N filter frames of different sizes to perform mean filtering and smoothing processing on the first thermodynamic map to obtain N second thermodynamic maps after filtering; wherein N is an integer greater than or equal to 1; extract the coordinate position with the largest thermodynamic value in each second thermodynamic map, and generate a selection box with the same size as the corresponding filter box with the coordinate position with the largest thermodynamic value on the cell image as the center; extract the cell edge in each selection box, and detect the circular feature of the cell based on the cell edge; screen the cells in the selection box based on the detected circular feature, count the number of red blood cells in each selection box, and use the selection box with the red blood cell number closest to the target number as the red blood cell detection area. This application introduces artificial intelligence technology into medical image analysis. Through deep learning algorithms, feature extraction and region division of red blood cell images can be performed, which can quickly and efficiently select red blood cell detection areas and effectively improve the efficiency of red blood cell analysis and detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0027] Figure 1This is an application scenario diagram of the method for selecting a red blood cell detection area provided in an embodiment of the present application;

[0028] Figure 2 is a flow chart of a method for selecting a red blood cell detection area according to an embodiment of the present application;

[0029] Figure 3 It is a schematic diagram of a first heat map obtained by processing a cell image using a pre-trained deep neural network in the red blood cell detection area selection method provided in an embodiment of the present application.

[0030] Figure 4 It is a schematic diagram of the effect of detecting the circular feature of cells based on the cell edge in the red blood cell detection area selection method provided in an embodiment of the present application;

[0031] Figure 5 A schematic diagram of the structure of a red blood cell detection area selection device provided in an embodiment of the present application;

[0032] Figure 6 A schematic diagram of the structure of a computer device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0034] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0035] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0036] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the" and the like involved in this application do not indicate a quantity limitation and may indicate the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0037] Figure 1 This is an application scenario diagram of the method for selecting a red blood cell detection area provided in one embodiment of the present application. Figure 1As shown, data can be transmitted between the server 101 and the mobile terminal 102 through the network. The mobile terminal 102 is used to collect cell images and transmit the cell images to the server 101. After receiving the cell images, the server 101 inputs the cell images into the pre-trained deep neural network to obtain the prediction results of the cell distribution of the cell images, and generates a first thermal map corresponding to the cell image based on the prediction results. In the first thermal map, the more uniform the cell distribution in the cell image, the higher the corresponding thermal value; the server 101 uses N preset filter frames of different sizes to perform mean filtering and smoothing processing on the first thermal map, and obtains N second thermal maps after filtering processing. ; Wherein, N is an integer greater than or equal to 1; the server 101 then extracts the coordinate position with the largest thermal value in each second thermal map, and generates a selection box with the same size as the corresponding filter box with the coordinate position with the largest thermal value on the cell image as the center; then the server 101 extracts the cell edge in each selection box, and detects the circular feature of the cell based on the cell edge; finally, the server 101 screens the cells in the selection box based on the detected circular feature, counts the number of red blood cells in each selection box, and uses the selection box with the red blood cell number closest to the target number as the red blood cell detection area. Wherein, the server 101 can be implemented by an independent server or a server cluster composed of multiple servers, and the mobile terminal 102 can be any image acquisition device.

[0038] The present application embodiment provides a method for selecting a red blood cell detection area, which is used to select a red blood cell detection area from a cell image. Figure 2 As shown, the method comprises the following steps:

[0039] Step S210, input the cell image into a pre-trained deep neural network, obtain a prediction result of the cell distribution of the cell image, and generate a first thermodynamic map corresponding to the cell image based on the prediction result. In the first thermodynamic map, the more uniform the cell distribution in the cell image, the higher the corresponding thermodynamic value.

[0040] Specifically, the pre-trained deep neural network in the present application can be based on the MobileNetV3 small network model as a prototype, and the MobileNetV3 small network model is trained using a large number of cell images to obtain the pre-trained deep neural network in the present application. In the present application, it is assumed that the pre-trained neural network model is represented by F(⋅;θ), where θ represents the network parameters; the cell image is represented by I, then the pre-trained deep neural network F(⋅;θ) is used to process the cell image I, and the generated first heat map H is represented as:

[0041]

[0042] The deep learning network is optimized by the maximum entropy loss function, which has the form:

[0043]

[0044] in, and are the true labels and predicted values.

[0045] In addition, in this application, a pre-trained deep neural network is used to predict the cell image. The schematic diagram of the first heat map generated based on the prediction result is shown in Figure 3 ,exist Figure 3 In the figure, the lighter the color, the higher the corresponding thermal value, which means the cell distribution is more uniform.

