Phagocytic cell AI identification training, detection and proportion detection analysis method

Abstract

The phagocytic cell AI identification training, detection and proportion detection analysis method comprises the following steps: preprocessing a sample to obtain a microscopic sample; shooting a microscopic examination sample to obtain a microscopic examination sample image; identifying and labeling phagocytic cells in the microscopic examination sample image to obtain a phagocytic cell labeling picture; and AI training is carried out by using the labeled picture to obtain an AI identification feature data set A, and the AI identification feature data set A comprises phagocytic cell AI identification feature data. Identifying the microscopic examination sample image by using an AI identification algorithm, identifying phagocytic cells in hydrothorax and ascites in a selected area S1 of the image, obtaining the number NUM1 of the phagocytic cells, and obtaining the number NUM2 of phagocytic cells; in a microscopic examination sample, the content of phagocytic cells in unit volume is equal to NUM1 / V1. The ratio of phagocytic cells is equal to NUM1 / NUM2.

Description

Technical Field

[0001] This application belongs to the technical field of formed element analysis based on microscopic images, specifically involving AI recognition training, detection, and proportion detection and analysis methods for phagocytic cells. Background Technology

[0002] Phagocytes: A collective term for cells in the body that have phagocytic function, mainly including the mononuclear phagocyte system and neutrophils.

[0003] Mononuclear phagocytic system: The mononuclear phagocytic system includes mononuclear cells that are free in the blood and macrophages that develop after entering various tissues.

[0004] Neutrophils include neutrophils in the blood and neutrophils that enter tissues or other sites from the blood, including degenerated neutrophils.

[0005] Macrophages are a type of phagocyte. Human phagocytes come in two types: large and small. Small phagocytes are neutrophils found in peripheral blood.

[0006] Phagocytic cells: In this application, phagocytic cells that are currently phagocytizing or have already phagocytosed foreign objects are referred to as phagocytic cells. In this application, phagocytic cells refer to phagocytic cells that exhibit the phenomenon of phagocytizing foreign objects.

[0007] Phagocytosis: When pathogenic microorganisms or other foreign substances invade the body, they can activate phagocytes. Phagocytocytes are chemotactic and actively migrate and gather towards pathogens and foreign substances by generating active amoeboid movements. First, they adsorb the foreign substances onto the cell surface. Then, the cell membrane of the adsorption area invaginates inward, extends pseudopodia to surround the foreign substances, and engulfs the cytoplasm to form phagocytic vesicles. Subsequently, the primary lysosomes in the cytoplasm fuse with the phagocytic vesicles to form phagolysosomes, which kill the pathogens and digest and decompose the foreign substances.

[0008] Phagocytic index: The phagocytic index is an indicator of phagocytic cell activity. Specifically, it is calculated by dividing the total number of bacteria phagocytosed by 100 neutrophils by 100, yielding the average number of bacteria phagocytosed by each white blood cell. This index reflects the activity of phagocytic cells in vitro. The phagocytic index can be assessed by comparing the phagocytic cell activity in the blood of patients and healthy individuals; it is the ratio of the phagocytic cell activity in the blood of patients to that in healthy individuals (non-immune). The method for determining the phagocytic index involves incubating a blood sample with equivalent antigen particles (usually microbial cells), preparing a whole blood smear, and measuring the average number of particles phagocytosed by phagocytic cells. The phagocytic index is derived by calculating the ratio of the average values ​​of the two blood samples.

[0009] Generally, the normal range for the phagocytic index of macrophages is 1.058 ± 0.049, and the reference range for phagocytic index of neutrophils is 1.06 ± 0.05. The phagocytic index generally refers to the phagocytic index of phagocytes, which can be classified into small phagocytes and large phagocytes. Small phagocytes generally refer to neutrophils in the blood, while large phagocytes typically refer to monocytes in the blood and macrophages in various tissues and organs.

[0010] Phagocytes typically act as the body's second line of defense. When pathogens penetrate the skin and mucous membranes, the first line of defense, phagocytes can cross blood vessels, gather at the sites of bacteria and viruses, and rapidly eliminate these foreign components to maintain the stability of the internal environment. The phagocytic index reflects the phagocytic capacity of phagocytes, and thus, to some extent, reflects the body's immune function.

[0011] However, the determination of the phagocytic index is quite complex and difficult, requiring highly specialized experimental procedures.

[0012] The technical problem to be solved is how to get rid of the above complicated operations and obtain relevant clinical diagnostic information in a more efficient, convenient and cheap way.

