Agglutinated platelet parameter, detection method and identification AI training method thereof

Abstract

In the agglutinated platelet parameter detection method and the identification AI training method thereof, a blood sample is preprocessed, and a microscopic examination sample is obtained and tiled; standing the platelets to sink to the bottom; shooting to obtain a microscopic examination sample image; according to the method, normal platelets and agglutinated platelets are identified and labeled, a labeled picture is obtained for AI training, an obtained AI feature data set A comprises normal platelet features, and an AI feature data set A comprises agglutinated platelet features. Identifying a microscopic examination sample image by using an AI identification algorithm, identifying agglutinated platelets in the blood sample in a selected area S1 of the image, and obtaining the total number NUMS1 of the agglutinated platelets in the blood sample in the selected image; and the AI recognition algorithm recognizes the agglutinated platelets in the blood sample according to the agglutinated platelet feature data set. The recognition of the agglutinated platelets is completed by utilizing advanced AI computing power, so that the analysis of the agglutinated platelets is more efficient and more accurate, and the dimension of platelet function analysis based on images is expanded.

Description

Technical Field

[0001] The present application belongs to the technical field of analysis of formed components based on microscopically magnified images, and particularly relates to an AI training method, calculation processing, and storage device for identifying aggregated platelets in blood samples, as well as an aggregated platelet parameter detection method, calculation processing, and storage device. Background Art

[0002] Platelets are a type of cell, a vital component of blood, primarily responsible for hemostasis and coagulation. Platelets are typically round or oval in shape, approximately 2 to 4 microns in diameter. They lack a nucleus but contain numerous organelles and platelet granules. Their primary function is to rapidly aggregate at the site of blood vessel damage, releasing substances such as platelet factor and platelet-derived growth factor, which promote vascular contraction and platelet aggregation, thereby forming a thrombus and preventing bleeding.

[0003] Platelets, however, are crucial to the body's hemostatic function. When blood is lost due to vascular trauma, the physiological hemostasis process of platelets can be roughly divided into two phases: the first phase primarily involves the rapid adhesion of platelets to the site of injury, where they aggregate into a soft hemostatic plug; the second phase primarily promotes blood coagulation and forms a firm hemostatic plug.

[0004] Platelets are crucial cells for maintaining the integrity of blood vessel walls and are the primary functional cells for early hemostasis. Clinically, platelet count is typically determined through routine blood tests, followed by other tests to determine the strength of platelet function. Clarifying platelet-related indicators will help doctors determine whether a patient is prone to bleeding or thrombosis, thereby guiding treatment. When platelet count decreases or function weakens, the body exhibits a tendency to spontaneously bleed, often associated with bleeding disorders such as essential thrombocytopenia and thrombasthenia. However, if platelet count increases or function strengthens, the body enters a hypercoagulable state, often associated with thrombotic disorders such as deep vein thrombosis, pulmonary embolism, and ischemic stroke.

[0005] The applicant filed a series of Chinese patents, such as

[0006] 1. CN2020112669290, “Cell analysis method and system and quantitative method and system”;

[0007] 2. CN2020112669182, “Cell Suspension Sample Imaging Method, System, and Kit”;

[0008] 3. CN2022104799126, “Fast focusing method for microscopic image acquisition device and microscopic image detection method”;

[0009] 4. CN2023100423151, “Blood sample imaging analysis system and method”;

[0010] 5. CN2023116834572, “AI training method and computational processing and storage device for leukocyte identification in blood samples” uses a new technical solution to measure the content of blood samples and formed elements in blood samples.

[0011] The above technical solution can identify and count normal platelets. However, it is difficult to accurately identify and count abnormal platelets. Abnormal platelet status often indicates platelet dysfunction and is highly valuable for the early prevention and diagnosis of cardiovascular and cerebrovascular diseases.

[0012] In the prior art, the detection of aggregated platelets is relatively difficult and complex. Generally, it is only a qualitative test without quantitative detection standards, making it difficult to quantitatively assess the severity of aggregated platelets. How to use the above technology to analyze aggregated platelets is a technical problem that this application needs to solve.

[0013] If AI recognition technology is used, not only can aggregated platelets be detected, but also aggregated platelets can be quantitatively detected. The technology in the series of patented technologies proposed by the applicant is used to quantitatively identify aggregated platelets. First, AI needs to be trained on aggregated platelets. How to obtain images of aggregated platelets in various states has become a technical difficulty, especially aggregated platelets in whole blood state, which is a technical problem that needs to be solved in this application.

[0014] Glossary:

[0015] Plasma (PRP) means: platelet-rich plasma (PRP).

[0016] The English abbreviation for aggregated platelets is APLT. Summary of the Invention

[0017] In this application, the inventors proposed that the AI ​​feature dataset A obtained by conducting AI training on images of blood sample suspensions containing aggregated platelets can effectively support AI detection for identifying aggregated platelets in blood samples, and use advanced AI computing power to complete the identification of aggregated platelets, making the analysis of aggregated platelets more efficient and accurate, and expanding the dimension of image-based platelet function analysis.