[0046] Step S220, using preset N filter frames of different sizes to perform mean filtering and smoothing processing on the first heat map, to obtain N second heat maps after filtering processing; wherein N is an integer greater than or equal to 1.

[0047] Specifically, the filter frame in the present application is a rectangular frame. Assuming that the filter frame in the present application is represented by K*K, the filter frame is used to perform mean filtering and smoothing processing on the first heat map H to obtain N second heat maps H' represented as follows:

[0048]

[0049] Step S230, extracting the coordinate position with the maximum thermal value in each second thermal map, and generating a selection box with the same size as the corresponding filter box with the coordinate position with the maximum thermal value on the cell image as the center.

[0050] Specifically, extract the coordinate position (x k， y k ) is calculated as follows:

[0051]

[0052] The coordinate position (x k， y k ) as the center to generate a selection box with the same size as the corresponding filter box as B k Finally, we get a set of N selection boxes of different sizes {B1, B2, ..., B N},in:

[0053]

[0054] Step S240, extracting the cell edge in each selection box, and detecting the circular feature of the cell based on the cell edge.

[0055] Specifically, the Canny edge detection algorithm can be used to extract the cell edge in each selection box, and the circular feature of the cell can be detected based on the cell edge by Hough transform. The effect diagram of the circular feature of the cell based on the cell edge detection is shown in FIG. Figure 4 .

[0056] Step S250, screening the cells in the selection box based on the detected circular features, counting the number of red blood cells in each selection box, and taking the selection box with the red blood cell number closest to the target number as the red blood cell detection area.

[0057] In this application, the cell radius can be determined based on the detected circular feature, and the cells in the selection box can be screened according to whether the cell radius meets the radius requirement of the red blood cells, and then the number of red blood cells in each selection box is counted, and finally the selection box with the red blood cell number closest to the target number is used as the red blood cell detection area. Specifically, it is assumed that the parameter equation based on the detected circular feature is expressed as:

[0058]

[0059] Assume that the red blood cell must satisfy the radius r min With r max The statistical formula for the number of red blood cells in each selection box is as follows:

[0060]

[0061] Furthermore, as one implementation method, the target number N may be returned. target Closest selection box B k , the error function is defined as follows:

[0062]

[0063] In the prior art, a suitable red blood cell detection area is selected by manual observation and counting, but manual observation and counting are easily affected by the subjective judgment of the observer and are time-consuming, resulting in the problem of low efficiency of red blood cell analysis and detection in the prior art.

[0064] In order to solve the above problems, the present application provides a method for selecting a red blood cell detection area, which is used to select a red blood cell detection area from a cell image, and obtains a prediction result of the cell distribution of the cell image by inputting the cell image into a pre-trained deep neural network, and generates a first thermodynamic map corresponding to the cell image based on the prediction result. In the first thermodynamic map, the more uniform the cell distribution in the cell image, the higher the corresponding thermodynamic value; the first thermodynamic map is respectively subjected to mean filtering and smoothing processing using preset N filter frames of different sizes to obtain N second thermodynamic maps after filtering processing; wherein N is an integer greater than or equal to 1; the coordinate position with the largest thermodynamic value in each second thermodynamic map is extracted, and a selection box with the same size as the corresponding filter box is generated with the coordinate position with the largest thermodynamic value on the cell image as the center; the cell edge in each selection box is extracted, and the circular feature of the cell is detected based on the cell edge; the cells in the selection box are screened based on the detected circular feature, the number of red blood cells in each selection box is counted, and the selection box with the red blood cell number closest to the target number is used as the red blood cell detection area. This application introduces artificial intelligence technology into medical image analysis. Through deep learning algorithms, feature extraction and region division of red blood cell images can be performed, which can quickly and efficiently select red blood cell detection areas and effectively improve the efficiency of red blood cell analysis and detection.

[0065] Figure 5 is a schematic diagram of a device for selecting a red blood cell detection area according to an embodiment of the present invention. Figure 5 As shown, a red blood cell detection area selection device 30 is provided, the device includes a generation module 31, a filtering module 32, an extraction module 33, a detection module 34 and a screening module 35;

[0066] A generating module 31 is used to input the cell image into a pre-trained deep neural network, obtain a prediction result of the cell distribution of the cell image, and generate a first thermal map corresponding to the cell image based on the prediction result. In the first thermal map, a region in the cell image with more uniform cell distribution has a higher corresponding thermal value;

[0067] A filtering module 32, used to perform mean filtering and smoothing processing on the first heat map using preset N filter frames of different sizes to obtain N second heat maps after filtering processing; wherein N is an integer greater than or equal to 1;

[0068] An extraction module 33 is used to extract the coordinate position with the maximum thermal value in each second thermal map, and generate a selection box with the same size as the corresponding filter box with the coordinate position with the maximum thermal value on the cell image as the center;

[0069] A detection module 34 is used to extract the cell edge in each selection frame and detect the circular feature of the cell based on the cell edge;

[0070] The screening module 35 is used to screen the cells in the selection box based on the detected circular features, count the number of red blood cells in each selection box, and use the selection box with the red blood cell number closest to the target number as the red blood cell detection area.