[0013] The applicant has filed a series of Chinese patents, such as: 1. CN2020112669290, "Cell Analysis Method and System and Quantitative Method and System"; 2. CN2020112669182, "Cell Suspension Sample Imaging Method and System and Reagent Kit"; 3. CN2022104799126, "Rapid Focusing Method and Microscopic Image Acquisition Method for Microscopic Image Acquisition Device"; 4. CN2023109924201, "Urine Formed Elements Analysis Method System and Staining Preparation Device Reagent Preparation Method"; using novel technical solutions and leveraging artificial intelligence to measure the content of formed elements in blood, urine, and fecal samples. However, there is still considerable room for exploration regarding phagocytic cells. Summary of the Invention

[0014] The technical problem this application aims to solve is to overcome the shortcomings of existing technologies in accurately quantifying phagocytic cells. It proposes a method to capture magnified microscopic images of suspended samples under a microscope, and then train AI-based image analysis technology to train AI for phagocytic cell recognition, thereby achieving AI-based quantitative analysis of phagocytic cells. This provides accurate quantitative information on the degree of inflammatory response for clinical use.

[0015] This application proposes the concept of phagocytic cells, which refers to cells that are engulfing or have already engulfed foreign objects or cells. The activity of phagocytic cells is assessed by detecting the ratio of phagocytic cells to phagocytic cells. This indicator, compared with the phagocytic index, more directly reflects the overall activity of phagocytic cells.

[0016] In the tangible object detection technology system proposed by the applicant, the ratio of phagocytic cells to phagocytic cells is easier to measure than the phagocytic index.

[0017] The technical solution of this application to solve the above problems is a training method for AI recognition of phagocytic cells. The sample is preprocessed to obtain a microscopic sample; the microscopic sample is photographed to obtain a microscopic sample image; phagocytic cells in the microscopic sample image are identified and labeled to obtain labeled images of phagocytic cells; AI is trained using the labeled images to obtain an AI recognition feature dataset A, wherein the AI ​​recognition feature dataset A includes AI recognition feature data of phagocytic cells.

[0018] Phagocytes in microscopic sample images are identified and labeled to obtain labeled images of phagocytes; AI is trained using the labeled images to obtain the aforementioned AI recognition feature dataset A, which includes the phagocyte AI recognition feature dataset.

[0019] Sample pretreatment includes any one of the following technical features: TA1: Adding a staining agent to the sample for staining; TA2: Adding a dry powder staining agent to the sample; TA3: Adding reference particles to the sample; TA4: Adding a staining agent to the sample, including a new methylene blue component; TA5: Mixing multiple samples, allowing them to settle and separate into layers, and then taking the bottom layer to obtain a microscopic sample; TA6: Centrifuging the sample to separate into layers, enriching the layers, and taking the bottom layer to obtain a microscopic sample; TA7: Taking samples from organisms with increased phagocytic cells; TA8: Sending the above microscopic sample images to the AI ​​training server via the network; TA9: Sending the above microscopic sample images to the AI ​​training server via the network for manual identification and annotation of the segmented images; TA10: The above samples include any one of blood, urine, feces, and body cavity fluid.

[0020] The above samples were obtained from laboratory animals, including the following steps: Step B10: Injecting the laboratory animal with stimulant A1; Step B11: Waiting time T1; Step B30: Extracting the sample.

[0021] The steps preceding step B30 are as follows: Step B20: Injecting irritant A2 into the peritoneal cavity of the experimental animal; Step B21: Waiting time T2; Step B30: Extracting samples including any one or more of the following: peritoneal fluid, blood, urine, and feces.

[0022] The above samples were obtained from laboratory animals, specifically mice.

[0023] The above-mentioned AI recognition training method for phagocytic cells includes any one of the following technical features: TB1: T1 is greater than 10 hours and less than 30 hours; TB2: Stimulant A1 is 6% starch broth; TB3: Stimulant A1 contains trypan blue; TB4: Stimulant A2 is a suspension of xenobiotic red blood cells.

[0024] The technical solution of this application to solve the above problems can also be a method for detecting and analyzing phagocytic cells. The sample is preprocessed to obtain a microscopic sample. The microscopic sample image is obtained by taking a picture of the microscopic sample and selecting the image area S1. The volume corresponding to the image area S1 is V1. The AI ​​recognition algorithm is used to identify the above microscopic sample image and identify the phagocytic cells in the pleural and peritoneal fluid in the selected area S1 of the image, and obtain the number of phagocytic cells NUM1. The above AI recognition algorithm includes a phagocytic cell AI recognition feature dataset. In the microscopic sample, the content of phagocytic cells per unit volume = NUM1 / V1.