[0018] The technical solution of the present application to solve the above-mentioned technical problems is an AI training method for identifying aggregated platelets, wherein a blood sample is preprocessed to obtain a microscopic sample; the microscopic sample is flattened; the platelets are allowed to settle to the bottom; the flattened microscopic sample is photographed to obtain a microscopic sample image; normal platelets and aggregated platelets in the microscopic sample image are identified and labeled to obtain a labeled image, and the labeled image is used for AI training to obtain an AI feature data set A; the AI ​​feature data set A includes normal platelet features, and the AI ​​feature data set A includes aggregated platelet features.

[0019] Pretreatment of blood samples includes any of the following technical features: TA1: adding a dye to the blood sample, staining the blood sample, and diluting the blood sample after staining to obtain a microscopic sample; TA2: adding a liquid dye to the blood sample, staining and diluting the blood sample to obtain a microscopic sample; TA3: centrifuging and stratifying the blood sample to obtain a microscopic sample, wherein the microscopic sample is platelet-containing plasma (PRP); TA4: centrifuging and stratifying the blood sample to obtain an intermediate sample, wherein the intermediate sample is platelet-containing plasma (PRP), and the intermediate sample is added to the blood sample to obtain a microscopic sample; the above-mentioned microscopic sample contains red blood cells.

[0020] A platelet agonist is added to the blood sample, intermediate sample or / and microscopic sample, and the platelet agonist promotes platelet aggregation.

[0021] Platelet agonists include one or more of epinephrine, adenosine diphosphate (ADP), collagen, ristocetin, and arachidonic acid (AA).

[0022] The above-mentioned AI training method for identifying aggregated platelets includes any one of the following technical features: TB1: performing image filtering processing on the above-mentioned microscopic sample images to filter out the red blood cell images; TB2: performing image filtering processing on the above-mentioned microscopic sample images to filter out the red blood cell images, and filtering out the red blood cells based on the color difference between the red blood cell and platelet images; TB3: performing image filtering processing on the above-mentioned microscopic sample images to filter out the red blood cell images, and filtering out the red blood cells based on the grayscale difference between the red blood cell and platelet images.

[0023] Preprocessing of blood samples includes any of the following technical features: TC1: After the microscopic sample is prepared, it is photographed after waiting for a period of time TW; TC2: the above blood sample comes from multiple animal individuals; TC3: multiple blood samples are mixed to obtain a microscopic sample; TC4: the above sample comes from one or more individuals with aggregated thrombocytosis.

[0024] Taking AI feature data set A as the feature data set, AI software is used to identify one or more images input into AI recognition to obtain recognition output images, which include recognized normal platelet images and agglutinated platelet images. The recognized output agglutinated platelet images are manually reviewed to form manual review mark images, which include marked normal platelet images and agglutinated platelet images. AI training is performed using manual review mark images to obtain AI feature data set B, which includes normal platelet features and agglutinated platelet features. The above-mentioned AI software runs locally or on a network server, and the above-mentioned feature data sets are placed locally or on a network server.

[0025] The technical solution of the present application can also be a computing and processing device, including any one of the following technical features: TD1: all or part of the above-mentioned AI training method for identifying aggregated platelets; TD2: the memory of the above-mentioned computing and processing device includes the above-mentioned AI feature data set A; TD3: the memory of the above-mentioned computing and processing device includes the above-mentioned AI feature data set B.

[0026] The technical solution of the present application may also be a data storage device, comprising any one of the following technical features: TE1: storing program code for executing all or part of the aforementioned AI training method for identifying aggregated platelets; TE2: storing the aforementioned AI feature dataset A; TE3: storing the aforementioned AI feature dataset B. A detection device, configured to execute part or all of the aforementioned AI training method for identifying aggregated platelets.

[0027] The technical solution of the present application can also be a method for detecting parameters of aggregated platelets, wherein a blood sample is preprocessed to obtain a microscopic sample; the microscopic sample is flattened; the flattened microscopic sample is photographed to obtain a microscopic sample image; the microscopic sample image is identified using an AI recognition algorithm to identify aggregated platelets in the blood sample in a selected area S1 of the image, and the total number NUMS1 of aggregated platelets in the blood sample in the selected image is obtained; the AI ​​recognition algorithm identifies aggregated platelets in the blood sample based on an aggregated platelet feature data set.

[0028] The above-mentioned aggregated platelet feature dataset is obtained by training after manually annotating aggregated platelet images; or the aggregated platelet feature dataset is obtained by training after images of normal platelets are obtained after being stimulated and aggregated by platelet agonists.

[0029] The above-mentioned platelet aggregation parameter detection method includes any one of the following technical features: TG1: the platelet aggregation parameter is the number of platelets in a unit volume of blood sample, and the blood sample has been pre-treated by dilution and diluted N times; the tile height of the microscopic sample is H; the volume V corresponding to the total number of platelets NUMS1 in the blood sample is = the selected area S1×H; the number of platelets in a unit volume of blood sample is = N×the total number of platelets in the blood sample is NUMS1 / the volume of the blood sample is V; TG2: the platelet aggregation parameter is the number of platelets in a unit volume of blood sample, and the blood sample has been pre-treated by concentration and concentrated N times, and the tile height of the microscopic sample is H; the volume V corresponding to the total number of platelets NUMS1 in the blood sample is = the selected area S1×H; the number of platelets in a unit volume of blood sample is = the total number of platelets in the blood sample is NUMS1 / (the volume of the blood sample is V×N).