[0071] The above-mentioned red blood cell detection area selection device 30 is used to select a red blood cell detection area from a cell image, and obtain a prediction result of the cell distribution of the cell image by inputting the cell image into a pre-trained deep neural network, and generate a first thermodynamic map corresponding to the cell image based on the prediction result. In the first thermodynamic map, the more uniform the cell distribution in the cell image, the higher the corresponding thermodynamic value; the first thermodynamic map is respectively subjected to mean filtering and smoothing processing using preset N filter frames of different sizes to obtain N second thermodynamic maps after filtering processing; wherein N is an integer greater than or equal to 1; the coordinate position with the largest thermodynamic value in each second thermodynamic map is extracted, and a selection box with the same size as the corresponding filter box is generated with the coordinate position with the largest thermodynamic value on the cell image as the center; the cell edge in each selection box is extracted, and the circular feature of the cell is detected based on the cell edge; the cells in the selection box are screened based on the detected circular feature, the number of red blood cells in each selection box is counted, and the selection box with the red blood cell number closest to the target number is used as the red blood cell detection area. This application introduces artificial intelligence technology into medical image analysis. Through deep learning algorithms, feature extraction and region division of red blood cell images can be performed, which can quickly and efficiently select red blood cell detection areas and effectively improve the efficiency of red blood cell analysis and detection.

[0072] In one embodiment, the detection module 34 is further configured to extract the cell edges in each selection box using a Canny edge detection algorithm.

[0073] In one embodiment, the detection module 34 is further configured to detect circular features of cells based on cell edges by using Hough transform.

[0074] In one embodiment, the pre-trained deep neural network is based on the MobileNetV3 small network model.

[0075] It should be noted that the above modules can be functional modules or program modules, and can be implemented by software or hardware. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0076] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store a set of preset configuration information. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the above-mentioned red blood cell detection area selection method is implemented.

[0077] In one embodiment, a computer device is provided, which may be a terminal. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for selecting a red blood cell detection area is implemented. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball, or touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0078] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0079] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0080] Inputting the cell image into a pre-trained deep neural network, obtaining a prediction result of the cell distribution of the cell image, and generating a first thermal map corresponding to the cell image based on the prediction result, wherein in the first thermal map, a region in the cell image where the cell distribution is more uniform has a higher corresponding thermal value;

[0081] Using preset N filter frames of different sizes to perform mean filtering and smoothing processing on the first heat map, respectively, to obtain N second heat maps after filtering processing; wherein N is an integer greater than or equal to 1;

[0082] Extract the coordinate position with the largest thermal value in each second thermal map, and generate a selection box with the same size as the corresponding filter box, with the coordinate position with the largest thermal value on the cell image as the center;

[0083] Extract the cell edge in each selection box, and detect the circular feature of the cell based on the cell edge;

[0084] The cells in the selection box are screened based on the detected circular features, the number of red blood cells in each selection box is counted, and the selection box with the red blood cell number closest to the target number is used as the red blood cell detection area.

[0085] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0086] The Canny edge detection algorithm is used to extract the cell edges in each selection box.

[0087] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0088] The circular features of cells are detected based on cell edges using Hough transform.

[0089] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0090] The pre-trained deep neural network is based on the MobileNetV3 small network model.

[0091] The above-mentioned storage medium is used to select a red blood cell detection area from a cell image, obtain a prediction result of the cell distribution of the cell image by inputting the cell image into a pre-trained deep neural network, and generate a first thermodynamic map corresponding to the cell image based on the prediction result. In the first thermodynamic map, the more uniform the cell distribution in the cell image, the higher the corresponding thermodynamic value; respectively use preset N filter frames of different sizes to perform mean filtering and smoothing processing on the first thermodynamic map to obtain N second thermodynamic maps after filtering; wherein N is an integer greater than or equal to 1; extract the coordinate position with the largest thermodynamic value in each second thermodynamic map, and generate a selection box with the same size as the corresponding filter box with the coordinate position with the largest thermodynamic value on the cell image as the center; extract the cell edge in each selection box, and detect the circular feature of the cell based on the cell edge; screen the cells in the selection box based on the detected circular feature, count the number of red blood cells in each selection box, and use the selection box with the red blood cell number closest to the target number as the red blood cell detection area. This application introduces artificial intelligence technology into medical image analysis. Through deep learning algorithms, feature extraction and region division of red blood cell images can be performed, which can quickly and efficiently select red blood cell detection areas and effectively improve the efficiency of red blood cell analysis and detection.