[0025] The AI ​​recognition algorithm was used to identify the above microscopic sample images and to identify phagocytes in the pleural and peritoneal fluid of the selected area S1 of the image, and to obtain the number of phagocytes NUM2; the above AI recognition algorithm includes a phagocyte AI recognition feature dataset; in the microscopic sample, the proportion of phagocytes = NUM1 / NUM2.

[0026] The technical solution of this application to solve the above problems can also be a method for detecting and analyzing the proportion of phagocytic cells. The sample is preprocessed to obtain a microscopic sample; the microscopic sample is photographed to obtain a microscopic sample image, and an image area S1 is selected; an AI recognition algorithm is used to identify the microscopic sample image, identifying phagocytic cells in the pleural and peritoneal fluid within the selected area S1 of the image, obtaining the number of phagocytic cells NUM1; an AI recognition algorithm is used to identify the microscopic sample image, identifying phagocytic cells in the pleural and peritoneal fluid within the selected area S1 of the image, obtaining the number of phagocytic cells NUM2; the AI ​​recognition algorithm includes a phagocytic cell AI recognition feature dataset; the proportion of phagocytic cells in the microscopic sample is NUM1 / NUM2.

[0027] The above analytical methods include any one of the following technical features: TG3: The above sample has been pretreated and its concentration has been concentrated N times. The content of phagocytic cells per unit volume of the sample = NUM1 / (V1×N); TG4: The above sample has been pretreated and its concentration has been diluted M times. The content of phagocytic cells per unit volume of the sample = (M×NUM1) / V1; TG5: The above phagocytic cells include any one or more of phagocytic neutrophils, phagocytic monocytes, and phagocytic macrophages; The above phagocytic cells include any one or more of neutrophils, monocytes, and macrophages.

[0028] The technical solution of this application to solve the above problems can also be a computing processing device, including any one of the following technical features: TH1: for running all or part of the above method; TH2: the memory of the above computing processing device includes all or part of the program code of the above method; TH3: the memory of the above computing processing device includes all or part of the data of the above method.

[0029] The technical solution of this application to solve the above problems can also be a data storage device, including any one of the following technical features: TK1: storing all or part of the program code for running the above method; TK2: storing all or part of the data of the above method; the data includes phagocytic cell AI recognition feature dataset and phagocytic cell AI recognition feature dataset.

[0030] The technical solution of this application to solve the above problems can also be a detection device used to run all or part of the above methods.

[0031] The beneficial technical effects of this application include that the AI ​​recognition feature dataset A includes AI recognition feature data of phagocytic cells, which can be used for targeted recognition training of phagocytic cells, facilitate data accumulation, and improve the accuracy of recognition.

[0032] The beneficial technical effects of this application include AI recognition feature dataset A, which includes a phagocyte AI recognition feature dataset, and separately labels and distinguishes phagocytes and phagocytes, and can accurately obtain the data set for AI recognition training of the two types of cells.

[0033] The beneficial technical effects of this application include that adding staining agents to prepare microscopic samples can stain and distinguish different formed elements, thereby improving the accuracy of identification.

[0034] The beneficial technical effects of this application include that the dry powder staining agent does not change the sample volume during staining, reducing the need for subsequent volume conversion steps and improving detection efficiency.

[0035] The beneficial technical effects of this application include that the addition of reference particles provides a reference for identification and counting, thereby improving the efficiency of subsequent identification and detection.

[0036] The beneficial technical effects of this application include that the staining agent includes a new methylene blue component, which can stain substances in the cell nucleus and improve the efficiency of identification and differentiation.

[0037] The beneficial technical effects of this application include that mixing multiple samples can improve the efficiency of AI recognition training for phagocytic cells and increase the probability of phagocytic cells appearing.

[0038] The beneficial technical effects of this application include improving the efficiency of AI recognition training for phagocytic cells and increasing the probability of phagocytic cells appearing by centrifuging and stratifying samples.

[0039] The beneficial technical effects of this application include: taking samples from biological individuals with an increase in phagocytic cells; and improving the efficiency of AI recognition training for phagocytic cells.