[0030] An AI recognition algorithm is used to identify the above-mentioned microscopic sample image, identify normal platelets in the blood sample in the selected area S1 of the image, and obtain the total number NUMS2 of normal platelets in the blood sample in the selected image; the above-mentioned AI recognition algorithm identifies normal platelets in the blood sample based on a characteristic data set of normal platelets in the blood sample; the ratio of the number of aggregated platelets to the number of normal platelets = NUM1 / NUM2.

[0031] The blood sample pretreatment includes diluting the blood sample with a liquid dye to obtain a microscopic examination sample, where the dilution multiple is greater than 50 and less than 800.

[0032] Complete microscopic examination and photography within time T13 after the microscopic examination sample is prepared.

[0033] The above-mentioned method for detecting parameters of aggregated platelets includes any one of the following technical features: TK1: performing image filtering processing on the above-mentioned microscopic sample image to filter out the red blood cell image; TK2: performing image filtering processing on the above-mentioned microscopic sample image to filter out the red blood cell image, and filtering out the red blood cells based on the color difference between the red blood cell and platelet images; TK3: performing image filtering processing on the above-mentioned microscopic sample image to filter out the red blood cell image, and filtering out the red blood cells based on the grayscale difference between the red blood cell and platelet images.

[0034] The technical solution of the present application may also include a computing device for executing all or part of the above-mentioned platelet aggregation parameter detection method, a data storage device for storing all or part of the program code for executing the above-mentioned platelet aggregation parameter detection method, and a detection device for executing part or all of the above-mentioned platelet aggregation parameter detection method.

[0035] The technical solution of the present application can also be a platelet aggregation parameter, which is used to evaluate the number of aggregated platelets in a blood sample. The number of aggregated platelets in a unit volume of blood sample = the number of aggregated platelets in the blood sample per unit volume of blood sample / the volume of blood sample.

[0036] The technical solution of the present application can also be an aggregation platelet parameter, which is used to evaluate the number of aggregated platelets in a blood sample. The aggregated platelet ratio = the number of aggregated platelets in a blood sample of the same volume / the number of normal platelets.

[0037] The technical effects of the above technical solution include: improving the efficiency of AI training data collection through diversified means and improving the efficiency of AI evolution.

[0038] The technical effects of the above technical solution include: using advanced AI computing power to complete the identification of aggregated platelets, making the analysis of aggregated platelets more efficient and more accurate.

[0039] The technical effects of the above technical solution include: encapsulating the trained AI feature data set A including the characteristics of agglutinated platelets into independent data, which can be sold separately, reducing the engineering difficulty of the entire industry and allowing new technologies to be promoted and implemented as soon as possible.

[0040] The technical effects of the above technical solution include: improving the image collection efficiency of aggregated platelets through a variety of preprocessing methods, which is conducive to accelerating the AI ​​training iteration process and improving efficiency.

[0041] The technical effects of the above technical solution include: the purpose of blood sample staining is to detect other cells in the blood sample, such as white blood cells. When training the AI ​​model, using the same staining can improve the accuracy of simultaneous identification of multiple cells and further improve system efficiency.

[0042] The technical effects of the above technical solution include: direct addition of liquid dye, which can achieve staining and dilution at the same time, and high sample preparation efficiency.

[0043] The technical effects of the above technical solution include: obtaining platelet-containing plasma (PRP) by centrifugal stratification, and improving the image collection efficiency of aggregated platelets.

[0044] The technical effects of the above technical solution include: platelet-containing plasma (PRP) is added to the blood sample as an intermediate sample to obtain a microscopic sample; the microscopic sample contains red blood cells, which can actively control the number and ratio of red blood cells, improve the compatibility and adaptability of training images, and various red blood cell concentration states can be set.

[0045] The technical effects of the above-mentioned technical solution include: the addition of platelet agonists further improves the efficiency of collecting images of aggregated platelets. AI training for aggregated platelets can be performed based on normal blood samples. It also facilitates providing samples with a variety of different aggregated platelet concentrations. The concentration of aggregated platelets can be controlled by adding different amounts of platelet agonists, thereby obtaining a rich variety of training samples in different states. Furthermore, it is applicable to a variety of platelet agonists, making them easily accessible.

[0046] The technical effects of the above technical solution include: filtering out red blood cell images, highlighting the presentation of aggregated platelets in the image, and improving the efficiency of AI training.

[0047] The technical effects of the above technical solution include: after the microscopic sample is prepared, it is photographed after waiting for a period of time TW. After waiting for a long enough time, the agonist can fully exert its effect and produce enough aggregated platelets for AI training.

[0048] The technical benefits of the above-mentioned technical solution include: by mixing blood samples from multiple individual animals to obtain microscopic samples, the efficiency of obtaining aggregated platelets is improved. Since not all samples contain aggregated platelets, the above-mentioned method can improve the efficiency of obtaining characteristic data.