[0092] It should be understood that the specific embodiments described herein are only used to explain the application, rather than to limit it. All other embodiments obtained by ordinary technicians in this field without creative work according to the embodiments provided in this application are within the protection scope of this application.

[0093] Obviously, the drawings are only some examples or embodiments of the present application. For ordinary technicians in the field, the present application can also be applied to other similar situations based on these drawings without creative work. In addition, it is understandable that although the work done in this development process may be complicated and lengthy, for ordinary technicians in the field, certain changes in design, manufacturing or production based on the technical content disclosed in this application are only conventional technical means and should not be regarded as insufficient content disclosed in this application.

[0094] The term "embodiment" in this application refers to a specific feature, structure or characteristic described in conjunction with the embodiment that can be included in at least one embodiment of the present application. The appearance of this phrase in various locations in the specification does not necessarily mean the same embodiment, nor does it mean that it is mutually exclusive with other embodiments and is independent or optional. It is clearly or implicitly understood by those of ordinary skill in the art that the embodiments described in this application can be combined with other embodiments without conflict.

[0095] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of patent protection. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the scope of protection of the present application. Therefore, the scope of protection of the present application shall be subject to the attached claims.

Claims

1. A method for selecting a red blood cell detection area, characterized in that: The method is used to select a red blood cell detection area from a cell image, and the method comprises the following steps: Inputting the cell image into a pre-trained deep neural network, obtaining a prediction result of the cell distribution of the cell image, and generating a first thermodynamic map corresponding to the cell image based on the prediction result, wherein a region in the cell image with more uniform cell distribution has a higher corresponding thermodynamic value; Using preset N filter frames of different sizes to perform mean filtering and smoothing processing on the first heat map, respectively, to obtain N second heat maps after filtering processing; wherein N is an integer greater than or equal to 1; Extract the coordinate position with the largest thermal value in each of the second thermal maps, and generate a selection box with the same size as the corresponding filter box with the coordinate position with the largest thermal value on the cell image as the center; Extracting the cell edge in each of the selection boxes, and detecting the circular feature of the cell based on the cell edge; The cells in the selection box are screened based on the detected circular features, the number of red blood cells in each selection box is counted, and the selection box with the red blood cell number closest to the target number is used as the red blood cell detection area.

2. The method for selecting a red blood cell detection area according to claim 1, characterized in that: The step of extracting the cell edge in each selection box comprises the following steps: The Canny edge detection algorithm is used to extract the cell edges in each selection box.

3. The method for selecting a red blood cell detection area according to claim 1, characterized in that: The method of detecting the circular feature of cells based on cell edges comprises the following steps: The circular features of cells are detected based on cell edges using Hough transform.

4. The method for selecting a red blood cell detection area according to claim 1, characterized in that: The pre-trained deep neural network is based on the MobileNetV3 small network model.

5. A red blood cell detection area selection device, characterized in that: The device comprises a generation module, a filtering module, an extraction module, a detection module and a screening module; The generating module is used to input the cell image into a pre-trained deep neural network, obtain a prediction result of the cell distribution of the cell image, and generate a first thermodynamic map corresponding to the cell image based on the prediction result, wherein in the first thermodynamic map, a region in the cell image where the cell distribution is more uniform has a higher corresponding thermodynamic value; The filtering module is used to perform mean filtering and smoothing processing on the first heat map using preset N filter frames of different sizes to obtain N second heat maps after filtering processing; wherein N is an integer greater than or equal to 1; The extraction module is used to extract the coordinate position with the maximum thermal value in each of the second thermal maps, and to generate a selection frame with the same size as the corresponding filter frame, with the coordinate position with the maximum thermal value on the cell image as the center; The detection module is used to extract the cell edge in each of the selection boxes, and detect the circular feature of the cell based on the cell edge; The screening module is used to screen the cells in the selection box based on the detected circular features, count the number of red blood cells in each selection box, and use the selection box with the red blood cell number closest to the target number as the red blood cell detection area.

6. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images