[0040] The beneficial technical effects of this application include that the aforementioned samples include any one of blood, urine, feces, and body cavity fluid, and different samples can correspond to corresponding AI recognition feature datasets A; such as blood AI recognition feature dataset A, urine AI recognition feature dataset A, feces AI recognition feature dataset A, and body cavity fluid AI recognition feature dataset A. Alternatively, various body fluids can be combined to form a larger AI recognition feature dataset A applicable to multiple body fluids.

[0041] The beneficial technical effects of this application include that by injecting stimulant A1 into the peritoneal cavity of experimental animals and obtaining samples after waiting for a corresponding period of time, the probability of phagocytic cells appearing is greatly increased, which can improve the efficiency of AI recognition training for phagocytic cells.

[0042] The beneficial technical effects of this application include that by injecting various foreign substances, such as stimulants A1 and A2, into the peritoneal cavity of experimental animals and obtaining samples after waiting for a corresponding period of time, the probability of phagocytic cells appearing is further increased, which can further improve the efficiency of AI recognition training for phagocytic cells.

[0043] The beneficial technical effects of this application include that the T1 time is long enough to ensure that the probability of phagocytic cells appearing is increased;

[0044] The beneficial technical effects of this application include that stimulant A1 contains trypan blue; it is easier to identify and can further improve the efficiency of AI recognition training for phagocytic cells.

[0045] The beneficial technical effects of this application include that stimulant A2, being a suspension of red blood cells from a different animal species, more easily increases the probability of phagocytic cells appearing, thereby further improving the efficiency of AI training for phagocytic cell recognition.

[0046] The beneficial technical effects of this application include that microscopic sample images are sent to an AI training server via the network, and the segmented images are manually identified and labeled via the network; this can further improve the training efficiency of AI recognition of phagocytic cells.

[0047] The beneficial technical effects of this application include the following: the content of phagocytic cells per unit volume = NUM1 / V1, which can provide quantitative analysis results of phagocytic cells and provide accurate quantitative reference information for clinical practice.

[0048] The beneficial technical effects of this application include the following: the proportion of phagocytic cells = NUM1 / NUM2, which further provides the proportion of phagocytic cells and phagocytes, providing accurate quantitative reference information for clinical use. Attached Figure Description

[0049] Figure 1 This is a flowchart illustrating the training method for AI recognition of phagocytic cells. Figure 1 ;

[0050] Figure 2 This is a flowchart illustrating the training method for AI recognition of phagocytic cells. Figure 2 ;

[0051] Figure 3 This is a flowchart illustrating the training method for AI recognition of phagocytic cells. Figure 3 ;

[0052] Figures 4 to 9 This is a schematic diagram of the sample preprocessing process. Figures 1 to 6 ;

[0053] Figure 10 This is a flowchart illustrating the phagocytic cell detection and analysis method. Figure 1 ;

[0054] Figure 11 This is a flowchart illustrating the method for detecting and analyzing the proportion of phagocytic cells. Figure 2 ;

[0055] Figure 12 This is a magnified microscopic illustration of phagocytic neutrophils, a type of phagocytic cell. Figure 1 ;

[0056] Figure 13 This is a magnified microscopic illustration of phagocytic neutrophils, a type of phagocytic cell. Figure 2 ;

[0057] Figure 14 This is a magnified microscopic illustration of a phagocytic macrophage, a type of phagocytic cell. Figure 1 ;

[0058] Figure 15 This is a magnified microscopic illustration of a phagocytic macrophage, a type of phagocytic cell. Figure 2 ;

[0059] Figure 16 This is a magnified microscopic illustration of a phagocytic macrophage, a type of phagocytic cell. Figure 3 ;

[0060] Figure 17 This is a magnified microscopic illustration of a phagocytic macrophage, a type of phagocytic cell. Figure 4 . Detailed Implementation

[0061] The embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0062] In current techniques, cytological examination of pleural effusion and ascites in clinical practice requires samples to be centrifuged at low speed for 5 minutes, followed by staining and microscopic examination using Reggi's or Diff-Quick staining methods. This procedure is cumbersome and time-consuming. Clinical staining and microscopic examination of serous cavity effusions can only determine the presence of phagocytic cells and estimate their proportion, making it difficult to accurately count cells and calculate the precise percentage of phagocytic cells.