[0049] The technical effects of the above technical solution include: the blood sample is derived from one or more individuals with increased agglutinated platelets, which can further improve the efficiency of obtaining agglutinated platelets.

[0050] The technical effects of the above technical solution include: the feature data set B evolved based on the AI ​​feature data set A is encapsulated into independent data, so that the ability of AI recognition of agglomerated platelets can be iterated and improved.

[0051] The technical effects of the above technical solution include: manually reviewing and marking images, including marking normal red blood cell images and marking agglutinated platelet images, which can reduce the probability of misidentifying normal red blood cells as agglutinated platelets and improve the accuracy of recognition.

[0052] The technical effects of the above technical solution include: the total number NUMS1 of aggregated platelets in the blood sample in the selected image is given, the basic parameters for quantitative analysis of aggregated platelets are given, and the AI ​​recognition algorithm is used, with high efficiency and accuracy.

[0053] The technical effects of the above technical solution include: the feature dataset of aggregated platelets in blood samples is obtained by training after manually annotating the images of aggregated platelets in blood samples, and the accuracy of AI recognition is more guaranteed.

[0054] The technical effects of the above technical solution include: the number of aggregated platelets in a unit volume of blood sample, based on classification recognition, combined with image-based sample quantification, to achieve accurate aggregated platelet counting, which takes the accurate quantitative analysis of formed elements in blood samples a step further, and can perform precise quantitative analysis of aggregated platelets, making the formed elements of blood samples that can be accurately quantitatively analyzed richer and more diverse.

[0055] The technical benefits of the above-mentioned technical solution include: calculating the percentage of aggregated platelets in a blood sample = the number of aggregated platelets in the same volume of blood sample / the number of normal platelets. This takes the analysis of formed elements in blood samples a step further, enabling accurate quantitative analysis of the percentage of aggregated platelets. This provides a more dimensional quantitative analysis basis for determining the severity of the disease. It also provides clinicians with comprehensive red blood cell analysis data, enriching the clinical data reference set. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is one of the steps in the AI ​​training method for identifying aggregated platelets;

[0057] Figure 2 This is the second step of the AI ​​training method for identifying aggregated platelets;

[0058] Figure 3 This is the third step of the AI ​​training method for identifying aggregated platelets;

[0059] Figure 4 This is one of the schematic diagrams of the blood sample pretreatment method;

[0060] Figure 5 This is the second diagram of the blood sample pretreatment method;

[0061] Figure 6 This is the third diagram of the blood sample pretreatment method;

[0062] Figure 7 This is the fourth diagram of the blood sample pretreatment method;

[0063] Figure 8 This is the fifth diagram of the blood sample pretreatment method;

[0064] Figure 9 This is the sixth diagram of the blood sample pretreatment method;

[0065] Figure 10 This is the seventh diagram of the blood sample pretreatment method;

[0066] Figure 11 This is the seventh diagram of the blood sample pretreatment method;

[0067] Figure 12This is one of the schematic diagrams of a blood sample containing aggregated platelets when imaged in a detection chip;

[0068] Figure 13 This is one of the schematic diagrams of the method for detecting parameters of aggregated platelets;

[0069] Figure 14 This is the second schematic diagram of the method for detecting platelet aggregation parameters;

[0070] Figure 15 This is the second schematic diagram of a blood sample containing aggregated platelets when imaged in a detection chip;

[0071] Figure 16 It is a schematic diagram of the principle of measuring platelet function in the prior art. DETAILED DESCRIPTION

[0072] The contents of this application are further described in detail below in conjunction with the accompanying drawings. It should be noted that the following is a description of the preferred embodiments of the present invention and does not constitute any limitation to the present invention. The description of the preferred embodiments of the present invention is only an illustration of the general principles of the present invention. The numbers such as "first", "second" and "A" and "B" involved in the present invention are only for the convenience of explanation and do not represent the order relationship in time or space. The combination of letters and numbers "TA", "TB" and "H" involved in the present invention are only for the convenience of explanation, and the specific meaning is determined by the specific words referred to.

[0073] Platelets are crucial cells for maintaining the integrity of blood vessel walls and are the primary functional cells for early hemostasis. Clinically, platelet count is typically determined through routine blood tests, followed by other tests to determine the strength of platelet function. Clarifying platelet-related indicators will help doctors determine whether a patient is prone to bleeding or thrombosis, thereby guiding treatment. When platelet count decreases or function weakens, the body exhibits a tendency to spontaneously bleed, often associated with bleeding disorders such as essential thrombocytopenia and thrombasthenia. However, if platelet count increases or function strengthens, the body enters a hypercoagulable state, often associated with thrombotic disorders such as deep vein thrombosis, pulmonary embolism, and ischemic stroke.

[0074] Aggregated platelets or platelets refer to platelets that appear aggregated in the blood, and their clinical significance is usually related to bleeding and coagulation abnormalities. The formation of aggregated platelets may lead to thrombosis or bleeding tendency. The number of aggregated platelets is greater than the upper limit of the reference range, indicating that the sample has coagulation, essential thrombocythemia, or chronic cell leukemia. Coagulation dysfunction: Aggregated platelets may cause coagulation abnormalities, thereby increasing the risk of bleeding. Thrombosis: In some cases, aggregated platelets may also cause thrombosis and increase the risk of cardiovascular disease. Inflammatory response: Some disease states or inflammatory responses may lead to an increase in aggregated platelets, reflecting the body's inflammatory state.

[0075] Existing hemacytometers utilize fluorescent staining, flow cytometry, Coulter impedance theory, and other principles to count and classify blood cells in mammals. These methods typically rely on the magnitude of electrical or optical signals, rather than image-based judgment, making it difficult to distinguish between aggregated platelets and other normal blood cells, such as red blood cells and white blood cells.

[0076] In the existing technology, only blood smear microscopy can provide a simple qualitative analysis of aggregated platelets. Blood smear microscopy is usually affected by manual smear pushing technology. In addition, due to the unevenness of the front and back of the smear and the large workload of counting, full smear counting is usually not performed, so the data repeatability of multiple smear pushes is poor. It is greatly affected by the ability of the interpreter. Moreover, due to the strong stress response of platelets, manual operation will affect their true state. Platelets are related to coagulation function, so they are very sensitive cells. There are many ways to be activated, such as low temperature, mechanical stress, and biochemical substances. For example, human platelets will be activated by ADP, arachidonic acid, collagen, etc. Therefore, after ex vivo, unless they are in a very suitable environment, they are not easy to activate and non-specific aggregation occurs. Therefore, the smear operation cannot truly reflect their true state.

[0077] In this application, blood suspension measurements are based on a relatively stable state where blood cells are immune to external forces and influences, preserving the true state of the blood to the greatest extent possible. Especially for sensitive parameters like platelet aggregation, the measurement conditions used in this application are uniquely and optimally close to the true state. Therefore, the platelet aggregation parameters provided are more accurate and reliable.

[0078] With the standardization and widespread use of platelet function assays, the ability to rapidly assess platelet function to support clinical decision making is increasing. In vitro platelet aggregation assays are becoming increasingly important in the assessment and intervention of acute trauma.

[0079] In the prior art, platelet function determination is also very troublesome and requires Figure 16 The dedicated optical analysis equipment shown has not been seen before. Platelet function analysis based on microscopic magnified images has not been seen before.

[0080] like Figure 1 A method for AI training to identify aggregated platelets comprises preprocessing a blood sample to obtain a microscopic sample; laying the microscopic sample flat; allowing the platelets to settle to the bottom; photographing the laid microscopic sample to obtain a microscopic sample image; identifying and labeling normal platelets and aggregated platelets in the microscopic sample image to obtain a labeled image, using the labeled image for AI training to obtain an AI feature dataset A; the AI ​​feature dataset A includes normal platelet features, and the AI ​​feature dataset A includes aggregated platelet features.

[0081] like Figure 2 ,exist Figure 1 On the basis of the scheme, AI feature data set A is used as the feature data set, and AI software is used to identify one or more pictures input into AI to obtain recognition output pictures, which include recognized normal platelet images and agglutinated platelet images. The recognized output agglutinated platelet images are manually reviewed to form manual review mark pictures, which include marked normal platelet images and agglutinated platelet images. AI training is performed using manual review mark pictures to obtain AI feature data set B, which includes normal platelet features and agglutinated platelet features. The above-mentioned AI software runs locally or on a network server, and the above-mentioned feature data sets are placed locally or on a network server.

[0082] Figure 1 In the , only one round of training was performed. Figure 2 The training can be repeated multiple times. Figure 2 In this process, AI recognition and verification are performed using AI feature dataset A, identifying and outputting normal platelets and aggregated platelets. The identified aggregated platelets are manually reviewed, labeled as normal platelets and aggregated platelets, and a second AI training session is conducted on this basis. The resulting iterative AI feature dataset B enriches the dataset and reduces the probability of misidentification of aggregated platelets. By adding manual review and secondary AI training, through multiple iterations, the feature dataset can be continuously optimized and recognition accuracy improved.

[0083] like Figure 3 , perform image filtering on the above-mentioned microscopic sample images to filter out the red blood cell images; the red blood cells can be filtered out based on the color difference between the red blood cell and platelet images; or the red blood cells can be filtered out based on the grayscale difference between the red blood cell and platelet images. Agglutinated platelets 10 9 / L and red blood cell count 10 12 / L are not of the same order of magnitude. Identifying one aggregated platelet requires identifying a large number of red blood cells at the same time, which is inefficient. By utilizing the staining differences between red blood cells and aggregated platelets, image processing methods are used to initially screen aggregated platelets, and then suspected platelets are identified, which reduces the complexity of the system, reduces the image processing links, and greatly improves the recognition efficiency.

[0084] The pretreatment of blood samples can be done by adding dyes to the blood samples, staining the blood samples, and then diluting the blood samples to obtain microscopic examination samples. Alternatively, liquid dyes can be added to the blood samples, staining the blood samples, and then diluting the blood samples to obtain microscopic examination samples. Figure 4 , diluent staining solution can be added to the blood sample to perform staining and dilution at the same time.