[0063] Human phagocytes come in two types: small and large. Small phagocytes are neutrophils found in peripheral blood. Large phagocytes are monocytes in the blood and macrophages in various organs and tissues. Phagocytes have the function of engulfing and digesting aging and dead cells and foreign substances in the body, and are one of the important mechanisms of the body's natural defense. Phagocytosis is an important experiment in medical microbiology and immunology. It is mainly used to observe the ability of neutrophils and macrophages to engulf foreign microorganisms and other particles, thereby reflecting the body's natural immune defense function.

[0064] Detecting phagocyte function plays an important role in assessing macrophage function and understanding the body's specific and non-specific immune functions.

[0065] Phagocytic cells and phagocytes can reflect the ability of phagocytes to engulf foreign substances and reflect the level of the body's immune capacity; phagocytosis rate (%) = number of phagocytic cells / number of counted phagocytes × 100.

[0066] Impaired phagocytic function of phagocytes: This is mainly seen in various malignant tumors, with phagocytosis rates often below 45%. It can increase after surgical resection and improvement, thus serving as a reference indicator for the diagnosis of malignant tumors and the efficacy of radiotherapy, chemotherapy, and immunotherapy. Decreased phagocytic rate of phagocytes: In some patients with weakened immune function, the phagocytic rate of macrophages is reduced, which can be used as an indicator for predicting infection occurrence, monitoring treatment efficacy, and assessing prognosis.

[0067] like Figure 1 and Figure 3In one embodiment of a training method for AI recognition of phagocytes, the sample is preprocessed to obtain a microscopic sample; the microscopic sample is photographed to obtain a microscopic sample image; phagocytes in the microscopic sample image are identified and labeled to obtain labeled images of phagocytes; AI is trained using the labeled images to obtain an AI recognition feature dataset A, wherein the AI ​​recognition feature dataset A includes AI recognition feature data of phagocytes.

[0068] It should be noted that in this application, after the microscopic sample preprocessing, the image is obtained by taking pictures of the microscopic sample in a suspended state, and the formed element analysis is performed based on the acquired image.

[0069] like Figure 2 and Figure 3 In one embodiment of a training method for AI recognition of phagocytes, phagocytes in microscopic sample images are identified and labeled to obtain labeled images of phagocytes; AI training is performed using the labeled images to obtain the aforementioned AI recognition feature dataset A, which includes a phagocyte AI recognition feature dataset.

[0070] like Figure 9 In one embodiment of a training method for AI recognition of phagocytic cells, a staining agent is added to the sample to stain it; the staining agent is a liquid staining agent.

[0071] like Figure 4 In one embodiment of an AI recognition training method for phagocytic cells, a staining agent is added to the sample; the staining agent is a dry powder staining agent.

[0072] like Figure 5 In one embodiment of a training method for AI-based recognition of phagocytic cells, after the sample is allowed to settle and separate into layers, the bottom layer is taken as an intermediate sample, and reference particles are added to the intermediate sample; a dry powder staining agent is added to the sample. The staining agent includes a new methylene blue component.

[0073] like Figure 6 In one embodiment of an AI recognition training method for phagocytic cells, a sample of an individual organism with an increased number of phagocytic cells is taken; the sample is centrifuged and layered, and enriched by centrifugation or static sedimentation and layering, and the bottom layer is taken to obtain a microscopic sample.

[0074] In some embodiments of the AI ​​recognition training method for phagocytic cells not shown in the accompanying figures, the microscopic sample images are sent to the AI ​​training server via the network, and the segmented images are manually identified and labeled via the network.

[0075] In an embodiment of the AI ​​recognition training method for phagocytic cells, the sample includes any one of blood, urine, feces, or body cavity fluid.

[0076] like Figure 7In one embodiment of a training method for AI recognition of phagocytic cells, the above-mentioned samples are from experimental animals, including the following steps: Step B10: Injecting stimulant A1 into the experimental animal; Step B11: Waiting time T1; Step B30: Extracting samples.

[0077] like Figure 7 In one embodiment of a training method for AI recognition of phagocytic cells, the above-mentioned samples are from experimental animals, including the following steps: Step B10: Injecting stimulant A1 into the experimental animal; Step B11: Waiting time T1; Step B20: Injecting stimulant A2 into the peritoneal cavity of the experimental animal; Step B21: Waiting time T2; Step B30: Extracting samples including any one or more of peritoneal fluid, blood, urine, and feces.

[0078] The above samples were obtained from laboratory animals, specifically mice. T1 was greater than 10 hours and less than 30 hours; Stimulant A1 was 6% starch broth; Stimulant A1 contained trypan blue; Stimulant A2 was a suspension of red blood cells from a different animal species.