[0085] like Figure 5, platelet agonists can be added after dilution and staining. Figure 6 After adding the platelet agonist, the dilution staining solution can be added for staining and dilution.

[0086] Pretreatment of blood samples, e.g. Figure 7 , blood samples can be centrifuged and layered to obtain microscopic samples, which are platelet-containing plasma (PRP). Figure 8 , a platelet agonist can be added to platelet-containing plasma (PRP).

[0087] Pretreatment of blood samples, e.g. Figure 9 The blood sample is centrifuged and layered to obtain an intermediate sample, which is platelet-containing plasma (PRP). The intermediate sample is added to the blood sample to obtain a microscopic examination sample; the microscopic examination sample contains red blood cells. Different amounts of red blood cells from the blood sample can be added as needed.

[0088] like Figures 5 to 9 Platelet agonists can be added to the blood sample, intermediate sample, and / or microscopic sample to promote platelet aggregation. Platelet agonists include one or more of epinephrine, adenosine diphosphate (ADP), collagen, ristocetin, and arachidonic acid (AA).

[0089] like Figure 4 After the microscopic sample is prepared, wait for TW before taking the photo. The range of TW is 1 minute to 5 minutes.

[0090] Multiple blood samples can be pooled to obtain a sample for microscopic examination. Figure 10 , the above blood samples come from multiple animal individuals. Figure 11 The samples were derived from one or more individuals with aggregated thrombocytosis. When multiple blood samples were pooled, some pretreatment and screening steps were omitted. These pretreatment and screening steps can prevent aggregation of blood cells from different individuals. For example, blood samples with the same blood type were selected.

[0091] A computing and processing device is used to run all or part of the above-mentioned AI training method for identifying aggregated platelets; the memory of the computing and processing device includes the above-mentioned AI feature data set A; the memory of the computing and processing device includes the above-mentioned AI feature data set B.

[0092] A data storage device storing program code for executing all or part of the aforementioned AI training method for identifying aggregated platelets; storing the aforementioned AI feature dataset A; storing the aforementioned AI feature dataset B. A detection device for executing part or all of the aforementioned AI training method for identifying aggregated platelets.

[0093] like Figure 13 and Figure 14 A method for detecting platelet aggregation parameters comprises the following steps: pre-processing a blood sample to obtain a microscopic sample; flattening the microscopic sample; photographing the flattened microscopic sample to obtain a microscopic sample image; using an AI recognition algorithm to recognize the microscopic sample image, identifying aggregated platelets in the blood sample in a selected area S1 of the image, and obtaining a total number NUMS1 of aggregated platelets in the blood sample in the selected image; and the AI ​​recognition algorithm identifies aggregated platelets in the blood sample based on an aggregated platelet feature dataset.

[0094] The above-mentioned aggregated platelet feature dataset is obtained by training after manually annotating aggregated platelet images; or the aggregated platelet feature dataset is obtained by training after images of normal platelets are obtained after being stimulated and aggregated by platelet agonists.

[0095] A platelet aggregation parameter is defined for evaluating the number of aggregated platelets in a blood sample. The number of aggregated platelets per unit volume of blood sample = the number of aggregated platelets per unit volume of blood sample / the volume of blood sample.

[0096] The above-mentioned platelet aggregation parameter is the number of aggregated platelets in a unit volume of blood sample. The above-mentioned blood sample has been diluted N times after dilution pretreatment; the tile height of the above-mentioned microscopic sample is H; the volume V corresponding to the total number of aggregated platelets NUMS1 in the above-mentioned blood sample = selected area S1×H; the number of aggregated platelets in a unit volume of blood sample = N×the total number of aggregated platelets NUMS1 in the blood sample / blood sample volume V.

[0097] The above-mentioned platelet aggregation parameter is the number of aggregated platelets in a unit volume of blood sample. The above-mentioned blood sample is concentrated N times after concentration pretreatment, and the tile height of the above-mentioned microscopic sample is H; the volume V corresponding to the total number of aggregated platelets NUMS1 in the above-mentioned blood sample = the selected area S1×H; the number of aggregated platelets in a unit volume of blood sample = the total number of aggregated platelets in the blood sample NUMS1 / (blood sample volume V×N).

[0098] A platelet aggregation parameter is used to assess the number of aggregated platelets in a blood sample, characterized by: aggregated platelet ratio = number of aggregated platelets in an equal volume of blood sample / number of normal platelets. An AI recognition algorithm is used to identify the microscopic sample image, identify normal platelets in the blood sample within a selected area S1 of the image, and obtain the total number of normal platelets NUMS2 in the blood sample in the selected image. The AI ​​recognition algorithm identifies normal platelets in the blood sample based on a dataset of normal platelet characteristics in the blood sample. The ratio of the aggregated platelet number to the normal platelet number is NUM1 / NUM2.

[0099] Blood sample pretreatment includes diluting the blood sample with a liquid dye to obtain a microscopic specimen, with the dilution factor being greater than 50 and less than 800. Microscopic examination and imaging are completed within time T13 after the microscopic specimen is prepared. Time T13 ranges from 10 minutes to 30 minutes. Time T13 can also range from 10 minutes to 120 minutes.