[0079] like Figure 10 In one embodiment of a method for detecting and analyzing phagocytic cells, the sample is preprocessed to obtain a microscopic sample; the microscopic sample is photographed to obtain a microscopic sample image, and an image area S1 is selected; the volume corresponding to the image area S1 is V1; an AI recognition algorithm is used to identify the above microscopic sample image, and phagocytic cells in the pleural and peritoneal fluid within the selected image area S1 are identified, obtaining the number of phagocytic cells NUM1; the above AI recognition algorithm includes a phagocytic cell AI recognition feature dataset; in the microscopic sample, the content of phagocytic cells per unit volume = NUM1 / V1. Under the condition of taking the image, the formed elements in the sample are in a suspended state in the microscopic sample.

[0080] The AI ​​recognition algorithm was used to identify the above microscopic sample images and to identify phagocytes in the pleural and peritoneal fluid of the selected area S1 of the image, and to obtain the number of phagocytes NUM2; the above AI recognition algorithm includes a phagocyte AI recognition feature dataset; in the microscopic sample, the proportion of phagocytes = NUM1 / NUM2.

[0081] like Figure 11A method for detecting and analyzing the proportion of phagocytic cells is disclosed. The sample is preprocessed to obtain a microscopic sample; the microscopic sample is photographed to obtain an image, and an image area S1 is selected; an AI recognition algorithm is used to identify the phagocytic cells in the pleural and peritoneal fluid within the selected area S1, obtaining the number of phagocytic cells NUM1; the AI ​​recognition algorithm is used again to identify the phagocytic cells in the pleural and peritoneal fluid within the selected area S1, obtaining the number of phagocytic cells NUM2; the AI ​​recognition algorithm includes a phagocytic cell AI recognition feature dataset; the proportion of phagocytic cells in the microscopic sample is calculated as NUM1 / NUM2.

[0082] The aforementioned AI-based phagocytic cell recognition feature dataset includes the AI-based phagocytic cell recognition feature dataset; the aforementioned AI-based phagocytic cell recognition feature dataset includes the aforementioned AI-based phagocytic cell recognition feature dataset.

[0083] The above samples were pretreated to concentrate the concentration by N times. The content of phagocytic cells per unit volume in the sample = NUM1 / (V1×N).

[0084] The above samples were pretreated and diluted M times. The content of phagocytic cells per unit volume in the sample is (M×NUM1) / V.

[0085] The aforementioned phagocytic cells include any one or more of the following: phagocytic neutrophils, phagocytic monocytes, and phagocytic macrophages.

[0086] like Figure 12 and Figure 13 A magnified microscopic diagram of phagocytic neutrophils. The left side is before labeling, and the right side is after labeling, showing the phenomenon of neutrophils phagocytosing foreign objects.

[0087] like Figures 14 to 15 A magnified microscopic diagram of phagocytic macrophages. The left side is before labeling, and the right side is after labeling, showing the phenomenon of phagocytic macrophages engulfing foreign objects.

[0088] A computing processing device for running all or part of the above methods.

[0089] The memory of the aforementioned computing processing device contains all or part of the program code for the aforementioned method.

[0090] The memory of the aforementioned computing processing device includes all or part of the data described in the above-described method.

[0091] A data storage device stores all or part of the program code for running the above-described methods; stores all or part of the data for the above-described methods; the data includes a phagocytic cell AI recognition feature dataset and a phagocytic cell AI recognition feature dataset.

[0092] A detection device for performing all or part of the above methods.

[0093] In one specific embodiment, the AI ​​recognition training method for pleural and peritoneal phagocytic cells includes the following steps:

[0094] Step 1: Using the centrifugation enrichment method, the peritoneal or pleural fluid samples of animals with exudative peritoneal or pleural effusions are centrifuged and enriched. The supernatant is removed, and the bottom sediment is added to the staining reagent.

[0095] Step 2 involves photographing peritoneal or pleural fluid samples from animals with exudative peritoneal or pleural effusions, manually preprocessing the images to obtain a large number of phagocytic cell images, and establishing a phagocytic cell database.

[0096] Step 3 involves labeling the images in the phagocytic cell database.

[0097] Step four involves uploading the processed images to the AI ​​training server for training, thereby obtaining an AI recognition model capable of identifying a database of phagocytic cells in pleural and peritoneal effusions.

[0098] Step 5 uses the trained AI recognition model for phagocytic cells in pleural and peritoneal fluid to identify and segment subsequent raw images, automatically obtains information and coordinates of phagocytic cells, and uses image algorithms for automatic annotation.

[0099] Step Six: The AI ​​model automatically identifies and labels the data library, then uploads it to the AI ​​server for enhanced training of the AI ​​model, thereby improving AI performance.

[0100] Macrophages, monocytes, and neutrophils in higher animals possess phagocytic function. They are widely distributed in tissues and blood, playing a crucial role in the body's non-specific immune function. When pathogenic microorganisms or other foreign substances invade the body, they activate phagocytes. Phagocytes exhibit chemotaxis, actively migrating and gathering towards pathogens and foreign substances through active amoeboid movements. They first adsorb the foreign substance onto the cell surface, then the cell membrane in the adsorption area invaginates, extending pseudopodia to surround the foreign substance and engulf it in the cytoplasm, forming a phagocytic vesicle. Subsequently, primary lysosomes in the cytoplasm fuse with the phagocytic vesicle to form a phagolysosome, killing the pathogen and digesting and decomposing the foreign substance.

[0101] The macrophage phagocytosis assay includes the following steps: (1) 24 hours before the experiment, inject 3 ml of 6% starch broth (containing trypan blue) into the peritoneal cavity of the mouse. (2) During the experiment, inject 1 ml of 1% canine / cat red blood cell suspension into the peritoneal cavity and gently rub the abdomen to disperse the canine / cat red blood cell suspension. (3) 20-30 minutes later, use a syringe to perform peritoneal puncture on the mouse and extract peritoneal fluid samples to observe the phagocytic phenomenon of phagocytes. (4) After observing phagocytes, take 150 μl of peritoneal fluid and add it to the staining reagent, and mix thoroughly. (5) Take pictures, collect and process data, and train.

[0102] like Figure 16 A magnified microscopic diagram of phagocytic macrophages obtained through the above animal experiments; clearly showing the phenomenon of macrophages phagocytosing red blood cells.

[0103] like Figure 17 In diseased animals, the primary phagocytic cells also show obvious macrophage phagocytosis of red blood cells. In this application... Figures 12 to 17 The images shown are of phagocytic cells that are currently engulfing prey.

[0104] The serial numbers such as "first," "second," etc., in this application are for convenience of expression only and do not necessarily indicate a sequential relationship in terms of size or time. The letter numbers of the steps are also for convenience of expression only and do not necessarily indicate a sequential relationship in terms of time.

[0105] As shown in the accompanying drawings, the above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of the invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for training AI-based recognition of phagocytic cells, characterized in that, The samples were pretreated to obtain microscopic samples; Take microscopic images of the samples; Phagocytic cells in microscopic sample images are identified and labeled to obtain labeled images of phagocytic cells. The phagocytic cells are cells that are engulfing or have already engulfed foreign objects or cells; AI is trained using labeled images to obtain AI recognition feature dataset A, which includes AI recognition feature data of phagocytic cells.

2. The AI ​​recognition and training method for phagocytic cells according to claim 1, characterized in that, Phagocytes in microscopic sample images are identified and labeled to obtain labeled images of phagocytes; The AI ​​recognition feature dataset A is obtained by training AI with labeled images, and the AI ​​recognition feature dataset A includes the phagocyte AI recognition feature dataset.

3. The AI ​​recognition and training method for phagocytic cells according to claim 1, characterized in that, Sample preprocessing includes any one of the following technical features: TA1: Add staining agent to the sample to stain it; TA2: Add staining agent to the sample; the staining agent is a dry powder staining agent. TA3: Add reference particles to the sample; TA4: Add staining agent to the sample, including neomethylene blue; TA5: Mix multiple samples, allow them to settle and separate into layers, and then take the bottom layer to obtain a microscopic sample. TA6: Centrifuge the sample into layers, enrich the layers, and take the bottom layer to obtain the microscopic sample; TA7: Samples were taken from organisms with an increased number of phagocytic cells; TA8: The microscopic sample images are sent to the AI ​​training server via the network; TA9: The microscopic sample images are sent to the AI ​​training server via the network, and the segmented images are manually identified and labeled via the network; TA10: The sample includes any one of blood, urine, feces, or body cavity fluid.

4. The AI ​​recognition and training method for phagocytic cells according to claim 1, characterized in that, The samples were obtained from laboratory animals, and the following steps were performed: Step B10: Inject stimulant A1 into the experimental animal; Step B11: Waiting time T1; Step B30: Sample extraction.