[0100] like Figure 3 The above-mentioned microscopic sample images are filtered to remove red blood cell images. Red blood cells can be filtered out based on the color difference between red blood cell and platelet images, or based on the grayscale difference between red blood cell and platelet images. Agglutinated platelets and red blood cells are not of the same magnitude in number. Identifying a single agglutinated platelet requires simultaneously identifying a large number of red blood cells, which is inefficient. By utilizing the staining difference between red blood cells and agglutinated platelets, image processing methods are used to initially screen for agglutinated platelets, and then identify suspected platelets. This reduces system complexity, reduces image processing steps, and significantly improves recognition efficiency.

[0101] like Figure 12 and Figure 15 , Figure 12 This is a diagram of a blood sample containing aggregated platelets when imaged in a test chip. Figure 15 As can be seen in the picture, there are red blood cells, platelets, and aggregated platelets.

[0102] A computing device for executing all or part of the above-mentioned platelet aggregation parameter detection method. A data storage device for storing all or part of the program code for executing the above-mentioned platelet aggregation parameter detection method. A detection device for executing part or all of the above-mentioned platelet aggregation parameter detection method.

[0103] The method of the present application can obtain accurate quantitative parameters of aggregated platelets, and can also obtain accurate parameters of aggregated platelets relative to normal platelets. The image-based identification of aggregated platelets has high accuracy. Whether the platelet aggregation rate is reduced or increased, the quantitative analysis of aggregated platelet parameters can provide a closer to real display, providing more diverse, rich and accurate auxiliary diagnostic information for clinical practice.

[0104] Although the present invention is illustrated and described based on the preferred embodiment and several alternatives, the invention is not limited by the specific description in this specification. Other additional replacement or equivalent components can also be used to practice the present invention.

Claims

1. A method for AI training to identify aggregated platelets, characterized in that: Blood samples were pre-processed to obtain microscopic specimens; Spread the sample for microscopic examination flatly; let the platelets settle to the bottom; photographing the flattened microscopic sample to obtain an image of the microscopic sample; Identify and label normal platelets and aggregated platelets in microscopic sample images to obtain labeled images. Use the labeled images for AI training to obtain AI feature dataset A. The AI ​​feature dataset A includes normal platelet features, and the AI ​​feature dataset A includes aggregated platelet features.

2. The AI ​​training method for identifying aggregated platelets according to claim 1, characterized in that: Pretreatment of blood samples, including any of the following technical features: TA1: Adding a dye to the blood sample, staining the blood sample, and diluting the blood sample after staining to obtain a microscopic examination sample; TA2: Add liquid dye to the blood sample, stain and dilute the blood sample to obtain a microscopic sample; TA3: The blood sample is centrifuged and layered to obtain a microscopic sample, which is platelet-containing plasma (PRP); TA4: The blood sample is centrifuged and layered to obtain an intermediate sample, which is platelet-containing plasma (PRP). The intermediate sample is added to the blood sample to obtain a microscopic sample; the microscopic sample contains red blood cells.

3. The AI ​​training method for identifying aggregated platelets according to claim 2, characterized in that: The blood sample, intermediate sample and / or microscopic sample are added with a platelet agonist, which promotes platelet aggregation.

4. The AI ​​training method for identifying aggregated platelets according to claim 3, characterized in that: Platelet agonists include one or more of epinephrine, adenosine diphosphate (ADP), collagen, ristocetin, and arachidonic acid (AA).

5. The AI ​​training method for identifying aggregated platelets according to claim 1, characterized in that: Include any one of the following technical features: TB1: performing image filtering processing on the microscopic sample image to filter out the red blood cell image; TB2: performing image filtering processing on the microscopic sample image to filter out the red blood cell image, wherein the red blood cells are filtered out based on the color difference between the red blood cell and platelet images; TB3: performing image filtering processing on the microscopic sample image to filter out the red blood cell image, wherein the red blood cells are filtered out based on the grayscale difference between the red blood cell and platelet images.

6. The AI ​​training method for identifying aggregated platelets according to claim 1, characterized in that: Pretreatment of blood samples, including any of the following technical features: TC1: After the microscopic sample is prepared, wait for TW before taking the photo; TC2: The blood samples come from multiple animal individuals; TC3: Mix multiple blood samples to obtain microscopic samples; TC4: The sample is derived from one or more individuals with coagulated thrombocytosis.

7. The AI ​​training method for identifying aggregated platelets according to any one of claims 1 to 6, characterized in that: Using AI feature dataset A as the feature dataset, AI software is used to identify one or more images input into the AI ​​recognition image to obtain a recognition output image, wherein the recognition output image includes a recognized normal platelet image and an agglutinated platelet image, and the recognition output agglutinated platelet image is manually reviewed to form a manually reviewed marked image, wherein the manually reviewed marked image includes a marked normal platelet image and an agglutinated platelet image. AI training is performed using the manually reviewed marked image to obtain AI feature dataset B, wherein AI feature dataset B includes normal platelet features, and AI feature dataset B includes agglutinated platelet features; The AI ​​software runs on a local or network server, and the feature data set is placed on a local or network server.

8. A computing and processing device, characterized in that: Include any one of the following technical features: TD1: used to run all or part of the AI ​​training method for identifying aggregated platelets according to any one of claims 1 to 9; TD2: The memory of the computing and processing device includes the AI ​​feature dataset A according to any one of claims 1 to 9; TD3: The memory of the computing and processing device includes the AI ​​feature data set B described in claim 9.

9. A data storage device, characterized in that Include any one of the following technical features: TE1: stores program codes for executing all or part of the AI ​​training method for identifying aggregated platelets according to any one of claims 1 to 9; TE2: stores the AI ​​feature dataset A described in any one of claims 1 to 9; TE3: stores the AI ​​feature dataset B described in any one of claim 9.

10. A detection device, characterized in that: Used to execute part or all of the AI ​​training method for identifying aggregated platelets according to any one of claims 1 to 9.

11. A method for detecting platelet aggregation parameters, characterized in that: Blood samples were pre-processed to obtain microscopic specimens; Laying the microscopic examination sample flat; photographing the laid microscopic examination sample to obtain an image of the microscopic examination sample; An AI recognition algorithm is used to identify the microscopic sample image, identify the aggregated platelets in the blood sample in the selected area S1 of the image, and obtain the total number NUMS1 of aggregated platelets in the blood sample in the selected image; the AI ​​recognition algorithm identifies the aggregated platelets in the blood sample based on the aggregated platelet feature data set.

12. The method for detecting platelet aggregation parameters according to claim 11, wherein: The aggregated platelet feature dataset is obtained by training after manually annotating aggregated platelet images; or the aggregated platelet feature dataset is obtained by training after normal platelets are aggregated by platelet agonist stimulation.

13. The method for detecting platelet aggregation parameters according to claim 11, characterized in that: Include any one of the following technical features: TG1: The platelet aggregation parameter is the number of aggregated platelets per unit volume of blood sample, where the blood sample is diluted N-fold after dilution pretreatment; the microscopic sample tile height is H; the total number of aggregated platelets NUMS1 in the blood sample corresponds to a volume V = selected area S1 × H; the number of aggregated platelets per unit volume of blood sample = N × the total number of aggregated platelets NUMS1 in the blood sample / blood sample volume V; TG2: The platelet aggregation parameter is the number of aggregated platelets per unit volume of blood sample. The blood sample is concentrated N-fold after concentration pretreatment, and the microscopic sample is tiled at a height of H. The total number of aggregated platelets NUMS1 in the blood sample corresponds to a volume V = selected area S1 × H. The number of aggregated platelets per unit volume of blood sample = the total number of aggregated platelets in the blood sample NUMS1 / (blood sample volume V × N).

14. The method for detecting platelet aggregation parameters according to claim 11, wherein: An AI recognition algorithm is used to identify the microscopic sample image, identify normal platelets in the blood sample in a selected area S1 of the image, and obtain the total number NUMS2 of normal platelets in the blood sample in the selected image; the AI ​​recognition algorithm identifies normal platelets in the blood sample based on a characteristic data set of normal platelets in the blood sample; the ratio of the number of aggregated platelets to the number of normal platelets = NUM1 / NUM2.

15. The method for detecting platelet aggregation parameters according to claim 11, characterized in that: The blood sample pretreatment includes diluting the blood sample with a liquid dye to obtain a microscopic examination sample, where the dilution multiple is greater than 50 and less than 800.

16. The method for detecting platelet aggregation parameters according to claim 11, wherein: Complete microscopic examination and photography within time T13 after the microscopic examination sample is prepared.

17. The method for detecting platelet aggregation parameters according to claim 13, wherein: Include any one of the following technical features: TK1: performing image filtering processing on the microscopic sample image to filter out the red blood cell image; TK2: performing image filtering processing on the microscopic sample image to filter out the red blood cell image, wherein the red blood cells are filtered out based on the color difference between the red blood cell and platelet images; TK3: performing image filtering processing on the microscopic sample image to filter out the red blood cell image, wherein the red blood cell is filtered out based on the grayscale difference between the red blood cell image and the platelet image.

18. A computing and processing device, characterized in that: Used to perform all or part of the method for detecting platelet aggregation parameters according to any one of claims 11 to 17.

19. A data storage device, characterized in that: Storing all or part of the program codes for executing the method for detecting platelet aggregation parameters according to any one of claims 11 to 17.

20. A detection device, characterized in that: Used to run part or all of the method for detecting platelet aggregation parameters according to any one of claims 11 to 17.

21. A platelet aggregation parameter for evaluating the number of aggregated platelets in a blood sample, characterized in that: The number of aggregated platelets per unit volume of blood sample = the number of aggregated platelets per unit volume of blood sample / the volume of blood sample.

22. A platelet aggregation parameter for evaluating the number of aggregated platelets in a blood sample, characterized in that: The proportion of aggregated platelets = the number of aggregated platelets in the same volume of blood sample / the number of normal platelets.

Images