5. The AI ​​recognition and training method for phagocytic cells according to claim 4, characterized in that, The sample is derived from laboratory animals, and the following steps are included before step B30: Step B20: Inject irritant A2 into the peritoneal cavity of the experimental animal; Step B21: Waiting time T2; Step B30: Sample collection includes any one or more of the following: peritoneal fluid, blood, urine, and feces.

6. The AI ​​recognition and training method for phagocytic cells according to claim 4, characterized in that, The samples were obtained from laboratory animals, specifically mice.

7. The AI ​​recognition and training method for phagocytic cells according to claim 5, characterized in that, Includes any one of the following technical features: TB1: T1 is greater than 10 hours and less than 30 hours; TB2: Irritant A1 is 6% starch broth; TB3: Irritant A1 contains trypan blue; TB4: Irritant A2 is a suspension of red blood cells from a different animal species.

8. A method for detecting and analyzing phagocytic cells, characterized in that, The samples were pretreated to obtain microscopic samples; Take a microscopic examination of the sample to obtain a microscopic image of the sample, and select the image area S1; the volume corresponding to the image area S1 is V1. The AI ​​recognition algorithm was used to identify the microscopic sample image and the phagocytic cells in the pleural and peritoneal fluid of the selected area S1 of the image were identified to obtain the number of phagocytic cells NUM1. The AI ​​recognition algorithm includes a dataset of AI recognition features for phagocytic cells; In microscopic examination of samples, the content of phagocytic cells per unit volume = NUM1 / V1.

9. The method for detecting and analyzing phagocytic cells according to claim 8, characterized in that, The AI ​​recognition algorithm was used to identify the microscopic sample image, and the phagocytes in the pleural and peritoneal fluid of the selected area S1 of the image were identified to obtain the number of phagocytes NUM2. The AI ​​recognition algorithm includes a dataset of phagocyte AI recognition features; In microscopically examined samples, the percentage of phagocytic cells = NUM1 / NUM2.

10. A method for detecting and analyzing the proportion of phagocytic cells, characterized in that, The samples were pretreated to obtain microscopic samples; Take a microscopic examination of the sample to obtain an image of the sample, and select the image area S1; The AI ​​recognition algorithm was used to identify the microscopic sample image and the phagocytic cells in the pleural and peritoneal fluid of the selected area S1 of the image were identified to obtain the number of phagocytic cells NUM1. The AI ​​recognition algorithm was used to identify the microscopic sample image, and the phagocytes in the pleural and peritoneal fluid of the selected area S1 of the image were identified to obtain the number of phagocytes NUM2. The AI ​​recognition algorithm includes a dataset of AI recognition features for phagocytic cells; The AI ​​recognition algorithm includes a dataset of phagocyte AI recognition features; In microscopically examined samples, the percentage of phagocytic cells = NUM1 / NUM2.

11. The analytical method according to any one of claims 8 to 10, characterized in that, Includes any one of the following technical features: TG1: The phagocytic cell AI recognition feature dataset includes the phagocytic cell AI recognition feature dataset as described in any one of claims 1 to 7; TG2: The phagocyte AI recognition feature dataset includes the phagocyte AI recognition feature dataset as described in any one of claims 2 to 7; TG3: The sample has been pretreated and its concentration has been increased by N times. The content of phagocytic cells per unit volume in the sample = NUM1 / (V1×N); TG4: The sample was pretreated and diluted M times. The content of phagocytic cells per unit volume of the sample = (M×NUM1) / V1. TG5: The phagocytic cells include any one or more of the following: phagocytic neutrophils, phagocytic monocytes, and phagocytic macrophages; TG6: The phagocytes include any one or more of neutrophils, monocytes, and macrophages.

12. A computing processing device, characterized in that, Includes any one of the following technical features: TH1: Used to perform all or part of the method according to any one of claims 1 to 7; TH2: The memory of the computing processing device includes all or part of the program code for the method described in any one of claims 1 to 7; TH3: The memory of the computing processing device includes all or part of the data described in any one of claims 1 to 7.

13. A data storage device, characterized in that, Includes any one of the following technical features: TK1: Stores program code for running all or part of the method according to any one of claims 1 to 7; TK2: Stores all or part of the data of the method according to any one of claims 1 to 7; the data includes a phagocytic cell AI recognition feature dataset and a phagocytic cell AI recognition feature dataset.

14. A detection device, characterized in that, Used to perform all or part of the method according to any one of claims 1 to 11.

Images